i

MENTAL HEALTH AND SUBSTANCE-RELATED TREATMENT UTILIZATION,
DROPOUT, AND CONTINUITY OF CARE AMONG DETAINED ADOLESCENTS:
A 14-YEAR LONGITUDINAL STUDY

A Dissertation
Submitted to the Faculty
of
Purdue University
by
Laura Morgan White

In Partial Fulfillment of the
Requirements for the Degree
of
Doctor of Philosophy

August 2015
Purdue University
Indianapolis, Indiana

ii

For my mother, who is and always will be my role model. You continue to be the ideal
exemplar of a strong, intelligent, and independent woman. Thank you for your endless
love, encouragement, validation, and patience.

iii

ACKNOWLEDGEMENTS

This dissertation would not be possible without the guidance of several
remarkable individuals. First and foremost, I would like to express my deepest gratitude
to my advisor, Dr. John McGrew, for his invaluable support and assistance during my
years in graduate school and in completing this dissertation. His mentorship was
paramount to providing a well-rounded education that allowed me to cultivate my
individuality and self-sufficiency as a research and clinician. Second, I want to
acknowledge my research mentor, Dr. Matthew Aalsma, one of the most brilliant,
humorous, and creative people with whom I have had the pleasure of knowing. His
passion for social justice, in conjunction with genuine humor and openness to different
ideas, sparked my interest in research and served as the primary motivation for this
project. I have little doubt that his influence has made me a better person, as well as an
independent thinker and passionate advocate for vulnerable populations. Third, I want to
wholeheartedly thank the other members of my committee, Dr. Alan McGuire and Dr.
Michelle Salyers, for their many suggestions and contributions to this dissertation. Thank
you for being so flexible and understanding with my tendency to complete tasks lastminute. Finally, I would like to thank my family, particularly my mother. An
unwavering source of support, my mother has always encouraged me and told me I could
accomplish anything. Thanks mom for always believing in me. This dissertation would
not exist without you.

iv

TABLE OF CONTENTS

Page
LIST OF TABLES ............................................................................................................ vii
LIST OF FIGURES ......................................................................................................... xiv
ABSTRACT ...................................................................................................................... xv
CHAPTER 1.

INTRODUCTION .................................................................................... 1

1.1

Detained Adolescents.............................................................................................. 1

1.2

Mental Health Concerns of Detained Adolescents ................................................. 3

1.3

Substance-Related Concerns of Detained Adolescents .......................................... 5

1.4

Why Care about these Concerns among Detained Adolescents? ........................... 7

1.5

Treatment Services for Detained Adolescents ........................................................ 8

1.6

Prevalence of Treatment Use among Detained Adolescents ................................ 10

1.7

Factors Related to Treatment Use among Detained Adolescents ......................... 12

1.8

Timing of Treatment Services and Continuity of Care ......................................... 18

1.9

Limitations of the Literature ................................................................................. 20

1.10

Purpose of Study ................................................................................................... 22

1.11

Study Aims and Research Questions .................................................................... 23
1.11.1

Research Questions for Aim 1 ................................................................24

1.11.2

Research Question for Aim 2 .................................................................27

1.11.3

Research Question for Aim 3 .................................................................27

1.11.4

Research Questions for Aim 4 ................................................................28

CHAPTER 2. METHOD ............................................................................................... 31
2.1

Data Abstraction and Data Linking ...................................................................... 31

2.2

Data Coding .......................................................................................................... 33

v
Page
2.3

Measures ............................................................................................................... 39

2.4

Recruitment and Participants ................................................................................ 41

2.5

Study Setting ......................................................................................................... 42

2.6

Data Analysis ........................................................................................................ 42

2.7

2.6.1

Data Analysis for Aim 1 .........................................................................44

2.6.2

Data Analysis for Aim 2 .........................................................................45

2.6.3

Data Analysis for Aim 3 .........................................................................46

2.6.4

Data Analysis for Aim 4 .........................................................................49

Statistical Power Analysis .....................................................................................52

CHAPTER 3.

RESULTS .............................................................................................. 54

3.1

Sample Demographics .......................................................................................... 54

3.2

Criminal History Outcomes .................................................................................. 55

3.3

Mental Health and Substance-Related Outcomes ................................................. 59

3.4

Prevalence of Pre-Detention Treatment Utilization .............................................. 62

3.5

Aim 1. Prevalence of Post-Detention Treatment Utilization ................................ 63
3.5.1

3.6

Aim 1. Quantity of Post-Detention Treatment Services ........................................68
3.6.1

3.7

Aim 1. Treatment Dropouts and Treatment Gaps by Groups .................77

Aim 3. Predictors of Treatment Utilization ...........................................................80
3.10.1

3.11

Aim 2. Intensity of Post-Detention Treatment Services by Groups .......74

Aim 1. Treatment Dropouts and Treatment Gaps .................................................77
3.9.1

3.10

Aim 2. Frequency of Post-Detention Treatment Services by Groups ....72

Aim 1. Intensity of Post-Detention Treatment Services ........................................73
3.8.1

3.9

Aim 2. Quantity of Post-Detention Treatment Services by Groups .......69

Aim 1. Frequency of Post-Detention Treatment Services .....................................71
3.7.1

3.8

Aim 2. Prevalence of Post-Detention Treatment Utilization by Groups 64

Aim 3. Predictors of Treatment Utilization by Cohorts .........................84

Aim 4. Patterns of Treatment Utilization: Time to Treatment ..............................90
3.11.1

Aim 4. Time to Treatment by Groups ....................................................91

3.11.2

Time to Treatment: Cox Regression Analyses .......................................94

vi
Page
3.11.3
3.12

Time to Treatment by Cohorts ................................................................96

Aim 4. Patterns of Treatment Utilization: Time to Termination .........................101
3.12.1

Time to Termination by Groups ...........................................................102

3.12.2

Time to Termination: Cox Regression Analyses ..................................104

3.12.3

Time to Termination by Cohorts ..........................................................106

CHAPTER 4.

DISCUSSION ...................................................................................... 110

4.1

Discussion ........................................................................................................... 110

4.2

Criminal History Outcomes ................................................................................ 110

4.3

Mental Health Outcomes .................................................................................... 113

4.4

Aim 1. Post-Detention Treatment Utilization ..................................................... 116

4.5

Aim 4. Patterns of Post-Detention Treatment Utilization................................... 119

4.6

Aim 1. Quantity of Post-Detention Treatment Services ..................................... 121

4.7

Aim 2. Group Differences: Demographic Groups .............................................. 125

4.8

Aim 2. Group Differences: Mental Health Groups ............................................. 130

4.9

Aim 2. Group Differences: Criminal History Groups ........................................ 134

4.10

Aim 3. Predictors of Treatment Utilization ........................................................ 137

4.11

Time Effect ......................................................................................................... 138

4.12

Study Limitations ................................................................................................ 140

4.13

Implications and Recommendations for Future Research .................................. 141

REFERENCES ............................................................................................................... 145
TABLES ......................................................................................................................... 160
FIGURES ........................................................................................................................ 255
APPENDIX ..................................................................................................................... 278
VITA ............................................................................................................................... 289

vii

LIST OF TABLES

Table ..............................................................................................................................Page
Table 3.1

Sample Demographics ............................................................................ 160

Table 3.2

Correlation Matrix: Demographic and Criminal History Variables ....... 161

Table 3.3

Prevalence of Detained Adolescents and Length of Stay for First
Detention Stay......................................................................................... 163

Table 3.4

Criminal History Outcomes: Criminal Charges ...................................... 164

Table 3.5

Mental Health and Criminal History Outcomes by Gender, Race,
and Age Cohorts ..................................................................................... 165

Table 3.6

Post-Detention Criminal History Outcomes: Recidivism,
Detention, and Incarceration ................................................................... 167

Table 3.7

Massachusetts Youth Screening Instrument-2nd Edition (MAYSI2) among Cohort Two ............................................................................. 168

Table 3.8

Correlation Matrix: Demographic and Mental Health Variables ........... 169

Table 3.9

Mental Disorders and Substance-Related Disorders among Sample ...... 170

Table 3.10

Prevalence of Pre-Detention and During Detention Treatment
Utilization: Subdivided by Treatment Type and Treatment Setting ....... 171

Table 3.11

Correlation Matrix: Demographics, Pre-Detention Treatment
Utilization, and During Detention Treatment Utilization ....................... 172

Table 3.12

Prevalence of Post-Detention Treatment Utilization: Subdivided by
Treatment Type and Treatment Setting .................................................. 174

Table 3.13

Correlation Matrix: Demographics and Post-Detention Treatment
Utilization ............................................................................................... 175

viii
Table
Table 3.14

Page
Prevalence of Post-Detention Treatment Utilization within Two
Years Post-Detention: Demographic Groups.......................................... 176

Table 3.15

Correlation Matrix: Mental Health Variables and Post-Detention
Treatment Utilization Variables .............................................................. 177

Table 3.16

Prevalence of Treatment Utilization within Two Years PostDetention: Mental Health Groups ........................................................... 179

Table 3.17

Correlation Matrix: Pre-Detention, During Detention, and PostDetention Treatment Utilization ............................................................. 180

Table 3.18

Correlation Matrix: Criminal History Variables and Treatment
Utilization within Two Years Post-Detention......................................... 182

Table 3.19

Prevalence of Treatment Utilization within Two Years PostDetention: Criminal History Groups ....................................................... 184

Table 3.20

Prevalence of Treatment Utilization within Two Years PostDetention: Cohort One versus Cohort Two ............................................ 185

Table 3.21

Quantity of Treatment Services within Two Years Post-Detention:
Demographic Groups .............................................................................. 186

Table 3.22

Quantity of Treatment Services within Two Years Post-Detention:
Mental Health Groups ............................................................................. 187

Table 3.23

Quantity of Treatment Services within Two Years Post-Detention:
Criminal History Groups......................................................................... 188

Table 3.24

Quantity of Treatment Services within Two Years Post-Detention:
Cohort One versus Cohort Two .............................................................. 189

Table 3.25

Frequency of Treatment Services within Two Years PostDetention ................................................................................................. 190

Table 3.26

Frequency of Treatment Services within Two Years PostDetention: Demographic Groups ............................................................ 191

Table 3.27

Frequency of Treatment Services within Two Years PostDetention: Mental Health Groups ........................................................... 192

ix
Table
Table 3.28

Page
Frequency of Treatment Services within Two Years PostDetention: Criminal History Groups ....................................................... 193

Table 3.29

Frequency of Treatment Services within Two Years PostDetention: Cohort One versus Cohort Two ............................................ 194

Table 3.30

Intensity of Treatment Services within Two Years Post-Detention:
Treatment Type, Treatment Setting, and User Intensity Level ............... 195

Table 3.31

Intensity of Treatment Services within Two Years Post-Detention:
Demographic Groups .............................................................................. 196

Table 3.32

Intensity of Treatment Services within Two Years Post-Detention:
Mental Health Groups ............................................................................. 198

Table 3.33

Intensity of Treatment Services within Two Years Post-Detention:
Criminal History Groups......................................................................... 200

Table 3.34

Intensity of Treatment Services within Two Years Post-Detention:
Cohort One versus Cohort Two .............................................................. 202

Table 3.35

Outpatient Treatment Dropouts and Treatment Gaps within Two
Years Post-Detention .............................................................................. 203

Table 3.36

Outpatient Treatment Dropouts and Treatment Gaps within Two
Years Post-Detention: Demographic Groups.......................................... 204

Table 3.37

Outpatient Treatment Dropouts and Treatment Gaps within Two
Years Post-Detention: Mental Health Groups ........................................ 205

Table 3.38

Outpatient Treatment Dropouts and Treatment Gaps within Two
Years Post-Detention: Criminal History Groups .................................... 206

Table 3.39

Outpatient Treatment Dropouts and Treatment Gaps within Two
Years Post-Detention: Cohort One versus Cohort Two .......................... 207

Table 3.40

Summary of Results: Group Differences in Prevalence and
Quantity of Treatment Services .............................................................. 208

Table 3.41

Summary of Results: Group Differences in Treatment Intensity,
Dropouts, and Gaps ................................................................................. 210

Table 3.42

Multicollinearity Statistics of Predictor Variables .................................. 212

x
Table
Table 3.43

Page
Model Summary Statistics for Logistic Regression Models
Predicting Treatment Utilization within Two Years Post-Detention ...... 213

Table 3.44

Odds Ratios and Associated Statistics for Logistic Regression
Models Predicting Treatment and Mental Health Treatment within
Two Years Post-Detention ...................................................................... 214

Table 3.45

Odds Ratios and Associated Statistics for Logistic Regression
Models Predicting Substance-Related Treatment and Either/Both
Treatment within Two Years Post-Detention ......................................... 215

Table 3.46

Odds Ratios and Associated Statistics for Logistic Regression
Models Predicting Outpatient Treatment and Non-Outpatient
Treatment within Two Years Post-Detention ......................................... 216

Table 3.47

Model Summary Statistics for Logistic Regression Predicting
Treatment within Two Years Post-Detention: Cohort One .................... 217

Table 3.48

Predictors Associated with Increased Likelihood (+) and
Decreased Likelihood (–) of Post-Detention Treatment Utilization:
Total Sample, Cohort One, and Cohort Two .......................................... 218

Table 3.49

Odds Ratios and Associated Statistics for Logistic Regression
Models Predicting Treatment and Mental Health Treatment within
Two Years Post-Detention: Cohort One ................................................. 219

Table 3.50

Odds Ratios and Associated Statistics for Logistic Regression
Models Predicting Substance-Related Treatment and Either/Both
Treatment within Two Years Post-Detention: Cohort One .................... 220

Table 3.51

Odds Ratios and Associated Statistics for Logistic Regression
Models Predicting Outpatient Treatment and Non-Outpatient
Treatment within Two Years Post-Detention: Cohort One .................... 221

Table 3.52

Model Summary Statistics for Logistic Regression Models
Predicting Treatment within Two Years Post-Detention: Cohort
Two ......................................................................................................... 222

xi
Table
Table 3.53

Page
Odds Ratios and Associated Statistics for Logistic Regression
Models Predicting Treatment and Mental Health Treatment within
Two Years Post-Detention: Cohort Two ................................................ 223

Table 3.54

Odds Ratios and Associated Statistics for Logistic Regression
Models Predicting Substance-Related Treatment and Either/Both
Treatment within Two Years Post-Detention: Cohort Two .................... 224

Table 3.55

Odds Ratios and Associated Statistics for Logistic Regression
Models Predicting Outpatient Treatment and Non-Outpatient
Treatment within Two Years Post-Detention: Cohort Two .................... 225

Table 3.56

Life Table: Time to First Post-Detention Treatment Utilization ............ 226

Table 3.57

Kaplan-Meier Survival Analyses for Time to First Treatment:
Demographic, Mental Health, and Criminal History Groups ................. 227

Table 3.58

Model Summary Statistics for Cox Regression Models Predicting
Time to First Post-Detention Treatment ................................................. 229

Table 3.59

Hazard Ratios and Associated Statistics for Cox Regression
Models Predicting Time to First Treatment and First Mental Health
Treatment ................................................................................................ 230

Table 3.60

Hazard Ratios and Associated Statistics for Cox Regression
Models Predicting Time to First Substance-Related Treatment and
First Either/Both Treatment .................................................................... 231

Table 3.61

Hazard Ratios and Associated Statistics for Cox Regression
Models Predicting Time to First Outpatient Treatment and First
Non-Outpatient Treatment ...................................................................... 232

Table 3.62

Model Summary Statistics for Cox Regression Models Predicting
Time to First Treatment: Cohort One ..................................................... 233

Table 3.63

Predictors of Cox Regression Analyses Associated with Increased
Risk (+) and Decreased Risk (–) of Time to First Treatment: Total
Sample, Cohort One, and Cohort Two.................................................... 234

xii
Table
Table 3.64

Page
Hazard Ratios and Associated Statistics for Cox Regression
Models Predicting Time to First Treatment and First Mental Health
Treatment: Cohort One ........................................................................... 235

Table 3.65

Hazard Ratios and Associated Statistics for Cox Regression
Models Predicting Time to First Substance-Related Treatment and
First Either/Both Treatment: Cohort One ............................................... 236

Table 3.66

Hazard Ratios and Associated Statistics for Cox Regression
Models Predicting Time to First Outpatient Treatment and First
Non-Outpatient Treatment: Cohort One ................................................. 237

Table 3.67

Model Summary Statistics for Cox Regression Models Predicting
Time to First Treatment: Cohort Two ..................................................... 238

Table 3.68

Hazard Ratios and Associated Statistics for Cox Regression
Models Predicting Time to First Treatment and First Mental Health
Treatment: Cohort Two........................................................................... 239

Table 3.69

Hazard Ratios and Associated Statistics for Cox Regression
Models Predicting Time to First Substance-Related Treatment and
First Either/Both Treatment: Cohort Two............................................... 240

Table 3.70

Hazard Ratios and Associated Statistics for Cox Regression
Models Predicting Time to First Outpatient Treatment and First
Non-Outpatient Treatment: Cohort Two ................................................ 241

Table 3.71

Life Table: Time to Termination from Outpatient Treatment ................ 242

Table 3.72

Time to Termination from Outpatient Treatment: Demographic,
Mental Health, and Criminal History Groups ......................................... 243

Table 3.73

Model Summary Statistics for Cox Regression Models Predicting
Time to Termination from Outpatient Treatment ................................... 245

Table 3.74

Hazard Ratios and Associated Statistics for Cox Regression
Models Predicting Time to Termination from Outpatient Treatment
and from Mental Health Outpatient Treatment ....................................... 246

xiii
Table
Table 3.75

Page
Hazard Ratios and Associated Statistics for Cox Regression
Models Predicting Time to Termination from Substance-Related
Outpatient Treatment and from Either/Both Outpatient Treatment ........ 247

Table 3.76

Model Summary Statistics for Cox Regression Models Predicting
Time to Termination from Outpatient Treatment: Cohort One .............. 248

Table 3.77

Predictors in Cox Regression Analyses Associated with Increased
Risk (+) and Decreased Risk (–) of Time to Termination from
Outpatient Treatment: Total Sample, Cohort One, and Cohort Two ...... 249

Table 3.78

Hazard Ratios and Associated Statistics for Cox Regression
Models Predicting Time to Termination from Outpatient Treatment
and from Mental Health Outpatient Treatment: Cohort One .................. 250

Table 3.79

Hazard Ratios and Associated Statistics for Cox Regression
Models Predicting Time to Termination from Substance-Related
Outpatient Treatment and from Either/Both Outpatient Treatment:
Cohort One .............................................................................................. 251

Table 3.80

Model Summary Statistics for Cox Regression Models Predicting
Time to Termination from Outpatient Treatment: Cohort Two.............. 252

Table 3.81

Hazard Ratios and Associated Statistics for Cox Regression
Models Predicting Time to Termination from Outpatient Treatment
and from Mental Health Outpatient Treatment: Cohort Two ................. 253

Table 3.82

Hazard Ratios and Associated Statistics for Cox Regression
Models Predicting Time to Termination from Substance-Related
Outpatient Treatment and from Either/Both Outpatient Treatment:
Cohort Two ............................................................................................. 254

Appendix Table
Table A.1.

Codebook for Variables, Variable Type, and Levels .............................. 278

xiv

LIST OF FIGURES

Figure .....................................................................................................................

Page

Figure 3.1

Survival Curve: Total Sample ..................................................................255

Figure 3.2

Survival Curve: Treatment Type .............................................................256

Figure 3.3

Survival Curve: Treatment Setting ..........................................................257

Figure 3.4

Survival Curve: Gender ...........................................................................258

Figure 3.5

Survival Curve: Race ...............................................................................259

Figure 3.6

Survival Curve: Age Cohorts ...................................................................260

Figure 3.7

Survival Curve: MAYSI-2 .......................................................................261

Figure 3.8

Survival Curve: Disorder Type ................................................................262

Figure 3.9

Survival Curve: Pre-Detention Treatment ...............................................263

Figure 3.10

Survival Curve: Violent Offender ............................................................264

Figure 3.11

Survival Curve: Recidivism .....................................................................265

Figure 3.12

Survival Curve: Re-Detained ...................................................................266

Figure 3.13

Survival Curve for Treatment Termination: Total Sample ......................267

Figure 3.14

Survival Curve for Treatment Termination: Treatment Type..................268

Figure 3.15

Survival Curve for Treatment Termination: Gender ...............................269

Figure 3.16

Survival Curve for Treatment Termination: Race ...................................270

Figure 3.17

Survival Curve for Treatment Termination: Age Cohorts .......................271

Figure 3.18

Survival Curve for Treatment Termination: MAYSI-2 ...........................272

Figure 3.19

Survival Curve for Treatment Termination: Disorder Type ....................273

Figure 3.20

Survival Curve for Treatment Termination: Prior Treatment ..................274

Figure 3.21

Survival Curve for Treatment Termination: Violent Offender ................275

Figure 3.22

Survival Curve for Treatment Termination: Recidivism .........................276

Figure 3.23

Survival Curve for Treatment Termination: Re-Detained .......................277

xv

ABSTRACT

White, Laura Morgan. Ph.D., Purdue University, August 2015. Mental Health and
Substance-Related Treatment Utilization, Dropout, and Continuity of Care among
Detained Adolescents: A 14-Year Longitudinal Study. Major Professor: John H.
McGrew.

Although approximately 60%-80% of detained adolescents have a psychiatric
disorder, little is known about their utilization of mental health and substance-related
treatment services upon release from detention. Given that treatment can potentially
reduce symptomology and recidivism, the study examined detained adolescents’ postdetention treatment utilization and longitudinal patterns of use. Data were abstracted
from the electronic juvenile justice records and medical records of 9664 detained
adolescents (62.7% male; 34.8% White, 65.2% Black; 72.6% with disorder) with
Medicaid coverage held in a Midwestern detention center at some time during 19982011. A series of statistical tests (e.g., chi-square, ANOVA, logistic regression, KaplanMeier survival analyses, Cox regression) were conducted to identify group differences in
treatment utilization during the 14-year follow-up period. Following detention release,
approximately 66.2% of adolescents were re-arrested and 54.9% were re-detained or
incarcerated. Treatment utilization within two years post-detention was 36.7%; 31.4%
obtained mental health treatment, 10.4% obtained substance-related treatment, 36.0%
obtained outpatient treatment, and 6.2% obtained non-outpatient treatment. Among
treatment users, 22.5% dropped out of treatment within 1-3 sessions and 40.6%
experienced gaps (>45 days) between treatment services. Treatment utilization was
significantly higher among males, White (vs. Black) adolescents, younger adolescents,
violent (vs. non-violent) offenders, recidivists (vs. non-recidivists), and adolescents with
mental disorders (vs. substance-related disorders). Variables associated with increased

xvi
likelihood of post-detention treatment included: male gender, psychiatric disorder(s), predetention arrest(s), charge severity, violent offender, incarceration, and pre-detention
treatment; age and Black race were associated with decreased likelihood of treatment. As
one of the only longitudinal studies to examine treatment utilization among detained
adolescents upon community reentry, findings suggest limited service utilization, as well
as treatment gaps and disparities. Future research should focus on the treatment needs of
detained adolescents, factors associated with disparities, and programs/policies to ensure
consistent identification, referral, and connection to care for detained adolescents.

1

CHAPTER 1. INTRODUCTION

1.1

Detained Adolescents

Approximately 1.65 million adolescents (≥18 years) are arrested in the United
States each year (Office of Juvenile Justice and Delinquency Prevention, 2013). Of
these, about 20%, or roughly 330,000 youths, are placed in short-term detention centers
or long-term prison facilities. These adolescents represent a vulnerable population,
marked by high rates of behavioral health concerns, mental disorders, and substancerelated disorders (Shufelt & Cocozza, 2006; Wasserman, McReynolds, Lucas, Fisher, &
Santos, 2002). Despite evidence that these problems can interfere with rehabilitation and
successful reintegration into the community (Calley, 2012; Cottle, Lee, & Heilbrun,
2001), research is limited regarding this population’s access to and use of treatment
services to address mental health and/or substance-related problems (Kataoka, Zima,
Dupre, Moreno, Yang, & McCracken, 2001; Teplin, Abram, McClelland, Washburn, &
Pikus, 2005). The few studies that have examined treatment utilization among detained
adolescents (DAs) report inconsistent findings, such that conclusions about treatment
prevalence, types of services, and characteristics of treatment users remain uncertain
(Herz, 2001; Johnson, Cho, Fendrich, Graf, Kelly-Wilson, Pickup, 2004). Furthermore,
the majority of studies have examined treatment utilization cross-sectionally (Rawal,
Romansky, Jenuwine, & Lyons, 2004; Sedlak & McPherson, 2010), so patterns in service
use over time, particularly upon community reentry, are not well understood. Thus,
longitudinal research that tracks detained adolescents over time is needed to fully
understand patterns of treatment utilization for this population.
The detained adolescent population is defined as any youth (10-18 years) residing
in a correctional facility, such as a detention center or juvenile prison (Yazzie, 2011).

2
Typically, these youths are placed in correctional facilities after being arrested, charged,
and/or adjudicated of criminal offenses (Gupta, Kelleher, Pajer, & Cuellar, 2005). The
type of facility in which a detained adolescent is placed depends upon the status of the
youth’s criminal case. Specifically, detained adolescents (DAs) who have been arrested
and/or charged with an offense(s) are typically placed in detention centers for short-term
stays of about 2-4 weeks, as these adolescents await their appearances in court (Desai,
Goulet, Robbins, Chapman, Migdole, & Hoge, 2006; Yazzie, 2011). Following court
proceedings, about 4%, or roughly 70,000 adolescents, are convicted of serious crimes
and placed in juvenile prisons for long-term stays of several months to several years
(Bureau of Justice Statistic, 2015; Yazzie, 2011). The prevalence of DAs in the United
States is estimated to be about 224 per 100,000 adolescents, or 0.22% of youth
nationwide (Sedlak & Bruce, 2010). The number of DAs held in detention or prison total
approximately 80,000 to 100,000 adolescents on any given day in the United States
(Rawal et al., 2004).
Compared to the general adolescent population, DAs are disproportionately male
and racial/ethnic minorities (Henggeler & Schoenwald, 2011; Sedlak, 2009). In fact,
according to recent census data, at least 75%-80% of DAs are male and about 60%-70%
of DAs are Black, Hispanic, or another racial/ethnic minority (Office of Juvenile Justice
and Delinquency Prevention, 2013; Sedlak & Bruce, 2010). More than half of all
juveniles detained in detention or prison are 16-17 years old during their stay in a
juvenile justice facility (Sedlak, 2009). In addition, the majority of DAs are raised in
families of low socioeconomic status, marked by high rates of poverty and
unemployment (Dembo, Pacheco, Schmeidler, Ramirez-Garmica, Guida, & Rahman,
1997; Rogers, Zima, Powell, & Pumariega, 2001; Shelton, 2005). When not incarcerated,
at least half of all DAs live in poverty and receive some form of government assistance or
public aid (Johnson et al., 2004; Robertson, Dill, Hussain, & Undesser, 2004). Evidence
suggests that socioeconomic status, rather than race or ethnicity, may be the critical factor
underlying the disproportionate number of minority adolescents involved in the juvenile
justice system, since Black and Hispanic adolescents are more likely to live in poverty
(Braverman, & Murray, 2011).

3
In conjunction with low socioeconomic status, many DAs experience unstable
housing and frequent out-of-home placements in foster care, group homes, residential
treatment centers, and/or homes of extended family members (Maschi, Hatcher,
Schwalbe, Rosato, 2008; Riley, 2014; Shelton, 2001). According to a national
representative sample of DAs in the US, only one in three adolescents live in a twoparent household prior to incarceration and a sizable proportion of detained youth
experience homelessness at least once during their adolescence (Sedlak & Bruce, 2010).
Further, as many as 20%-25% of DAs are current or expectant parents, compared to only
2% of non-detained adolescent males and 6% of non-detained adolescent females
(Braverman, & Murray, 2011; Sedlak, 2009). Finally, evidence suggests that lesbian,
gay, bisexual, and transgendered youth are likely to be over-represented within the
detained adolescent population (Curtin, 2002; Schaffner, 1998). For example, one study
found that approximately 20%-33% of detained females identify as bisexual or lesbian in
sexual orientation, whereas only 4%-10% of females within the general adolescent
population identify as such (Schaffner, 1998).

1.2

Mental Health Concerns of Detained Adolescents

The detained adolescent population suffers from an increased incidence of poor
mental health, as evidenced by high rates of mental disorders, severe mental illness
(SMI), trauma history, suicidal behavior, and comorbidity of disorders (Sedlak &
McPherson, 2010; Shufelt & Cocozza, 2006). The majority of DAs meet criteria for at
least one mental disorder, with estimates ranging from 60-80% of the population (Teplin,
Abram, McClelland., Dulcan, Mericle, 2002; Teplin, Abram, McClelland, Mericle,
Dulcan, & Washburn, 2006; Teplin, Welty, Abram, Dulcan, & Washburn, 2012),
compared to only 15%-20% of the general adolescent population (Hoeve, McReynolds,
& Wasserman, 2013; Vincent, Grisso, Terry, & Banks, 2008). With respect to specific
disorders, a recent meta-analysis of 32 studies reported the following prevalence rates for
DAs: 10.6% of males and 29.2% of females have major depressive disorder, 11.7% of
males and 18.5% of females have attention deficit/hyperactivity disorder, and 52.8% of
males and 52.8% of females have a behavior-related disorder (e.g., conduct disorder,

4
oppositional defiant disorder) (Fazel, Doll, & Langstrom, 2008). Furthermore, as many
as half of the DA population meets criteria for an anxiety-related disorder, 2%-5% of
adolescents meet criteria for obsessive compulsive disorder, and about 12%-16% of
adolescents meet criteria for an eating disorder (Sedlak & Bruce, 2010; Teplin et al.,
2006). Given the high frequency of mental disorders among detained adolescents,
comordibity of disorders is common. Almost 50% of all DAs meet diagnostic criteria for
at least two disorders and 15%-25% of DAs meet criteria for three or more disorders
(Abram, Teplin, McClelland, Dulcan, 2003; Braverman & Murray, 2011).
Just as mental disorders are common among DAs, the risk for severe mental
illness (SMI), such as psychosis, delusional thinking, and early-onset schizophrenia, is
also disproportionately high within this population (Shufelt, & Cocozza, 2006). Detained
youth are about ten times more likely than non-detained peers to experience symptoms of
SMI, such as auditory or visual hallucinations, delusions, paranoia, and disordered
thinking (Bath, Clark, & Low, 2013; Fazel et al., 2008). In fact, an estimated 12%-17%
of adolescents experience hallucinations and about 40%-45% experience disordered
thinking (Sedlak & McPherson, 2010). Such symptomology is strongly related to the
development of SMI, and the studies that have examined mental illness among detained
youth estimate that 3.3% of male DAs and 2.9% of female DAs meet criteria for
psychotic disorders or schizophrenia (Fazel et al., 2008; Teplin et al., 2012).
Problems with trauma are also quite prevalent among adolescents held in juvenile
justice facilities (Sedlak, & McPherson, 2010; Teplin et al., 2002). Although the severity
of trauma history varies, the majority of DAs endorse experiencing at least one traumatic
event, such as neglect, malnourishment, physical abuse, sexual abuse, assault, or
witnessing extreme violence (Sedlak, & McPherson, 2010; Steiner, Garcia, & Matthews,
1997). According to the most recent Survey of Youth in Residential Placement (SYRP),
which sampled 7,073 adolescents held in juvenile justice facilities, 23% of adolescents in
detention and 25% of adolescents in prison endorsed experiencing three or more
traumatic events (Survey of Youth in Residential Placement, 2003). Not surprisingly,
epidemiological studies estimate that as many as one in three DAs have Posttraumatic
Stress Disorder (Steiner et al., 1997; Teplin et al., 2002). Moreover, about one in five

5
DAs endorse frequent, intense suicidal ideation and one in ten DAs report a history of
suicidal behaviors and/or attempts (Hussey, Drinkard, & Flannery, 2007; Shufelt &
Cocozza, 2006).

1.3

Substance-Related Concerns of Detained Adolescents

Over the past two decades, the number of adolescents using drugs and/or alcohol
has steadily increased (Jones, Heflinger, & Saunders, 2007) and this has been particularly
true among juveniles involved in the juvenile justice system (Dembo, et al., 1997;
Soenksen, Stein, Brown, Stengel, Rossi, & Lebeau, 2015). Approximately 70%-90% of
DAs admit to regular use of alcohol and drugs, compared to only about 35% of nondetained adolescents (Bostwick, & Ashley, 2009; Sedlak, 2009). When interviewed
about substance use, DAs report that drugs and alcohol are readily available and easy to
access within their communities, often resulting in early experimentation with substances
prior to the age 10, as well as regular use of multiple substances (Bath et al., 2013;
Dembo et al., 1997). Many DAs acknowledge that substance use leads to problems at
home, school, work, the legal system, and with friends, but continue to use substances to
gain or maintain peer acceptance, deal with problems, improve their mood, have fun,
and/or waste time (Sedlak, 2009; Survey of Youth in Residential Placement, 2003). At
least half of all DAs meet formal criteria for a substance-related disorder (Aarons,
Brown, Hough. 2004; Hussey et al., 2007) and about 25% of DAs have two or more
substance-related disorders (Young, Dembo, & Henderson, 2007). Alcohol, tobacco, and
marijuana represent the most commonly used substances among DAs, although about
20%-25% of DAs engage in regular experimentation with other drugs like cocaine,
ecstasy, crystal methamphetamine, inhalants, heroin, and/or prescription medications
(Jones et al., 2007; Sedlak & McPherson, 2010). Furthermore, as many as 65% of DAs
are comorbid, meaning they meet criteria for both mental disorders and substance-related
disorders (Hussey et al., 2007; Johnson et al., 2004).
Evidence suggests that some demographic variables are related to mental health
and substance-related problems within the detained adolescent population. Specifically,
detained females appear to be at greater risk for mental health problems than their male

6
counterparts. A meta-analysis based on 70,423 adolescents from 283 juvenile justice
facilities found that females were 1.8 to 2.4 times more likely to have clinically
significant elevations on various mental health symptom scales, as measured by a mental
health screener (Vincent et al., 2008). Additionally, studies indicate that, except for
psychotic disorders, prevalence rates for mental disorders are higher among detained
females than detained males (Abram, et al., 2003; Hussey, Drinkard, Falletta, &
Flannery, 2008). Not surprisingly, detained females regularly report greater severity and
frequency of mental illness symptoms, lower overall functioning, and poorer mental
health status (Bostwick, & Ashley, 2009; Ryan, Williams, & Courtney, 2013).
Several reasons to explain gender discrepancies among DAs have been proposed,
mostly related to the differential treatment of males and females within the juvenile
justice system. For example, females are less likely to be arrested than males (Bostwick,
& Ashley, 2009); about 3 in 100 adolescent females were arrested in 2010, compared to
about 7-8 in 100 adolescent males (Office of Juvenile Justice and Delinquency
Prevention, 2013). In addition, judges are less likely to place females in correctional
facilities and more likely to assign them to probation or other diversion programs
(Bostwick, & Ashley, 2009; Vincent et al., 2008). Thus, detained females tend to be the
most problematic and deviant females involved in the juvenile justice system, with severe
mental health problems (Hussey et al., 2008; Veysey, 2003), whereas detained males are
not necessarily the most deviant males involved in the juvenile justice system.
Evidence is mixed regarding racial and ethnic differences in overall rates of
mental health problems among detained adolescents. Some studies suggest White DAs
have significantly higher mental health needs (Lopez-Williams, Vander Stoep, Kuo, &
Stewart, 2006) and are more likely than Black DAs or Hispanic DAs to meet criteria for
one or more psychiatric disorders (Abram, et al., 2003; Teplin et al., 2012). In contrast,
other studies have found that minority DAs have more significant mental health problems
than White DAs (Lo, Howell, & Cheng, 2003; Rawal et al., 2004), or failed to find
significant racial/ethnic differences in the proportion of DAs with mental disorders
(Dalton, Evans, Cruise, Feinstein, & Kendrick, 2009). Despite such contradictory
findings, there is some evidence of race/ethnic differences for specific diagnoses

7
(Pumariega, Atkins, Rogers, Montgomery, Nybro, Caesar, & Millus, 1999). Black and
Hispanic adolescents are more likely to be diagnosed with affective disorders, anxiety
disorders, and/or psychotic disorders, whereas White adolescents are more likely to be
diagnosed with attention deficit-hyperactivity disorder, conduct disorders, and/or
substance-related disorders (Getahun, Jacobsen, Fassett, Chen, Demissie, & Rhoads,
2013; Teplin et al., 2012; Vincent et al., 2008). In one of the most comprehensive studies
of mental health concerns among the DA population, Vincent and colleagues (2008)
examined the MAYSI-2 results for over 70,000 youth in 283 facilities. Results showed
that White DAs were more likely than minority DAs to endorse concerns related to
alcohol/drug use, suicidal ideation, and somatic complaints, Black DAs were more likely
to endorse psychosis-related symptoms, and no differences were found for endorsement
of mood and anxiety symptoms. Based on such findings, Vincent and colleagues (2008)
concluded that the extant literature is currently too limited to draw firm conclusions about
possible differences in mental health problems across race/ethnicity groups within the
detained adolescent population.

1.4

Why Care about these Concerns among Detained Adolescents?

As noted above, strong empirical evidence suggests that DAs suffer
disproportionately high rates of mental health concerns, mental disorders, and substancerelated disorders. Such mental health problems are concerning because they are
associated with recidivism (Cottle et al., 2001), or the repetition of criminal activity that
typically involves re-offending and/or getting re-arrested with a new charge (Myner,
Santman, Cappelletty, & Perlmutter, 1998). Recidivism is common among DAs, with an
estimated 40%-70% of adolescents getting re-arrested following release from detention or
prison (Gordon, Diehl, & Anderson, 2012; Grunwald, Lockwood, Harris, & Mennis,
2010). As a result, many of these juveniles return to correctional facilities and become
stuck in a “revolving door” within the juvenile justice system, in which they are
repeatedly arrested, detained, released, re-arrested, and re-detained (Harrison, 2001;
Maschi et al., 2008). Unfortunately, the end result of this revolving door is often longterm incarceration in the adult prison population (Hussey et al., 2008). Not surprisingly,

8
longitudinal studies show that adolescents with the highest risk of incarceration in adult
prison are those with a history of incarceration as a juvenile (Myner et al., 1998).
Unfortunately, mental health problems increase the risk of recidivism and
contribute to the revolving door problem and multiple incarceration stays (Calley, 2012;
Grunwald et al., 2010). According to one meta-analysis of 23 studies and 15,265
adolescents, affective disorders (e.g., anxiety, depression) represent one of the strongest
predictors of juvenile recidivism (Cottle, et al., 2001). Substance abuse, trauma, neglect,
conduct problems, ADHD, and untreated mental health symptoms can also greatly
increase the risk of recidivism (Gordon et al., 2012; Ryan, et al., 2013). Interestingly, a
recent longitudinal study found that the relationship between mental health problems and
recidivism differed by race/ethnicity, with a strong association between conduct problems
and property-related recidivism among White detained youth, compared to an association
between emotional problems and recidivism among Moroccan detained youth and an
associated between hyperactivity and recidivism among Surinamese detained youth
(Colins, Boonmann, Veenstra, van Domburgh, Buffing, Doreleijers, & Vermeier, 2013).
Colins and colleagues (2011) also discovered that drug abuse disorders, either alone or
comorbid with mental disorders, were linked directly to increased probability of
recidivism. Clearly, DAs with mental health and substance-related concerns face an
elevated risk of recidivism. Thus, addressing these concerns is a critical factor in
reducing the elevated risk for recidivism and preventing continued involvement in the
revolving door of the juvenile justice system.
1.5

Treatment Services for Detained Adolescents

Given the strong relationship between mental health, recidivism, and
incarceration, mental health services represent a promising avenue for addressing both
mental health concerns and the revolving door problem (Lipsey, 2011; Lopez-Williams,
et al., 2006). Blueprints for Healthy Youth Development, a research center designed to
identify, test, and disseminate evidence-based interventions for high-risk youth, reviewed
over 600 delinquency, drug, and violence prevention and intervention programs
(Henggeler & Sheidow, 2012). The researchers involved in the review identified

9
multisystemic therapy, functional family therapy, and multidimensional treatment foster
care as three gold-standard mental health interventions with strong empirical support for
treating detained adolescents. Positive outcomes for these treatments include increased
self-esteem, reduced psychiatric symptoms, reduced substance use, improved family
functioning, decreased association with deviant peers, reduced number of re-arrests,
reduced severity of charges, and delayed time until re-arrest (Cuellar, McReynolds, &
Wasserman, 2006; Henggeler & Schoenwald, 2011). Key treatment components entail
behavioral and cognitive-behavioral strategies that aim to improve the function of the
adolescent and the adolescent’s family. Such treatments are typically community-based
and maintain high fidelity through quality assurance checks (Henggeler & Schoenwald,
2011; Lipsey, 2011).
After conducting a large meta-analysis of current treatments for detained youth,
Lipsey (2009) concluded that treatment services focused on behavioral and interpersonal
skills tend to be highly successful for promoting reductions in violence and recidivism.
Specifically, family functional therapy (FFT), community residential programs, and
multisystem therapy (MST) services showed significant impacts on recidivism (Darnell
& Schuler, 2015; Gordon et al., 2012). For example, the first community-based
experimental study of FFT produced real-world reductions in recidivism of 50% at 1-year
follow-up among adolescents who successfully completed the entire FFT intervention
(Sexton & Turner, 2010). However, such reductions in recidivism are not a universal
finding and poorly implemented interventions have failed to demonstrate significant,
long-term impacts on repeated criminal activity (Lipsey, 2011; Schwalbe, Gearing,
MacKenzie, Brewer, & Ibrahim, 2012).
Although there is strong empirical support for the efficacy of specific mental
health interventions, the evidence to support substance use treatments is limited (Desai et
al., 2006). Based upon findings of a meta-analysis of current alcohol/drug treatments for
detained youth, Williams and Chang (2000) were unable to identify any specific
intervention with strong empirical evidence for producing long-term reductions in
substance use among DAs. At best, these researchers were able to conclude that some
treatment services appear to be better than no treatment services (Desai et al., 2006;

10
Williams & Chang, 2000). However, researchers have identified several key components
for substance-related interventions that appear to be related to positive outcomes;
components include substance use screening, thorough assessment of substance use,
motivation techniques, cognitive-behavioral and/or family-based treatment techniques,
graduated sanctions, and completion incentives (Bath et al., 2013). Treatment outcomes
among DAs also tend to be better when substance-related interventions have high
treatment fidelity, long duration (at least 6 weeks), focus on development of interpersonal
skills, provide family psychoeducation, and include active family participation (Bath et
al., 2013).

1.6

Prevalence of Treatment Use among Detained Adolescents

Despite the existence of several empirically-supported mental health interventions
(e.g., FFT, MST) that promote good outcomes for the detained adolescent population
(Darnell & Schuler, 2015; Henggeler & Schoenwald, 2011; Schwalbe et al., 2012), the
majority of DAs do not receive treatment services. Prevalence estimates for mental
health treatment utilization among DAs vary considerably across studies, ranging from as
low as 3% (Winter 1991) to as high as 76% (Lyons, Griffen, Quintenz, Jenuwine, Shasha,
2003). Clearly, there is a lack of clear consensus on service utilization rates for DAs
within the literature. To better understand service utilization, I conducted a meta-analysis
of treatment utilization rates among DAs, based on 32 studies with 34 distinct samples
totaling 21,039 adolescents (White, Aalsma, Salyers, & McGrew, under review). Results
indicated that approximately 37.8% of DAs obtained at least one mental health and/or
substance-related treatment service at some point in time, with 29.3% of DAs obtaining
mental health services (excluding substance-related services) and 26.0% of DAs
obtaining substance-related services (excluding mental health services). Furthermore,
less than 60% DAs with mental health needs (e.g., diagnosed mental disorders) received
any type of mental health and/or substance-related treatment. Findings from the metaanalysis suggest that about one in three DAs obtain treatment services, whereas two in
three DAs have prominent mental health and/or substance-related problems that likely
require treatment services.

11
Among the small percentage of DAs who utilize treatment services, the most
common services appear to be outpatient treatment in the community (Lyons et al.,
2003). Based upon the studies included my meta-analysis referenced above, an estimated
20%-40% of DAs have attended at least one mental health outpatient session (Hussey et
al., 2007; Hussey et al., 2008; Rawal et al., 2004) and about 15%-35% have attended at
least one substance-related outpatient session at some point in time (Hussey et al., 2007;
Paskar, 2008). Interestingly, when restricting services to only court-ordered outpatient
treatment, rates for mental health outpatient treatment among DAs decrease to
approximately 8%-10% (Aalsma, Tong, Lane, Katz, & Rosenman, 2012b; Novins,
Duclos, Martin, Jewett, & Manson, 1999), whereas rates for substance-related outpatient
treatment increase to approximately 38%-40% (Novins et al., 1999). In addition to
traditional outpatient treatment, an estimated 18%-28% of DAs with substance-related
disorders participate in community-based non-professional programs, such as Alcoholics
Anonymous or Narcotics Anonymous (Johnson et al., 2004), and as many as 15%-25% of
juvenile justice-involved youth seek religious counseling to address mental health and/or
substance-related concerns (Novins et al., 1999).
Similar to the general adolescent population, prevalence rates for non-outpatient
services among detained youth are notably low (Hoeve et al., 2013). The few studies to
examine non-outpatient treatment among juvenile justice-involved adolescents indicate
about 10%-25% of DAs utilize inpatient services (Lopez-Williams et al., 2006;
McPherson, 1991; Robertson et al., 2004) and 2%-16% reside in long-term residential
treatment facilities for mental health concerns (Lyons, Baerger, Quicgley, Erlich, &
Griffith, 2001; Novins et al., 1999); about 3%-12% utilize inpatient services (Lyons et al.,
2001; Paskar, 2008) and 2%-10% reside in long-term residential treatment facilities for
substance-related problems (Novins et al., 1999; Rawal et al., 2004). Since most studies
have focused on the prevalence of treatment utilization among DAs (Bullis, Yovanof, &
Havel, 2004; Lopez-Williams et al., 2006; Rogers, Pumariega, Atkins, & Cuffe, 2006;
Shelton 2005), rather than the quantity or frequency of treatment utilization, conclusions
about mean number of inpatient stays, average number of outpatient sessions, frequency
between services, or gaps in treatment services cannot be made.

12
Despite limited information about the quantity and frequency of service use by
DAs, researchers have examined the cost of treating adolescents with a history of
detention placements and mental health diagnoses. Specifically, Hussey and colleagues
(2008) estimated that the average cost of treating a detained youth with high-quality,
evidence-based treatment is approximately $10,274, with treatment services including
individual therapy, group counseling, family counseling, case management, in-home
visits, and housing (if needed) (Hussey et al., 2008). Given the high costs associated with
recidivism (e.g., multiple court appearances, multiple detention/prison stay) that place a
large financial burden on the juvenile justice system, researchers have concluded that
providing evidence-based services to DAs represents an efficient and cost-effective
option for managing DAs (Braverman & Murray, 2011; Hussey et al., 2008). However,
few DAs obtain high-quality, wrap-around treatment services that are empirically
supported. Rather, the treatment services available for DAs are often low quality and/or
ineffective (Aarons et al., 2004; Teplin et al., 2005); only 5.0% of eligible DAs
participate in evidence-based interventions each year (Henggeler & Schoenwald, 2011).
This markedly low prevalence estimate reflects a missed opportunity, since treatment can
potentially improve symptomology, enhance individual and family functioning, and
reduce recidivism (Lipsey, 2009; Lyons et al., 2003; Sexton & Turner, 2010).

1.7

Factors Related to Treatment Use among Detained Adolescents

Altogether, the low rates of treatment utilization, in conjunction with low-quality
services, indicate that DAs represent an underserved population (Curtin, 2002; Johnson et
al., 2004; Teplin et al., 2005). Several key factors and treatment barriers at the systemlevel and individual-level have been proposed as possible reasons to explain the poor
engagement in treatment among DAs (Abram, Paskar, Washburn, & Teplin, 2008; Kates,
Gerber, & Casey, 2014; Lopez-Williams et al., 2006; Rogers et al., 2001). With regards
to system-level factors, limited availability of treatment services within facilities, as well
as failure to connect DAs to treatment upon community reentry, represent serious
treatment barriers. Currently, all juvenile justice facilities are legally mandated to
provide behavioral healthcare to adolescents residing within facilities (Braverman &

13
Murray, 2011). As recommended by the American Academy of Pediatrics, the National
Commission on Correctional Health Care (NCCHC) has established standards for mental
health and substance-related treatment for detained youth (Gallagher & Dobrin, 2007).
However, facilities are not required to obtain accreditation with the NCCHC, so there is
no consistent mechanism for monitoring services, keeping facilities accountable, or
providing funding to facilities to maintain high standards of care (Pajer, Kelleher, Gupta,
Rolls, & Gardner, 2007; Yazzie 2011). As a result, most juvenile justice facilities lack
the budget and/or resources to provide psychiatric treatment and tend to offer the bare
minimum of care (Desai et al., 2006; Mulvey, Schubert, & Chung, 2007; Riley, 2014).
For example, NCCHC guidelines recommend that facilities administer mental health
screeners within 14 days of intake, provide 24-hour emergency mental health services,
and provide DAs the right to request mental health treatment services daily (Braverman
& Murray, 2011), but less than half of facilities are compliant with these standards
(Braverman & Murray, 2011). As of 2007, only 53 of approximately 3500 juvenile
justice facilities were fully accredited with the NCCHC (Gallagher & Dobrin, 2007).
Even when juvenile justice agencies attempt to comply with NCCHC standards,
the treatment services can vary greatly across facilities. Only a small percentage (25%30%) of facilities have established plans for 24-hour mental healthcare and only 40%55% of facilities offer mental health counseling (Pajer et al., 2007; Young et al., 2007).
Furthermore, 30%-40% of facilities for detained youth do not use mental health and/or
substance-related screeners (Young et al., 2007) and one-third of the facilities that use
screeners rely on correctional staff to administer these assessments, despite no training in
behavioral health (Desai et al., 2006; Sedlak, 2009). Much like the low prevalence of
evidence-based treatment, failure to implement and/or appropriately use mental health
screeners within the juvenile justice system represents a missed opportunity, since
screeners help detention staff identify DAs with mental health and/or substance-related
needs, determine appropriate services during (and after) detention, and improve
communication between staff and DAs (Rogers et al., 2001; Williams, Grisso, Valentine,
& Remsburg, 2008). It should be noted, however, staff within the juvenile justice system
sometimes misinterpret findings of screeners or fail to identify DAs with mental health

14
concerns, perhaps due to limited resources, lack of training, and/or high caseloads
(Wasserman, McReynolds, Whited, Keating, Musabegovic, & Huo, 2008). As a result,
DAs may be overlooked and not connected to appropriate mental health and substancerelated services within juvenile justice facilities and/or in the community (Hoeve et al.,
2013; Towberman, 1994). Thus, while screeners serve as vital tools for initial
assessment, screening for mental health concerns is not sufficient on its own to bring
about positive treatment outcomes and successful rehabilitation for DAs (Cauffman,
2004). Mental health screeners should never serve as replacements for treatment
services, nor viewed as signs of treatment engagement or treatment success (Rogers et al.,
2001; Williams et al., 2008).
In addition to limited services available in juvenile justice facilities, other systemlevel factors that impacts treatment utilization among DAs include staff behaviors,
attitudes, and knowledge (Holloway, Brown, Suman, & Aalsma, 2013). While
adolescents are being detained, detention staff are largely responsible for determining
who receives services and the types of services being offered. Unfortunately, evidence
suggests that negative attitudes among providers can result in biased and differential
treatment based on adolescents’ gender, race, or other characteristics (Lopez-Williams et
al., 2006; Towberman, 1994). Once DAs are released from detention, juvenile probation
officers play an integral role in helping DAs obtain appropriate mental health and
substance-related services in the community (Holloway et al., 2013; Riley, 2014). This is
because many DAs do not know who, where, or how to receive services, so probation
officers serve as a source of information or “gateway providers” for assisting detained
youth upon community reentry (Stiffman, Pescosolido, & Cabassa, 2004). Not
surprisingly, juvenile probation officers’ knowledge and competency to identify mental
health concerns, familiarity with available resources, and personal attitudes towards
rehabilitation versus punishment have been directly linked to treatment service
engagement among the DAs on their caseloads (Howell, 2003; Stiffman et al., 2004;
Wasserman et al., 2008).
Another system-level barrier related to treatment utilization among detained
youth involves insurance coverage. Approximately 15%-20% of DAs are uninsured and

15
cannot afford to utilize mental health services in the community (Gupta et al., 2005;
Maschi et al., 2008), leaving the juvenile justice system as one of the only avenues to
access treatment services (Pumariega et al., 1999; Rogers et al., 2006). This is
problematic, given that services within detention facilities are not always available and/or
high quality. For DAs with insurance, many receive services through Medicaid programs
(Braverman & Murray, 2011). While such programs help adolescents in the community,
about half the states in the United States suspend or terminate Medicaid coverage upon
placement in a correctional facility (Cuellar, Kelleher Rolls, & Pajer, 2005; Gupta et al.,
2005). Termination of coverage not only hinders adolescents already receiving services
in detention, but also prevents service use during the critical period immediately after
detention, due to a wait time of approximately 45-90 days for re-enrollment into
Medicaid (Liddle, Dakof, Henderson, & Rowe, 2011). The delay in access to community
treatment services post-detention disrupts continuity of care and has been shown to
negatively impact the treatment progress that adolescents make during incarceration
(Gupta et al., 2005; Liddle et al., 2011). In addition, confusion associated with the
application process for re-enrolling in Medicaid prevents many adolescents from
successfully re-applying for Medicaid coverage post-detention (Gupta et al., 2005). Thus,
except for the small percentage of adolescents (15%-20%) with private insurance, most
DAs with mental health needs face significant system-level barriers that interfere with
obtaining effective mental health services both during and immediately following
detention (Shelton, 2005).
With regard to individual-level factors associated with service use among DAs,
gender and race/ethnicity have been shown to impact treatment utilization (Kataoka et al.,
2001; Lopez-Williams et al., 2006). Regarding gender, rates of treatment-seeking,
treatment referral and service utilization tend to be significantly higher among
incarcerated females than males (Abram, et al., 2008; Veysey, 2003). In addition,
detained females are more likely to participate in several types of services (e.g.,
individual, group, family) and remain in treatment longer than their male counterparts
(Hussey et al., 2008; Pumariega et al., 1999). Regarding race/ethnicity, DAs from
racial/ethnic minority groups tend to be significantly less likely than White DAs to be

16
referred for treatment, placed in mental health treatment facilities, and utilize services
(Dalton et al., 2009; Novins et al., 1999; Samuel, 2015). To illustrate the strong
association between race/ethnicity and treatment utilization, one study of detained youth
found that race was the only significant predictor of mental health treatment, with White
adolescents being more likely to receive treatment than Black adolescents (Shelton,
2005).
The disparity in treatment access across racial groups is particularly problematic
when considering that minority adolescents are much more likely to be arrested and
incarcerated than White adolescents (Towberman, 1994). Such bias results in
overrepresentation of Black adolescents throughout the juvenile justice system; Black
adolescents account for only 15% of the adolescent population nationwide (Sedlak &
Bruce, 2010), but represent 26% of arrested youth, 31% of referrals to juvenile court,
40% of youth placed in residential facilities, and 44% of detained adolescents (PoeYamaga & Jones, 2007; Rawal et al., 2004). As addition evidence of bias, research
shows that Black adolescents with mental illness are 6 times more likely to be placed in
detention/prison than similarly-aged White adolescents with mental illness (Isaacs, 1992),
whereas White adolescents are 4 times more likely to receive a mental health treatment
placement (Herz, 2001). When considering gender and race together, White females
DAs are most likely to obtain mental health services, whereas Black male DAs are least
likely to obtain treatment (Abram et al., 2008). In fact, White females are approximately
8 times more likely than Black males, and 2.5 times more likely than Black females and
White males, to receive mental health treatment post-detention (Herz, 2001).
Besides gender and race/ethnicity, age is also related to service utilization within
the detained adolescent population. There appears to be a negative relationship, in that
the likelihood of service utilization post-detention decreases as age increases (Aalsma et
al., 2012b; Lopez-Williams, 2006). Thus, younger DAs are more likely to obtain mental
health referrals, receive mental health placements, be connected to services, and utilize a
variety of service types (e.g., inpatient, outpatient, family) than older DAs of similar
gender or racial/ethnic background (Herz, 2001; Vander Stoep, Evens, & Taub, 1997).
Current findings suggest a possible two-tiered approach within the juvenile justice

17
system, in which younger offenders are placed on a treatment-focused rehabilitation track
and older offenders are placed on a punishment-focused incarceration track (Herz, 2001).
Another individual-level factor associated with service utilization among DAs
includes criminal history. Specifically, adolescents who commit more serious and/or
violent crimes (e.g., battery, use of weapon, sexual assault) are more likely to obtain
referrals and to utilize treatment, compared to adolescents who commit less serious
crimes (e.g., trespassing, possession of marijuana) (Paskar, 2008; Shelton, 2005; Vander
Stoep et al., 1997). It has been suggested that serious offenders have higher rates of
treatment services because these adolescents tend to receive heightened attention, which
may include increased efforts to achieve rehabilitation through provision of treatment
services (Kataoka et al., 2001). Similarly, DAs with repeat offenses (i.e., recidivists) and
repeat detention stays (i.e., re-detained/incarcerated) are more likely to use mental health
and substance-related treatment services than first time offenders (Kataoka et al., 2001;
Lopez-Williams, et al., 2006; Vander Steoep et al., 1997), perhaps due to more
opportunities during multiple contacts with the juvenile justice system to be correctly
identified as having behavioral health concerns, placed in treatment services, and/or
referred for community services.
Some studies indicate that DAs with mental health and/or substance-related needs
are more likely to be referred and connected to treatment services (Brannan & Heflinger,
2005; Hoeve et al., 2013), resulting in higher rates of service utilization among DAs with
psychiatric disorders. In contrast, other studies have failed to show a strong relationship
between mental health concerns and treatment utilization (Lopez-Williams et al., 2006;
Novins et al., 1999). In fact, evidence suggests that demographic factors and insurance
status serve as stronger predictors of treatment utilization than actual mental health needs
or diagnoses. However, conclusions about service use for DAs with mental health and
substance-related needs are difficult to make because DAs are not consistently screened
or identified as having needs (Cauffman, 2004; Herz, 2001). Moreover, the method in
which studies gather data about treatment utilization can impact conclusions, given that
studies using interviews with DAs tend to find higher rates of psychiatric disorders in
conjunction with higher rates of treatment utilization (Novins et al., 1999), compared to

18
studies using surveys of non-detained youth (e.g., parents, detention staff) or review of
records that find low rates of treatment utilization (Aalsma et al., 2015; Lyons et al.,
2001; Rawal et al., 2004).
Finally, history of treatment services may play a role in treatment utilization by
detained youth. Studies of help-seeking behaviors and treatment utilization (Cauce, &
Srebnik, 2003; Samuel, 2015) among different adolescent populations (e.g., those with
chronic medical illness, development disabilities, involvement in foster care), indicate
that engagement in mental health treatment is strongly tied to prior use of services (Kates
et al., 2014), particularly if the prior services were viewed as helpful, necessary, easy to
access, and affordable (Abram et al., 2008). Given such findings, it is likely that DAs
with a history of engagement in mental health and/or substance-related treatment prior to
detention are more likely to continue using services during detention and following
release from detention. The few studies to examine service utilization both pre- and postdetention among DAs support this possibility, with a significantly higher percentage of
DAs with a history of treatment versus DAs without treatment engaging in communitybased services upon release from detention (Hussey et al., 2007; Hussey et al., 2008;
Shelton, 2005).

1.8

Timing of Treatment Services and Continuity of Care

Clearly, multiple system-level and individual-level factors impact treatment
utilization among detained adolescents. The timing of treatment utilization is also
crucial. While pre-detention treatment appears to be quite important, post-detention
treatment may represent the most important time-point for helping DAs manage mental
health concerns, as well as successfully reintegrate into the community and avoid
recidivism (Riley, 2014; Samuel, 2015). Unfortunately, a limited number of studies have
explored service use among DAs upon reentry into the community, with prevalence
estimates ranging from about 6% (Teplin et al., 2005) to about 40% (Trupin, Turner,
Stewart, & Wood, 2004). In my meta-analysis, prevalence rates for treatment utilization
were 27.7% for any treatment utilization, 34.3% for mental health treatment utilization,
and 29.0% for substance-related treatment utilization, with decreasing service use as time

19
in the community increased. In addition to poor utilization of community-based services,
longitudinal studies also show large gaps in the receipt of treatment services (Hussey et
al., 2008), as well as high rates of recidivism and re-incarceration (Paskar, 2008; Ryan et
al., 2013). Based on such findings, it appears that DAs may be failing to obtain services,
terminating early from services, and/or experiencing large gaps in treatment services over
time (Bullis, Yovanoff, Mueller, & Havel, 2002). DAs are not receiving continuity of
care or consistent receipt of services across settings and over time, which contradicts the
established guidelines for evidence-based interventions like MST and FFT (Henggeler, &
Sheidow, 2012; Lipsey, 2009; Lipsey, 2011). Moreover, discontinuity of care and
treatment dropout have been linked to numerous negative outcomes for adolescents and
the mental healthcare system, including increased mental health impairment, more severe
criminal activity, and increased financial burden on the healthcare system (Corning &
Malofeeva, 2004; Luk, Staiger, Mathai, Wong, Birleson, & Adle, 2001; Oruche, Downs,
Holloway, Draucker, & Aalsma, 2014).
Given the negative outcomes linked to treatment dropout and discontinuity, it is
imperative that researchers examine these issues among the DA population. Since few
studies have examined service use over time, little is known about the factors related to
treatment utilization and continuity once DAs return to the community. Similar to other
adolescent populations, early termination of treatment and/or gaps in treatment services
may occur due to a variety of reasons (Abram et al., 2008; Hoeve et al., 2013). As
mentioned previously, DAs may experience system-level barriers related to lack of
insurance coverage or provider bias (Cuellar et al., 2005; Gupta et al., 2005); other
potential factors may include transportation difficulties, lack of knowledge about
how/where to obtain services, lack of available providers, moving outside of provider
network, or placement in a long-term correctional facility (Abram et al., 2008; Chapman
et al., 2006; Kazdin & Mazurick, 1994). Alternatively, disruptions in treatment services
may stem from DAs intentionally dropping out or avoiding services due to negative
attitudes about treatment, distrust in mental health providers, inaccurate beliefs that one
does not need treatment, and/or embarrassment of obtaining treatment (Abram et al.,
2008; Moore, McArthur, & Saunders, 2013; Samuel, 2015). Interestingly, when

20
interviewed about mental health treatment, DAs endorsed prominent mental health
stigma and inaccurate beliefs that their mental health and substance-related problems
would simply “go away” without the use of treatment services (Abram et al., 2008).
Given the high rate of recidivism and re-incarceration within the DA population,
poor continuity of treatment services may be partially explained by DAs’ frequent
placements in detention centers or prison facilities. No study (that I have found) has
specifically examined whether DAs intentionally drop out of community-based services
or are forced to terminate services in response to placement in detention/prison or other
treatment barriers. While it is likely a combination of several factors, the important point
is that many DAs are not receiving mental health or substance-related services postdetention and are failing to benefit from the positive impacts of community-based mental
health interventions (Hoeve et al., 2013; Lyons et al., 2003).

1.9

Limitations of the Literature

Despite growing recognition that detained youth has significant mental health
needs and disproportionately high rates of psychiatric disorders (Shufelt & Cocozza,
2006; Yazzie, 2011), there are serious gaps in the literature that prevent full
understanding of how these needs are being met with treatment services. One of the
biggest limitations of the current literature is the lack of data about treatment utilization
beyond prevalence rates. Currently, most studies only report prevalence estimates (i.e.,
whether DAs ever obtained services), with little data about quantity, frequency, or
intensity of treatment services. Thus, no thorough examination (that I have found) has
studied the specific characteristics of service use, such as timing of services (e.g., pre-,
during, post-detention), number of services (e.g., mean, range, length of stay), and
frequency between services (e.g., days between sessions) (Kataoka et al., 2001; Shelton,
2005). Without information about quantity of services, it is difficult to determine
whether DAs are obtaining treatment that satisfies the established guidelines for
evidence-based interventions like MST or FFT (Henggeler & Sheidow, 2012). In
addition, it is possible that one group of DAs may have higher prevalence rates for
treatment utilization (e.g., White DAs compared to Black DAs), but may actually utilize

21
less number of services. Such findings could lead to new conclusions about the
intersection between individual-level factors (e.g., gender, race) and treatment utilization,
but such conclusions cannot be made given the current state of the literature.
With regards to intensity of service use, evidence suggests that a small percentage
(25%-35%) of DAs qualify as intensive service users who are receiving most of the
mental health services provided to the population (Cauffman, Scholle, Mulvey, &
Kelleher, 2005; Pumariega et al., 1999). Research within the psychiatric rehabilitation
literature suggests that high service users negatively impact the quality and effectiveness
of the mental healthcare delivery system by using a disproportionate amount of resources,
services, provider time, etc., thereby resulting in inferior treatment for the rest of the
population. The DA population may be experiencing similar issues, but more research is
needed to quantify the existence of intense service users and their use of services in
relation to the entire detained adolescent population.
The dearth of longitudinal research represents another serious limitation with the
current literature. Research focused on mental health treatment utilization among DAs
has been predominately based on cross-sectional studies, which yield snap-shots of
service use at one point in time, but no information about patterns of service use over
time. Such estimates do not capture the scope and/or magnitude in which DAs are likely
being overlooked and underserved, particularly regarding whether their needs are being
met at different time points within the juvenile justice system (e.g., during detention,
within 1 year post-detention). Further, the majority of the cross-sectional studies have
examined mental health and/or substance-related services prior to detention, rather than
treatment utilization upon community reentry. The few studies to examine treatment
services after release from detention have relied on relatively short-term follow-up
periods of 1-2 years. To my knowledge, only a few studies (Paskar, 2008; Teplin et al.
2012) have tracked detained youth for three or more years. Additionally, longitudinal
studies have generally examined community-based outpatient services (Aalsma et al.,
2012b; Bullis et al., 2002; Trupin et al., 2004), with limited exploration of alternative,
more intensive treatment services like visits to emergency departments of medical centers
or long-term stays within inpatient facilities.

22
Finally, perhaps due to the limited number of longitudinal studies, the literature
has largely overlooked crucial issues related to treatment dropout and continuity of care
over time (Abram et al., 2008; Gupta et al., 2005). As a result, there is a poor
understanding of the time-point in which DAs typically drop out of treatment, the average
number of sessions preceding treatment dropout, potential reasons for dropping out of
treatment, the prevalence of gaps between services, potential reasons for treatment gaps,
and the characteristics of adolescents who tend to remain in treatment versus drop out
from treatment. Examination of such issues is crucial for identifying system-level
problems, as well as developing policies and programs to retain DAs in treatment,
prevent treatment gaps, and ensure that positive treatment outcomes like improved mental
health symptoms, reduced recidivism, and successful reintegration occur.

1.10 Purpose of Study
Despite limitations in the literature, studies consistently indicate that rates of
mental health and substance-related treatment utilization among the detained adolescent
population are quite low compared to the high rates of psychiatric disorders (Paskar,
2008; Shelton, 2005; Teplin et al., 2005). In part, these low rates are due to various
treatment barriers, such poor availability of services, provider attitudes/behaviors,
demographic factors, criminal history background, and availability of services. As a
result, there appears to be a significant disconnect between the identification of DAs with
mental health concerns, connection of DAs to treatment services, and engagement of DAs
in appropriate post-detention services within the community (Hoeve et al., 2013; Mulvey
et al., 2007; Rogers et al., 2001). Given the gaps in the literature, the scope of this
problem is not well understood. The literature offers mostly cross-sectional studies that
do not adequately examine the characteristics (e.g., prevalence, quantity, intensity) of
treatment services for DAs upon release from detention, as well as over time (Mulvey et
al., 2007). The few longitudinal studies that exist have not adequately measured
treatment utilization among DAs over long periods of time, preventing full understanding
of service utilization patterns, continuity of care patterns, and potential differences in
treatment services between groups of DAs (Herz, 2001; Mulvey et al., 2007).

23
In response to the gaps in the literature, I conducted this exploratory, longitudinal
study of mental health and substance-related treatment utilization among a large sample
of detained youth experiencing their first detention stay at some point during 1998 and
2011. The study tracked DAs for fourteen years, in order to better understand the
prevalence of post-detention treatment at different time points, the characteristics of
services use over time, the role of individual-level factors (e.g., demographics, criminal
history) in treatment utilization, the characteristics of treatment dropouts, continuity of
care patterns, and the complex relationships between treatment utilization and outcomes
like recidivism and re-incarceration (Aalsma et al., 2012b; Trupin et al., 2011).
Ultimately, conclusions from this study are intended to expand the literature and identify
gaps and disparities in treatment services for DAs, so that appropriate recommendations
can be made to ensure positive outcomes for the detained adolescent population.

1.11 Study Aims and Research Questions
The specific aims of the study were as follows: 1) Conduct a cross-sectional
exploration of post-detention treatment utilization (within two years of detention release),
2) Examine group differences in post-detention treatment utilization across demographic
groups, mental health groups, and criminal history groups, 3) Identify predictors of postdetention treatment utilization (within two years of detention release), and 4) Examine
longitudinal patterns of post-detention treatment utilization (e.g., time to first service,
time to dropout) among detained adolescents. Given the limited number of longitudinal
studies that currently provide data related to treatment utilization within the DA
population, specific hypotheses were not developed or tested; rather, this research was
exploratory in nature. In lieu of hypotheses, a series of research questions associated
with study aims were proposed and subsequently quantified, assessed, and answered to
achieve study aims. The research questions associated with study aims were as follows:

24
1.11.1 Research Questions for Aim 1
1. Prevalence of treatment utilization. What are the cross-sectional prevalence rates
for post-detention treatment utilization at different points in time among detained
adolescents?
a. Any treatment services. What is the prevalence of post-detention
treatment utilization?
(1)

Treatment services were subdivided and analyzed for the following
three treatment types: mental health services, substance-related
services, and either/both services.

(2)

Treatment services were analyzed within the following postdetention follow-up time frames: 2-years, 18-months, 1-year, 6months, and 1-month.

b. Outpatient services. What is the prevalence of post-detention outpatient
treatment utilization?
(1)

Treatment services were subdivided and analyzed for the following
three treatment types: mental health services, substance-related
services, and either/both services.

(2)

Treatment services were analyzed within the following postdetention time frames: 2-years, 18-months, 1-year, 6-months, and
1-month.

c. Non-outpatient services. What is the prevalence of post-detention nonoutpatient treatment utilization?
(1)

Treatment services were subdivided and analyzed for the following
three treatment types: mental health services, substance-related
services, and either/both services.

(2)

Treatment services were analyzed within the following postdetention time frames: 2-years, 18-months, 1-year, 6-months, and
1-month.

25
2. Characteristics of treatment services. What are the characteristics (e.g., amount,
frequency, length) of treatment services utilized among detained adolescents
within two years post-detention?
a. Quantity of treatment services. What is the total number (mean and range)
of treatment services utilized within two years post-detention?
(1)

Treatment services were subdivided and analyzed across the three
treatment types (i.e., mental health services, substance-related
services, and either/both services) and two treatment settings (i.e.,
outpatient and non-outpatient).

b. Frequency of outpatient treatment services. How frequently (in days) are
outpatient sessions utilized per month? How many days (mean and range)
elapses between outpatient sessions?
(1)

Treatment services were subdivided and analyzed for the following
three treatment types: mental health services, substance-related
services, and either/both services.

c. Length of stay of inpatient treatment services. What is mean length of
stay in days (number and range) for inpatient hospitalizations?
(1)

Treatment services were subdivided and analyzed for the following
three treatment types: mental health services, substance-related
services, and either/both services.

3. Intensity of treatment services. What is the intensity of treatment utilization
among detained adolescents within two years post-detention?
a. Service user type. For the three treatment types (i.e., mental health
services, substance-related services, and unknown/both services), what
proportion of DAs are: non-users, users of one service type, users of two
service types, or users of all three services types?
b. Service user setting. For the three treatment settings (i.e., outpatient,
inpatient, and emergency department), what proportion of DAs are: nonusers, users of one service setting, users of two service settings, or users of
all three services settings?

26
c. Outpatient user intensity. Based on Cauffman and colleagues (2005)
intensity of service categories for outpatient service utilization, what
proportion of DAs are: non-users, low users (1-2 sessions), low-tomoderate service users (4-7 sessions), moderate users (8-12 sessions),
moderate-to-high users (13-17 sessions), high users (18-22 sessions), or
extreme users (≥23 sessions)?
d. Non-outpatient user intensity. What proportion of DAs are: non-users,
low users (1 visit/stay), low-to-moderate users (2 visits/stays), moderate
users (3-4 visits/stays), high users (5-6 visits/stays), or extreme users (≥7
visits/stays)?
4. Treatment dropouts and treatment gaps. What are the prevalence rates for
outpatient treatment dropout (i.e., termination within 1-3 sessions) and treatment
gaps (i.e., 46-120 days between sessions) among detained adolescents within two
years post-detention?
a. Treatment dropouts. What is the prevalence of dropout from outpatient
treatment within two years post-detention among detained adolescents?
(1)

Treatment services were subdivided and analyzed for the following
three treatment types: mental health services, substance-related
services, and either/both services.

b. Reasons for dropouts. How many dropouts may be explained by the
following events occurring during the same month of dropout: placement
in detention or prison, stay in inpatient hospital, or no explanation (i.e.,
appears to be true dropout)?
c. Treatment gaps between services. What is the prevalence of gaps between
outpatient treatment services within two years post-detention among
detained adolescents?
(1)

Treatment services were subdivided and analyzed for the following
three treatment types: mental health services, substance-related
services, and either/both services.

27
d. Reasons for treatment gaps between services. For adolescents with ≥1
treatment gap(s), how many gaps may be explained by the following
events occurring during the same month of the gap: placement in detention
or prison, stay in inpatient hospital, or no explanation (i.e., appears to be
true gap)?

1.11.2 Research Question for Aim 2
1. Treatment services by groups. Does post-detention treatment utilization
significantly differ across groups of detained adolescents? Questions 1-4 of Aim
1 were re-run to examine whether prevalence rates, service characteristics,
intensity of services, dropout, and treatment gaps within two years post-detention
significantly differed across the following groups:
a. Demographic groups: Gender (male vs. female), race (Black vs. White),
and age (younger, mid-age, and older).
b. Mental health groups: Positive screen on Massachusetts Youth Screening
Instrument-2nd Edition (yes/no), disorder type (mental disorder, substancerelated disorder, and comorbid disorder), and pre-detention treatment
(yes/no).
c. Criminal history groups: Violent offender (yes/no), recidivist (yes/no), and
re-detained/incarcerated (yes/no).

1.11.3 Research Question for Aim 3
1. Predictors of service use. What are the significant predictors of treatment
utilization within two years post-detention among detained adolescents? The
following predictors and outcomes were examined:
a. Predictors: Demographic variables (gender, race, age), criminal history
variables (number of pre-detention arrests, length of detention stay, charge
severity, number of charges, violent offender [yes/no], Risk Assessment
Inventory score), and mental health variables (positive screen on MAYSI2 [yes/no], conduct-related disorder [yes/no], non-conduct mental disorder

28
[yes/no], substance-related disorder [yes/no], number of disorders, predetention outpatient treatment [yes/no], and pre-detention non-outpatient
treatment [yes/no]).
b. Outcomes: Any treatment, mental health treatment, substance-related
treatment, either/both treatment, outpatient treatment, and non-outpatient
treatment.
c. Sub-analyses: Cohort one (i.e., detained during 1998-2005) versus cohort
two (i.e., detained during 2006-2011).

1.11.4 Research Questions for Aim 4
1. Patterns of service utilization. What are the patterns of service utilization over
time among detained adolescents?
a. Time to first treatment utilization. What is the median length of time (in
days) to first treatment utilization post-detention?
b. Time to termination from outpatient treatment. What is the median length
of time (in days) involved in continuous outpatient treatment postdetention?
2. Group differences in treatment utilization. How does post-detention service
utilization patterns over time differ across DA groups? Question 1 of Aim 4 was
re-run to examine whether patterns significantly differed across the following
groups:
a. Demographic groups: Gender (male vs. female), race (Black vs. White),
and age (younger, mid-age, and older).
b. Mental health groups: Positive screen on Massachusetts Youth Screening
Instrument-2nd Edition (yes/no), disorder type (mental disorder, substancerelated disorder, and comorbid disorder), and pre-detention treatment
(yes/no).
c. Criminal history groups: Violent offender (yes/no), recidivist (yes/no), and
re-detained/incarcerated (yes/no).

29
3. Predictors of time to first treatment utilization. What are the significant predictors
of time to first post-detention treatment utilization? The following predictors and
outcomes were examined:
a. Predictors: Demographic variables (gender, race, age), criminal history
variables (number of pre-detention arrests, length of detention stay, charge
severity, number of charges, violent offender [yes/no], redetained/incarcerated [yes/no], Risk Assessment Inventory score), and
mental health variables (positive screen on MAYSI-2 [yes/no], conduct
disorder [yes/no], non-conduct mental disorder [yes/no], substance-related
disorder [yes/no], number of disorders, pre-detention outpatient treatment
[yes/no], and pre-detention non-outpatient treatment [yes/no]).
b. Outcomes: Time to first treatment, time to first mental health treatment,
time to first substance-related treatment, time to first either/both treatment,
time to first outpatient treatment, and time to first non-outpatient
treatment.
c. Sub-analyses: Cohort one (i.e., detained during 1998-2005) versus cohort
two (i.e., detained during 2006-2011).
4. Predictors of treatment termination. What are the significant predictors of time to
terminate from post-detention outpatient treatment? The following predictors and
outcomes were examined:
a. Predictors: Demographic variables (gender, race, age), criminal history
variables (number of pre-detention arrests, length of detention stay, charge
severity, number of charges, violent offender [yes/no], redetained/incarcerated [yes/no], Risk Assessment Inventory score), and
mental health variables (positive screen on MAYSI-2 [yes/no], conductrelated disorder [yes/no], non-conduct mental disorder [yes/no],
substance-related disorder [yes/no], number of disorders, pre-detention
outpatient treatment [yes/no], and pre-detention non-outpatient treatment
[yes/no]).

30
b. Outcomes: Time to termination from any outpatient treatment, time to
termination from mental health outpatient treatment, time to termination
from substance-related outpatient treatment, and time to termination from
either/both outpatient treatment.
c. Sub-analyses: Cohort one (i.e., detained during 1998-2005) versus cohort
two (i.e., detained during 2006-2011).

31

CHAPTER 2. METHOD

2.1 Data Abstraction and Data Linking
Data for this study were abstracted from two primary electronic databases: the
Marion County juvenile justice system (QUEST database) and the Regenstrief Medical
Record System. The Marion County Juvenile Justice system (QUEST) was chosen as a
primary data source because it contains the case files/records of all adolescents in contact
with the juvenile justice system, including youth who are arrested, detained, waived to
adult court, and/or committed to long-term prison. QUEST did not start electronically
storing data until 1998, so information prior to that date was not available for this study.
Altogether, records of all adolescents with a history of being arrested and/or detained at
least once between 1998 and 2011 were abstracted from QUEST.
The Regenstrief Medical Record System (RMRS) served as the second main data
source for the study. RMRS was specifically chosen because it serves as one of the
largest and oldest medical records systems in existence, with treatment records for over
1.5 million patients dating back to 1973 (McDonald et al., 1999). The RMRS and its next
generation successor, the Indiana Network for Patient Care (INPC), connects over 90% of
hospital-system care in the Indianapolis area, including Wishard Health Services since
1972, Indiana University Hospital and Riley Hospital for Children since 1989, Methodist
Hospital since 1995, Community Hospital since 1999, St. Vincent Hospital since 2000,
and St. Francis Hospital since 2002. Furthermore, all Indiana Medicaid data from 2001
have been successfully transferred and entered into INPC. Thus, the RMRS contains the
medical history data for most areas and sectors of the Indianapolis population and likely
contains the medical records for the majority of detained adolescents listed within the
QUEST system.

32
Due to issues related to confidentiality and size of data files, the Marion County
Juvenile Justice system (QUEST), Regenstrief Medical Record System (RMRS), and
Indiana Network for Patient Care (INPC), are stored and managed by Regenstrief
Institute, Inc., a private, non-profit research organization affiliated with the Indiana
University School of Medicine. A team of data analysts from Regenstrief Institute, Inc.
linked the case records for all detained adolescents within QUEST to the medical records
within RMRS and INPC using the probabilistic linkage software Recmatch (Grannis,
Overhage, Hui & McDonald, 2003). Recmatch linked files using a multi-step
probabilistic matching process. First, the parameters for the file search were limited to
year of birth, so that only files of individuals born between 1984 and 2001 (i.e.
adolescents ≤18 years during detention) were included. Second, a data analyst formatted
unique participant identifiers (e.g., social security number, last name, middle name, first
name, date of birth, gender, and race) from QUEST and RMRS/INPC databases into a
consistent fashion. These files were entered into the RecMatch software and matched
based on the unique participant identifiers (Grannis et al., 2003).
Prior to linking, the original Marion County Juvenile Justice system contained the
case records of approximately 50,000 juvenile justice-involved youth and the Regenstrief
Medical Record System and Indiana Network for Patient Care contained the medical files
of over 300,000 youth (United States of America Census Bureau). Upon completing the
matching process outlined above, the RecMatch output yielded 74,835 pairs or matches
between QUEST records and medical records, consisting of 42,148 unique QUEST
individuals and 74,823 unique RMRS/INPC individuals. On average, each QUEST
individual was linked to 1.78 RMRS/INPC individuals, due to either data entry errors
made by data analysts or software linking errors (i.e. inability of RecMatch to recognize
that files should be linked). To illustrate a software linking error, a youth named “John
Doe” in QUEST may be linked by RecMatch to two files in RMRS/INPC, one of a youth
named “John Doe” and another named “John A. Doe.” The two files in RMRS/INPC are
most likely the same person and should be linked together, but the RecMatch software
errs on the side of caution and keeps the files separate, rather than incorrectly joining two
different people. RecMatch is not able to recognize that these files belong to the same

33
person and incorrectly creates two pairs of files for John Doe. To address data linking
problems with RecMatch, the software program RecFam was used to analyze the pairs of
files (Grannis et al., 2003). RecFam reduced the 74,835 pairs to 40,015 “families” of
pairs, meaning the program deduced that the 74,835 pairs reflect 40,015 unique
individuals. A data analyst from Regenstrief Institute, Inc. validated RecFarm results by
reviewing the RecFam “families” of pairs and re-grouped pairs when necessary. This
final step resulted in the identification of 41,798 unique individuals with retrievable data
from both QUEST and RMRS/INPC databases.
As part of two larger grants examining behavior health among juvenile justiceinvolved youth (i.e., HRSA/MCHB R40MC08721 and HRSA/MCHB T7100008), data
analysts from Regenstrief Institute, Inc. were provided instructions regarding data
abstraction. De-identified datasets containing requested data were made available on a
private password-protected server accessible to members of Dr. Matthew Aalsma’s
research team, including me. For the purposes of this study, I accessed four, separate
Statistical Analysis System (SAS) files containing data regarding the following: 1) Arrest
history, detention stays, and recidivism, 2) Prison stays, 3) Mental health and substancerelated needs/diagnoses, 4) Treatment services (See next section for more details). To
maintain confidentiality, SAS datasets did not contain the names, addresses, or insurance
policy numbers of individuals; rather, each unique individual was assigned a file number
that served as the participant’s primary identification number. All datasets contained the
files number for each potential participant, thereby enabling me to use the file number as
the key variable to link and merge the four SAS datasets into one primary Statistical
Package for the Social Sciences-Version 22.0 (SPSS) dataset.

2.2 Data Coding
Data abstracted from the Marion County Juvenile Justice system and the
Regenstrief Medical Record System databases were reviewed and coded according to a
fixed protocol with coding rules for all variables. Please see Appendix Table A.1 for
codebook containing list of variables and coding rules. Overall, variables were
categorized and coded within the following four broad domains: 1) Demographics, 2)

34
Criminal history, 3) Mental health and substance-related needs/diagnoses, and 4)
Treatment services. The following briefly describes the key variables within each coding
domain.
Demographic variables. Data coded within this domain included participants’
date of birth, date of death (if applicable), gender, race/ethnicity, age upon intake of first
detention stay, and insurance status. With the exception of date of death and insurance
status, QUEST served as the primary source of information for these variables. For
race/ethnicity, participants were coded as American Indian/Alaskan, Asian/Pacific
Islander, Black/African American, Non-white Hispanic, Multi-racial, White, or Other.
Based on work by Teplin and colleagues (2006), age was coded into three age cohorts:
younger DAs (age 11-13), mid-age DAs (age 14-15), and older DAs (age 16-18). The
insurance status for all participants was coded as private insurance, Medicaid/Medicare,
Wishard Advantage, no insurance/self-pay, workers’ compensation, and
unknown/missing.
Criminal history variables. Data abstracted from QUEST and coded within this
domain included information pertaining to referrals, charges, detention stays, prison
stays, recidivism, and criminal-related risk (see Appendix Table A.1 for more details).
Referrals were defined as any contacts or arrests within juvenile justice system (Bureau
of Justice Statistics, 2015); variables were coded according to date of first referral with
the juvenile justice system, age at first referral, referral prior to first detention stay
(yes/no), and number of referrals prior to first detention stay. Criminal charges were
defined as formal accusations of crime made by a governmental authority, but not
necessarily convictions of these crimes (Bureau of Justice Statistics, 2015). Criminal
charges were coded according to type of charge (e.g., Criminal Trespass, Theft,
Possession of Marijuana), level of charge (e.g., Felony, Misdemeanor), class of charge
(e.g., A, B, C), violence-related charge (e.g., Battery, Rape, Kidnapping; yes/no ), and
referring agency making charge (e.g., Indianapolis police department, Marion County
sheriff). The most severe charge (out of all charges) was rated on a severity scale of 1 to
5, with 1 = status offense/probation violation, 2 = non-violent misdemeanor, 3 = violent
misdemeanor, 4 = non-violent felony, and 5 = felony. Additionally, total number of

35
charges and whether the adolescent was a violent offender (yes/no) at time of first
detention were coded for each participant. It should be noted that all charges associated
with a participant’s first detention stay were coded within this section. Additional
charges following first detention stay were coded under recidivism.
Regarding detention stays, variables associated with first detention stay were
coded according to date of entry, date of release, length of stay (in days), reason for stay
(e.g., Outright arrest, Probation violation, Warrant arrest), and reason for release (e.g.,
Released on home detention, Released to parent, Released to community adjustment). As
part of a sub-analysis, the year in which participants were first detained was coded into
two year cohorts: cohort one (detained during 1998 – 2005) and cohort two (detained
during 2006 – 2011). If participants had more than one detention stay, each subsequent
detention stay was coded according to date of entry, date of release, length of stay, and
timing (in days) since release from first detention. Other detention-related variables
coded within this domain included re-detained (yes/no), timing of second detention (e.g.,
within 2 weeks, within 1 month), number of detention stays within two years of first
detention, and total number of detentions. Similar to detention stays, all prison stays
within DOC facilities were coded by date of entry, date of release, length of prison stay
(in days), and timing since release from first detention stay. The total number of prison
stays was abstracted and coded; participants were also coded as incarcerated within two
years of first detention (yes/no) and incarcerated at any time during study follow-up
(yes/no).
With regards to recidivism, variables were coded according to re-arrest/new
charge within 6 months post-detention (yes/no), type of new charge, level of new charge
(e.g., Felony, Misdemeanor), severity of new charge (on scale of 1-5), number of distinct
recidivism events, and recidivism rate (e.g., number of recidivism events out of total
number of detentions). Finally, participants’ scores on the Risk Assessment Inventory
(RAI) were coded as a measure of criminal-related risk (i.e., likelihood to recidivate).
Total RAI score ranging from 0 to 24 were abstracted, with higher scores representing
higher risk. Based on these scores, participants were coded as low-risk (i.e., scores of 0-5
for males; 0-7 for females), medium-risk (i.e., scores of 6-12 for males; 8-13 for

36
females), and high-risk (i.e., scores of ≥13 for males; ≥14 for females). (Please see
Measures section below for more information about the RAI and its scoring protocol).
Mental health variables. Data coded within this domain included information
pertaining to participants’ mental health needs, mental disorders, and substance-related
disorders. The mental health needs of participants were measured via the Massachusetts
Youth Screening Instrument-2nd Edition (MAYSI-2), a self-report screener designed to
identify mental health and substance-related concerns among juvenile justice-involved
youth (Grisso & Barnum, 1998). Total scores and the seven subscales scores from the
MAYSI-2 were abstracted from QUEST, with higher scores indicating higher mental
health and/or substance-related needs. Scores associated with each subscale were coded
as falling in the caution range (yes/no), as well as the warning range (yes/no). An
adolescent was coded as having a positive screen (yes/no) on the MAYSI-2 if his/her
suicidal ideation subscale score fell in the caution or warning range or ≥2 subscales fell in
the caution or warning range (Grisso, Barnum, Fletcher, & Peuschold, 2001). (Please see
Measures section below for more details).
Contained with RMRS and IPN databases, the International Classification of
Diseases and Related Health Problems–Ninth Revision (ICD-9) codes for mental
disorders and substance-related disorders were abstracted. Based on these codes,
participants were coded as having any disorder(s) (yes/no), any mental disorder(s)
(yes/no), a conduct disorder(s) (yes/no), a non-conduct mental disorder(s) (yes/no), any
substance disorder(s) (yes/no), and comorbid for both types of disorders (yes/no). In
addition, all participants were coded according to disorder type (e.g., just mental
disorder(s), just substance-related disorder(s) and number of disorders. Mental disorders
were coded according to the following classifications: conduct-related disorders (e.g.,
conduct disorder, oppositional defiant disorder), mood disorders (e.g., major mood
disorder, bipolar disorder), anxiety disorders (e.g., generalized anxiety disorder,
obsessive-compulsive disorder, post-traumatic stress disorder), attentiondeficit/hyperactivity disorder, adjustment disorders or somatization disorders, psychosisrelated disorders (e.g., schizophrenia, delusional disorder), and other disorders (e.g.,
eating disorder, sexual dysfunctions, sleep wake disorder). Substance-related disorders

37
were coded according to the following classifications: alcohol-related disorders (e.g.,
alcohol use disorder, alcohol intoxication), cannabis-related disorders (e.g., cannabis use
disorder, cannabis intoxication), drug-related disorders (e.g., inhaling use disorder, opiate
use disorder, stimulant use disorder), and other disorders (e.g., fetal alcohol syndrome,
using drugs during pregnancy). Finally, the date of diagnosis was derived from medical
records, so that variables could be coded according to date of diagnosis (i.e., date in
which ICD-9 code first appeared in record) and timing of diagnosis (e.g., pre-, during, or
post-detention).
Treatment services. Data contained within the medical records of the Regenstrief
Medical Records Systems served as the primary source of information for this domain.
All treatment services were abstracted and coded according to a multi-step process. First,
treatment services were coded according to service type. Specifically, any treatment
service associated with ICD-9 codes for mental disorders was coded as mental health, any
treatment service associated with ICD-9 codes for substance-related disorders was coded
as substance-related, and any treatment service associated with ICD-9 codes for both
mental disorders and substance-related disorders or not associated with any ICD-9 codes
was coded as either/both (i.e., not clear if treatment focused on mental health concerns,
substance-related concerns, or both concerns). Second, treatment services were coded for
timing in relation to first detention stay; thus, treatment services were coded as predetention, during detention, or post-detention. Third, treatment services were coded
according to the following settings: outpatient treatment in a community-based center,
inpatient stay in hospital/residential facility, or visit to the emergency department (ED)
for behavioral health concerns. After determining that a relatively small number of DAs
utilized inpatient treatment (n = 1396, 14.5%) and ED visits (n = 1891, 19.5%), compared
to outpatient treatment (n = 6437, 66.61%), inpatient treatment and ED visits were
subsequently collapsed into one variable coded as non-outpatient treatment. Thus, all
treatment services utilized by participants were coded according to type (i.e., mental
health, substance-related, either/both), timing (i.e., pre-, during, or post-detention), and
setting (i.e., outpatient, non-outpatient).

38
Once treatment services were coded, numerous treatment-related variables were
derived and coded. Specifically, participants were coded as treatment users (yes/no),
mental health treatment users (yes/no), substance-related treatment users (yes/no),
either/both treatment users (yes/no), pre-detention users (yes/no), during detention users
(yes/no), post-detention users (yes/no), outpatient users (yes/no), and non-outpatient users
(yes/no). Participants were also coded by service type user (e.g., just mental health, just
substance use, two types) and service setting user (e.g., just outpatient, two types, all
three types). (See Appendix Table A.1 for more details). With regards to non-outpatient
treatment utilization, coded variables included prevalence rates for different groups of
DAs (e.g., males, females, younger), number of non-outpatient service visits, length of
stay (in days) for inpatient services, and time (in days) between non-outpatient services.
Intensity of non-outpatient treatment utilization within two years post-detention was
coded according to the following intensity levels: non-user, low user (1 visit/stay), lowto-moderate user (2 visits/stays), moderate user (3-4 visits/stays), high user (5-6
visits/stays), and extreme user (≥7 visits/stays).
Similarly, outpatient treatment utilization was coded for prevalence rates of
different groups of DAs, number of total outpatient sessions, time (in days) between
sessions, and number of sessions per month. Based on Cauffman and colleagues (2005)
intensity of service categories for outpatient treatment, intensity of outpatient treatment
utilization within two years post-detention was coded according to the following intensity
levels: non-user (0 sessions), low user (1-2 sessions), low-to-moderate service user (4-7
sessions), moderate user (8-12 sessions), moderate-to-high user (13-17 sessions), high
user (18-22 sessions), and extreme user (≥23 sessions). Participants were also coded as
outpatient dropouts (yes/no) if they terminated outpatient treatment within 1-3 sessions,
with no return to treatment for ≥24 months (Luk et al., 2001). Potential reasons to
explain outpatient dropouts were coded according to one of the following events
occurring during the same month of treatment dropout: placement in detention or prison,
placement in an inpatient facility, or no placement (i.e., no clear explanation). Finally,
participants were coded as having a gap in outpatient treatment (yes/no) if they
experienced an absence of treatment for 45-120 days between two consecutive sessions

39
of the same outpatient treatment type. Potential reasons for gaps between outpatient
treatment sessions were coded using the same categories as reasons for outpatient
dropouts.

2.3 Measures
Risk Assessment Inventory (RAI). The RAI represents a basic tool used during
the risk screening process, which typically occurs prior to or upon entry into a detention
center (Steinhart, 2006). Trained evaluators rate an adolescent using a written checklist
of criteria, in which points are assigned for different criminal history domains; domains
include most serious offense, additional offenses, supervision status (e.g., probation,
electronic monitoring), pending petitions/disposition, warrant history, prior offenses,
mitigating factors (e.g., school, family support), and aggravating factors (e.g., no
community ties, runaways, prior escapes). Points are summed to produce a total score
that represents overall criminal-related risk, or likelihood to recidivate. Scores can then
be used to categorize adolescents into low-risk, medium-risk, and high-risk offenders
(Latessa, Lovins, & Makarios, 2013). Since the Risk Assessment Inventory is intended
to suit local needs, this instrument tends to vary in scope, format, and scoring criteria
across sites (Steinhart, 2006). Therefore, members of the Indiana Risk Assessment Task
Force at the Indiana Judicial Center coordinated with the University of Cincinnati Center
for Criminal Justice Research to examine, validate, and norm the risk assessment process
for juvenile justice-involved youth in Indiana (Latessa et al., 2013). The following
standards were established to categorize risk: scores of 0-5 for males and 0-7 for females
represent low-risk, scores of 6-12 for males and 8-13 for females represent medium-risk,
and scores of ≥13 for males and ≥14 for females represent high risk.
In addition to level of risk, an RAI score serves as an objective standard to
determine placement for an arrested youth, such as placement in a secure juvenile justice
facility, non-secure alternative program (e.g., mental health treatment facility), or home
release. The use of standardized RAI scores has proven effective in reducing subjective,
biased placement decisions, as well as decreasing total admissions to juvenile justice
facilities and curbing the costs and liabilities associated with detaining adolescents who

40
do not require a stay in a secure detention facility (Steinhart, 2006). Since the Marion
County juvenile detention center (i.e., the site for this study) does not follow specific cutoff criteria to determine appropriate placement for adolescents, the total RAI score was
used as a general measure of criminality, with higher scores indicating more severe
criminal activity. It should be noted that the RAI was not incorporated into the risk
screening process at the Marion County juvenile detention center until 2006, so only
adolescents detained between 2006 and 2011 obtained RAI scores.
Massachusetts Youth Screening Instrument-2nd Edition (MAYSI-2). The
MAYSI-2 is a 52-item T/F mental health screener designed to identify juvenile-justice
involved youth with special mental health needs (Grisso & Barnum, 1998). The scale is
divided into seven subscales: Alcohol/Drug Use, Angry/Irritable, Depressed/Anxious,
Suicidal Ideation, Somatic Complaints, Traumatic Experiences, and Thought
Disturbance. Subscale scores are interpreted as falling in the normal, caution, or warning
range. Individuals are considered to screen positive on the measure if the suicidal
ideation subscale score falls within the caution or warning range or at least two subscales
fall in the caution or warning range (Grisso et al., 2001). A positive screen on this scale
is not interpreted as an official diagnosis, but as an indication of serious mental health
problems and likely treatment needs. The instrument has been normed and validated for
juvenile-justice involved youth, and has good internal consistency and convergent
validity (Archer, Simonds-Bisbee, Spiegel, Handel, & Elkins, 2010). Since the legal
mandate requiring the use mental health screenings in all detention facilities in Indiana
did not occur until 2006, only adolescents detained between 2006 and 2011 completed
the MAYSI-2 upon intake into the Marion County juvenile detention center.
International Classification of Diseases and Related Health Problems-9th Revision
(ICD-9). The International Classification of Diseases and Related Health Problems is a
standardized, worldwide classification system of health problems that contains codes for
diseases, disorders, symptoms, and abnormal findings (Center for Disease Control and
Prevention). ICD-9 codes, which include codes for mental disorders and substancerelated disorders, are consistently used in all healthcare facilities and allow for direct
comparison of individuals over time and across settings (Center for Disease Control and

41
Prevention). In contrast to the delayed implementation of the RAI and MAYSI-2 that did
not occur until 2006, ICD-9 codes were used throughout the entire time frame, thereby
eliminating any potential problems related to inconsistent coding for disorders across
participants.

2.4 Recruitment and Participants
Since de-identified data from participants’ case records were gathered via
electronic databases, no direct recruitment of participants or informed consent process
was conducted. Participants were required to meet the following inclusion criteria: 1) 1118 years of age upon entry into a juvenile justice facility, 2) at least one stay at a juvenile
justice facility (i.e., detention center or prison facility) due to being arrested for a criminal
offense, 3) medical records data within the Regenstrief Medical Record System, and 4)
electronic records within the Marion County Juvenile Justice system. Adolescents were
excluded from the study if they were involved in the justice system for non-criminal
offenses (e.g., custody dispute, victim of child neglect, or witness) or were arrested, but
not incarcerated for the arrests. Participants were not required to have been charged
and/or convicted of the arrest(s) that resulted in their detention, nor were there any
exclusions associated with criminal history (e.g., prior contact with the juvenile justice
system, number of prior contacts) or mental health status (e.g., severe mental illness,
history of prior treatment).
As mentioned in the Data Sources and Data Linking section above, the RecMatch
process successfully yielded a total of 41,798 individuals. Of these potential participants,
a total of 416 (1.0%) individuals were excluded due to being outside the required age
range, 21,866 (52.3%) individuals were excluded because they had contact with the
juvenile justice system but were not detained or incarcerated, and 2,126 (5.1%)
individuals were excluded for having blank medical records with no information about
mental health status or treatment utilization. After excluding these individuals, a total of
17,398 (41.6%) detained adolescents remained in the sample. Though not initially
proposed as part of the inclusion criteria, an additional 7,734 (18.5%) adolescents with
private insurance or no medical insurance were subsequently removed from the sample

42
because treatment utilization data for these individuals were largely missing in a nonrandom and biased manner. Since treatment utilization represents the primary outcome
of interest for this study, including these individuals would have likely resulted in biased
findings and increased chances of Type 1 and Type 2 errors (Yoo, 2010); thus, these
adolescents were removed prior to data analysis. The final sample therefore contained a
total of 9,664 detained adolescents with Medicaid insurance who met full inclusion
criteria for the study.

2.5 Study Setting
Given that data was derived from the Marion County Juvenile Justice system, all
participants were detained in the Marion County juvenile detention center. This shortterm facility, serves as the only juvenile detention center for Marion County, houses
adolescents who have been arrested for an offense or an arrest warrant, as well as
adolescents serving a sentence imposed by the juvenile court (Hoskins, 2008). Prior to
2006, the Marion County juvenile detention center detained a maximum of 144
adolescents at a time, but currently detains approximately 112 adolescents at a time
(Hoskins, 2008).

2.6 Data Analysis
Following data linking and coding, basic descriptive statistics (e.g., mean, range,
standard deviation) and frequency distributions were run for all variables to check for
data entry errors, coding errors, non-normal distributions, outliers, and missing values.
To address data entry errors, the original datasets provided by Regenstrief Institute, Inc.
were reviewed to determine the source of the errors and whether errors could be fixed. If
errors were present in the original datasets, I coordinated with Regenstrief data managers
to address errors that were able to be fixed. For example, the dates for detention entry
and/or release for some individuals were incorrect (e.g., -1999 rather than 1999), but easy
to fix. When possible, coding errors were corrected prior to conducting main analyses.
Variables exhibiting non-normal distributions (e.g., skewness ≥3 and kurtosis ≥5) were
log-transformed. These log-transformed variables were used for all parametric statistical

43
tests (e.g., ANOVAs, t-tests) that require variables to have normal distributions. The
original variables, with the exclusion of outliers (e.g. ≥3 standard deviations from the
mean), were used for non-parametric statistical tests (e.g., logistic regression, chi-square
tests) that do not require variables to have normal distributions. (Allison, 2013).
Descriptive statistics (e.g., means, standard deviations, medians, ranges, minimum
values, maximum values, distributions, outliers) were conducted to examine basic sample
demographics, including the distribution and prevalence of gender, race/ethnicity, and
age during first detention. Interactions between these demographic variables were
performed to determine the prevalence of White males, White females, Black males,
Black females, Other Minority males, and Other Minority females within each age
cohort. Once calculated, the number of Other Minority DAs in the sample (n = 630,
6.5%) was disproportionately smaller than Black DAs (n = 5667, 58.6%) and White DAs
(n = 3367, 34.8%), so these adolescents were dropped from analyses focused on direct
comparisons by race, but included in all other analyses. These DAs were excluded,
rather than collapsed into one minority group with Black DAs, because prior research
(Lopez-Williams et al, 2006; Lyons et al., 2003; Rawal et al., 2004) has shown
significant differences in key outcomes for Black DAs versus non-Black minority DAs.
Finally, adolescents with missing values were eliminated from analyses on a case-by-case
basis, using pairwise deletion.
Basic descriptive statistics were also run to find the means, standard deviations,
medians, ranges, minimum values, maximum values, and outliers for continuous criminal
history variables, such as age at first contact with juvenile justice system, number of
arrests prior to detention, number of charges, length of detention stay (in days), number
of prison stays, and RAI scores. Frequency counts were conducted to determine the
prevalence of categorical criminal history variables, including number of youth with
felony charges, drug-related charges, violent offenses, multiple detention stays, as well as
rated as low-risk, medium-risk, and high-risk offenders.
With respect to mental health and substance-related variables, the means, standard
deviations, and ranges of the seven MAYSI-2 subscale scores were calculated, as well as
the prevalence of DAs who scored in caution and/or warning ranges for each subscale.

44
Based on these results, the mean (and range) of elevated subscales in the caution and/or
warning ranges and the prevalence of DAs who screened positive on the MAYSI-2 were
determined. In addition, prevalence rates were calculated for the number of DAs with
mental disorders, substance-related disorders, comorbidity, diagnoses prior to detention,
diagnoses after release from detention, and meeting criteria for different types of
disorders (e.g., anxiety disorders, alcohol-related disorders, conduct-related disorders).
Analyses were also conducted to calculate the mean number of different disorders, as
well as mean days to obtain diagnosis in relation to first detention stay.

2.6.1 Data Analysis for Aim 1
Focused statistical analyses were performed to answer the research questions
associated with the four study aims. To address research question one of Aim 1 (i.e.,
conduct a cross-sectional exploration of post-detention treatment utilization), frequency
tables displaying the number and percentage of DAs utilizing services were generated to
determine prevalence rates for the following types of treatment: any treatment, mental
health treatment, substance-related treatment, and either/both treatment, as well as the
following treatment settings: outpatient treatment and non-outpatient treatment. For each
treatment type and treatment setting, analyses were re-run to determine prevalence rates
for pre-detention treatment utilization, as well as treatment utilization at 2-years, 18months, 1-year, 6-months, and 1-month follow-up from release from detention.
It should be noted that the total sample size decreased as length of post-detention
follow-up increased, due to attrition. Specifically, an increasing number of participants
had to be dropped from data analysis due to attrition-related issues including lack of
complete follow-up data, death during the follow-up period, and/or incarceration for
≥50% of the follow-up time period. Adolescents incarcerated for long-term stays were
eliminated from time-restricted analyses (e.g., within one year post-detention, two years
post-detention) because treatment utilization data while incarcerated in prison were
missing, so including these adolescents would have created skewed results that
underestimated treatment prevalence rates. Although these incarcerated youth were
excluded from time-restricted analyses, they were included in overall analyses for total

45
study time frame (i.e., survival analyses). Further, due to attrition and disparate followup periods across participants that ranged from 1 year to 14 years post-detention, analyses
for Study Aims 1, 2, and 3 were restricted to the two-year period following release from
first detention stay; thus, all participants had the same time frame to receive services.
Consistent with data analyses already described, analyses for question two of Aim
1 entailed descriptive statistics to determine the quantity (i.e., mean, standard deviation,
and ranges) of unique treatment services used by DAs, as well as mean time (in days)
between services, number of services per month, and length of stay (in days) for inpatient
treatment. For research questions 3-4 of Study Aim 1, frequency tables were generated to
determine prevalence rates for types of services users (e.g., just mental health, two
treatment types), setting for service user (e.g., outpatient, inpatient), intensity of service
users (e.g., low users, moderate users, extreme users), outpatient treatment dropouts, and
gaps between outpatient treatment services. When applicable, analyses were re-run to
examine the prevalence and quantity of treatment services for each of the three treatment
types and two treatment settings.

2.6.2 Data Analysis for Aim 2
To answer the research questions associated with Aim 2 (e.g., examine group
differences in post-detention treatment utilization), a series of statistical tests were
conducted to identify significant differences in treatment services across groups of DAs.
The following groups served as the independent variables: gender (male vs. female), race
(White DAs vs. Black DAs), age cohorts (younger age, mid-age, and older age), MAYSI2 (did not take the MAYSI-2, positive screen, non-positive screen), disorder type (mental
disorder, substance-related disorder, comorbid), pre-detention treatment (yes/no), violent
offender (yes/no), recidivist (yes/no), and re-detained/incarceration after first detention
(yes/no). The following outcomes calculated in Aim 1 served as the dependent outcome
variables for Aim 2: 1) prevalence of treatment utilization, 2) total number of services, 3)
number of outpatient sessions per month, 4) time between treatment services, 5) length of
stay (in days) for inpatient treatment services, 6) type of service user, 7) intensity of
treatment services, 8) treatment dropouts, and 9) gaps between outpatient sessions.

46
Analyses focused on one independent variables and one dependent variable at a
time. Statistical tests were selected to appropriately match the type of independent and
dependent variables. For independent variables consisting of only two categorical groups
(e.g., gender [male/female], violent offender [yes/no]), independent t-tests were used to
test for differences between groups when dependent variables were continuous (e.g.,
number, frequency of session); Mann-Whitney U-tests were used when dependent
variables were ordinal (e.g., intensity of service user) and 2x2 chi-square tests (χ2) were
used when dependent variables were categorical (e.g., treatment [yes/no], dropout
[yes/no]). For independent variables consisting of three groups (e.g., age [younger, midage, and older], disorder type [mental health, substance-related, and comorbid], one-way
analysis of variance (ANOVAs) tests were used to test for differences between groups
when dependent variables were continuous, Kruskal-Wallis ANOVA tests were used
when dependent variables were ordinal, and 3x2 chi-square tests were used when
dependent variables were categorical. If the overall ANOVA or Kruskal-Wallis ANOVA
produced a significant finding (p ≤ .01), a series paired t-tests were run to identify which
of the three groups significantly differed from each other. Similarly, when 3x2 chisquare tests were significant, subsequent 2x2 chi-square tests were conducted to identify
group differences. To control for alpha inflation and reduce the likelihood of Type 1
errors (Altman, 2000), significance levels for all statistical tests for Aim 1 and 2 were set
at p ≤ .01.

2.6.3 Data Analysis for Aim 3
Hierarchical logistic regression was used to answer the research questions posed
in Aim 3 (e.g., identify predictors of post-detention service utilization). Prior to running
any regression analyses, bivariate correlations between predictor variables and
collinearity statistics for all predictor variables were examined to check for
multicollinearity. Strong correlations (r ≥ 0.75), large inverse inflation factors (≥5.0, p
≤ .10), and/or small tolerance estimates (≤.20) associated with a predictor were
considered indicators of multicollinearity (O’Brien, 2007). Predictors showing evidence

47
of multicollinearity were excluded from analyses. Outlier and missing values were
also removed using pairwise deletion and not included in the regression analyses.
Altogether, a series of six separate hierarchical logistic regression models were
run, with one outcome per model. The models focused on predicting one of the
following outcomes within two years of release from detention: 1) any treatment
utilization (yes/no), 2) any mental health treatment utilization (yes/no), 3) any substancerelated treatment utilization (yes/no), 4) any either/both treatment utilization (yes/no), 5)
any outpatient treatment utilization (yes/no), and 6) any non-outpatient treatment
utilization (yes/no). Based upon prior research that has used logistic regression analyses
to predict treatment utilization (Johnson et al., 2004; Lopez-Williams et al., 2006; Teplin
et al., 2005), predictor variables were entered via three stages. Predictors for stage one
included the following demographic variables: male gender (yes/no), Black race (yes/no),
and age upon detention entry. Predictors for stage two entailed the following criminal
history variables: number of arrests prior to detention, number of charges at detention
entry, charge severity (1 to 5), violent offender (yes/no), and length of first detention stay.
Predictors for the third stage included the following mental health variables: conductrelated disorder (yes/no), non-conduct mental disorder (yes/no), substance-related
disorder (yes/no), total number of disorders, pre-detention outpatient treatment (yes/no),
and pre-detention non-outpatient session (yes/no).
It should be noted that recidivism (yes/no) and re-incarceration (yes/no) were
initially intended to be included as predictors in stage two, but were eliminated from
analysis due to uncertainty about timing in relationship to treatment utilization. As
predictors of treatment utilization, these events needed to occur before service use.
However, due to the nature of the data, the dates for recidivism events were not known;
rather, the time frame (e.g., three months, six months) in which the recidivist event
occurred was known. Based on rough estimates, only a small percentage (16.2%) of
recidivism events occurred prior to post-detention treatment utilization, so recidivism was
unable to serve as a predictor variable. Similarly, the majority of re-detentions or reincarcerations (66.1%) occurred after treatment utilization, making this variable
inappropriate to serve as predictor variable for post-detention treatment utilization.

48
Scores from the Risk Assessment Inventory (RAI) and Massachusetts Youth
Screening Instrument-2nd Edition (MAYSI-2) were also intended to be included as
predictors of treatment utilization. However, only adolescents detained after 2005
completed these measures. To examine the potential importance of RAI scores and a
positive screen (yes/no) on the MAYSI-2, additional logistic regression analyses were
performed comparing results for adolescents in cohort one (i.e., detained prior to the
implementation of these measures) versus cohort two (i.e., detained post-implementation
of these measures). For analyses of cohort one, the six separate hierarchical logistic
regression models were re-run using the same predictors outlined above. For cohort two,
the six models were re-run with RAI scores included as a predictor in stage two predictor
and positive screen on the MAYSI-2 (yes/no) as a predictor in stage three.
After entering the predictors into the regression models, the following model
summary statistics were examined to compare the goodness of fit for each stage of the
analyses: -2 Log Likelihood, Nagelkerke R2, chi-square test, and classification
percentage. With regards to interpreting these model summary statistics, the value of the
-2 Log Likelihood is not informative on its own, but can be used to compare model
stages, with smaller numbers indicating a better fitting model (Allison, 2013). The
Nagelkerke R2 is a pseudo-R2 value, meaning it cannot be interpreted as R2 values are in
linear regression analyses (i.e., the percent of variance in the dependent variable
explained by the predictors) (Allison, 2013). Instead, this statistic reflects the percent of
improvement compared to a null model with no predictors (i.e., model predicts the mean
of the dependent variable based on no information or predictors) (Menard, 2000). Since
there are no clear standards for what constitutes a strong Nagelkerke R2 value, these R2
values are interpreted by comparing values across different stages; larger R2 statistics
indicate better fitting models (UCLA: Institute for Digital Research and Education). The
chi-square test examines the probability of obtaining results if there is no effect of the
predictor variables on the dependent variable. Significant findings therefore indicate that
the model is significantly different than the hypothesized null model, and that the
independent variables are significant in predicting the dependent variable (UCLA:
Institute for Digital Research and Education). Finally, the classification percentage refers

49
to the overall percentage of cases correctly predicted by the regression model using the
specified predictors, with higher percentages indicating a more accurate predictive model
(UCLA: Institute for Digital Research and Education). In addition to model summaries,
log-odds estimates, adjusted odds ratios, 95% confidence intervals for the odds ratios,
and significance values for all independent predictors at stage three were calculated to
identify the predictors that significantly (p ≤ .05) impacted the likelihood of treatment
utilization.

2.6.4 Data Analysis for Aim 4
Survival analyses were used to answer the research questions associated with Aim
4 (e.g., examine longitudinal patterns of post-detention treatment service utilization).
First, two separate Kaplan-Meier survival analyses were run to examine the following
outcomes: 1) connection to care, or time to first post-detention treatment utilization and
2) retention in outpatient services, or time to termination from outpatient treatment
services. For the first outcome, the origin of time was defined as the date of release from
first detention stay and the endpoint was defined as date of first treatment service (or end
of study time frame). Since the origin of time differed across DAs, follow-up times
ranged from 0 years (i.e. released in 2011 at the end of data collection) to 14 years (i.e.
released from detention in 1998 at the beginning of data collection). The metric of time
was continuous, measured in days since release from detention. Participants who did not
utilize treatment were deemed non-users and treated as censored cases (Cloyes, Wong,
Latimer, Abarca, 2010; Corning & Malofeeva, 2004). For the second survival analysis,
the origin of time was defined as the date of first outpatient treatment session and the
endpoint was the date of final outpatient session. The metric of time for these analyses
was also continuous, measured in days since participation in first outpatient session.
Again, participants who did not utilize outpatient treatment were treated as non-users and
censored cases (Corning & Malofeeva, 2004).
Descriptive statistics for the survival analysis models were generated and
examined via life tables, which show the event histories of participants from the
beginning to the end of data collection (Bewick, Cheek, & Ball, 2004). Given the

50
extended length of follow-up (i.e., 14 years), time for the life tables were divided into 6month time intervals. For each time interval, the following statistics were generated:
number of treatment users during time interval, censored cases (i.e., non-users), risk set
(i.e., number of adolescents eligible to experience the outcome), probability of outcome
(i.e., treatment utilization, treatment termination), and hazard rate (Bewick et al., 2004;
Corning & Malofeeva, 2004).
To address question two of Aim 4, analyses were re-run to examine group
differences. Specifically, Kaplan-Meier survival analyses were conducted to determine
whether the two main outcomes significantly differed (p < .01) for the following
independent variables: male gender (yes/no), Black race (yes/no), age cohorts (younger
age [≤13 years], mid-age [14-15 years], and older age [≥16 years]), MAYSI-2 (did not
take the MAYSI-2, positive screen, non-positive screen), disorder type (mental disorder,
substance-related disorder, comorbid), violent offender (yes/no), repeat offender (yes/no),
repeat detention/incarceration (yes/no), pre-detention treatment (yes/no), treatment type
(mental health, substance-related either/both), and setting (outpatient, non-outpatient).
The median time to event (i.e., treatment utilization, dropout) for each group was
calculated (Willie, 2012).
Survival curves for each group were directly compared using several chi-square
tests, including the Log Rank, Breslow/Wilcoxon, and Tarone-Ware (Bewick et al., 2004;
Bouliotis & Billingham, 2011). The Log Rank chi-square represents the most common
and frequently used test for identifying differences in survival analysis outcomes (Willie,
2012), with significant results indicating that survival curves differ significantly across
groups in the long-term. This test can be limited in examining survival curves that
intersect over time, which typically yield non-significant results, even though curves may
be significantly different at other follow-up time points (Bouliotis & Billingham, 2011).
Thus, alternative chi-square tests include weighted log-rank tests, such as the
Breslow/Wilcoxon test and Tarone-Ware test. The Breslow/Wilcoxon test is more
sensitive to differences in early follow-up periods, with significant results indicating
survival curves differ significantly across groups in the short-term (Willie, 2012). The
Tarone-Ware chi-square test considers the overall survival curve and has been shown to

51
be superior to the Log Rank test because it can better compare variables with more than
two levels (Willie, 2012). Significant results for the Tarone-Ware tests indicate that
survival curves differ significantly across groups in the middle portion of the follow-up
time, with the magnitude of results typically falling in between the results for the Log
Rank test and Breslow/Wilcoxon test. In addition to examining the chi-square tests, the
survival curves for different groups for each independent variable were generated and
displayed via graphs to visually examine longitudinal patterns of treatment utilization.
To identify significant predictors of treatment utilization over time, Cox
proportional hazards regression models with time-dependent variables were conducted.
Cox proportional hazards regression analyses can be interpreted like other regression
analyses, in that multiple variables can be entered into the model in stages to determine
whether they significantly impact the risk of an outcome (Bewick et al., 2004). Similar to
the logistic regression analyses, the following variables were entered via three separate
stages: male gender (yes/no), Black race (yes/no), age, number of pre-detention arrests,
number of charges upon detention entry, charge severity (1 to 5), length of detention stay,
violent offender (yes/no), conduct-related disorder (yes/no), non-conduct mental disorder
(yes/no), substance-related disorder (yes/no), pre-detention outpatient treatment (yes/no),
and pre-detention non-outpatient treatment (yes/no). In addition, being redetained/incarcerated (yes/no) was included as a criminal history variable in stage two.
Since this event occurred after release from detention and possibly after treatment
utilization, the variable was not included as a predictor in the logistic regression analyses.
However, Cox proportional hazards regression models are able to control for timing
effects by treating such variables as time-dependent, in which the value of the variable is
expected to change (i.e., from no [0] to yes [1]) within the same time frame as the
occurrence of the outcome (Corning & Malofeeva, 2004). The model handles timedependent variables by excluding participants via pairwise deletion who experience the
independent variable after the dependent variable (i.e., re-detained after treatment
utilization), while still retaining participants for the other components of the Cox
regression analyses pertaining to the other independent variables (i.e., hazard ratios for
independent variables like age, race, charge severity) (Corning & Malofeeva, 2004).

52
Similar to logistic regression analyses, additional analyses were conducted to
examine differences between DAs within cohort one versus cohort two. Analyses for
cohort one included the same predictors; analyses for cohort two included the addition of
RAI scores in stage two and MAYSI-2 positive screen (yes/no) in stage three. Several
goodness-of-fit statistics were examined to determine model fit, including -2 Log
Likelihood and chi-square tests (Corning & Malofeeva, 2004). Finally, the hazard ratios,
standard errors, and significant values for each predictor variable were calculated to
identify significant variables (p ≤ .05) associated main outcomes. Hazard ratios are
interpreted like odds ratio, with the exception that hazard ratios indicate the risk of an
outcome at any time during the 14-year follow-up period for one group compared to
another group, whereas odds ratios indicate the likelihood of an outcome by the endpoint
of a follow-up period (Bewick et al., 2004). Hazard ratios are considered to remain
constant over time (Corning & Malofeeva, 2004), so that the risk of an outcome like
dropping out of treatment is the same within two years of detention release as it is within
ten years of detention release. All data analyses were conducted using the software
program SPSS0-Version 22.0 and all study procedures were approved by the institutional
review board at Indiana University-Purdue University Indianapolis.

2.7 Statistical Power Analysis
The overall sample for the study included 9,664 adolescents, with approximately
8915 adolescents included in time-restricted analyses for treatment services within two
years of detention release. The primary statistical tests conducted for this this study
included t-tests, chi-square tests, ANOVAs, Pearson’s correlations, logistic regression
analyses, and Cox regression analyses. According to G*power statistical power analysis
software (Erdfelder, Faul, & Buchner, 1996), with a stringent alpha level of p < .01 and a
power level of ≥.95, the sample must contain a minimum of 5956 adolescents to have
enough statistical power to detect small significant effects (≥.10) using t-tests, at least
2268 adolescents to detect small effects using chi-square tests, at least 2070 adolescents
to detect small significant effects using ANOVAs between 3 groups, at least 595
adolescents to detent small significant effects using correlations, and at least 10861

53
adolescents to detect small effects for logistic regression analyses. Smaller samples are
required to detect significant medium (≥.30) or large (≥.50) effects using the various
statistical tests (G*power). Given the sample includes more than 9,000 youth, the study
contained enough statistical power to find significant results for small, medium, and/or
large effects, with the exception of being underpowered to detect small effects via
regression analyses.

54

CHAPTER 3. RESULTS

3.1 Sample Demographics
Please see tables and figures referenced within this section for full results. The
sample of participants who met full inclusion criteria included 9,664 detained adolescents
(DAs) with Medicaid coverage. A total of 6062 (62.7%) were male, 3367 (34.8%) were
White, 5667 (58.6%) were Black, and 630 (6.5%) were Other Minority races/ethnicities
(e.g., non-White Hispanic, American Indian/Alaskan, Asian/Pacific Islander). When
considering gender and race/ethnicity together, DAs included 2010 White males (32.2%
of males; 59.7% of White youth; 20.8% of sample), 3631 Black males (59.7% of males;
63.9% of Black youth; 37.5% of sample), 199 Other Minority males (7.1% of males;
68.4% of Other youth; 2.1% of sample), as well as 1357 White females (37.7% of
females; 40.3% of White youth; 14.0% of sample), 2046 Black females (56.8% of
females; 36.1% of Black youth; 21.2% of sample), and 321 Other Minority females
(5.5% of females; 31.6% of Other; 4.5% of sample).
Results revealed significant effects for race x gender (Table 3.1), with a positive
correlation between male gender and minority race (r = .04, p < .001) (Table 3.2). While
the overall sample included 62.7% males and 37.4% females, analyses found higher
percentages than expected for White females (40.3%) than White males (59.7%), and
higher percentages for Black males (63.9%) than Black females (36.1%; χ2 = 15.86, p
< .001) and higher percentages for Other Minority males (68.4%) than Other females
(31.6%; χ2 = 20.30, p < .001). As noted in the methods section, Other Minority DAs (n =
630, 6.5%) in the sample were dropped from subsequent analyses focused on
race/ethnicity, but included in all other analyses. Upon excluding Other Minority DAs,
the resulting sample for analyses examining racial differences included 9034 adolescents
(62.3% male; 62.7% White, 37.3% Black).

55
DAs were approximately 14.70 years (SD = 1.64, range = 11-18) upon entry into
their first detention stay. Females (M = 14.83, SD = 1.45) were significantly older than
males (M = 14.62, SD = 1.68) upon detention entry (t(1) = 38.51, p < .001); white DAs
(M = 14.80, SD = 1.64) were significantly older than Black DAs (M = 14.64, SD = 1.65;
t(1) = 18.29, p < .001). When divided into age cohorts, the sample consisted of 2406
(24.9%) younger DAs aged 11-13 years, 3925 (40.6%) mid-age DAs aged 14-15 years,
and 3334 (34.5%) older DAs aged 16-18 years (Table 3.1). Chi-square tests (χ2) revealed
significant effects for Gender x Age, with higher percentages of males (26.8%) than
females (21.7%) in the younger age cohort compared to lower percentages of males
(59.8%) than females (65.8%) in the mid-age cohort (χ2 = 22.02, p < .001), and lower
percentages of males (33.2%) than females (36.7%) in the older age cohort (χ2 = 4.74, p
< .001). Analyses also revealed significant effects for Race x Age; there were significant
differences in the proportion of White DAs (51.5%) versus Black DAs (55.0%) in the
young cohort compared to the proportion of White DAs (48.5%) versus Black DAs
(45.0%) in the mid-age cohort (χ2 = 7.80, p = .005) and compared to White DAs (61.6%)
versus Black DAs (56.3%) in the old age cohort. Results were not significantly different
for the proportion of males versus females (χ2 = 1.42, p = .231) or the proportion of Black
DAs versus White DAs (χ2 = .167, p = .171) in the mid-age cohort compared to the older
age cohort.

3.2 Criminal History Outcomes
Altogether, 9304 DAs (96.3%) had contact with the juvenile justice system prior
to first detention. For these DAs, time between first contact and first detention averaged
304.72 days (SD = 491.59, Mdn = 68.0, range = 0-840.15). The average age of first
contact with the juvenile justice system was 14.10 years (SD = 1.82, range = 6.0-18.0)
and the average number of contacts prior to first detention was 3.22 (SD = 2.77, range =
0.0-44.0). As shown in the correlation matrix of Table 3.2, demographic factors were
significantly related to numerous criminal history variables. Significant correlations were
found between male gender and age of first contact (r = -.90, p < .001) and number of
pre-detention arrests (r = .16, p < .001), between Black race and age of first contact (r

56
= .07, p < .001) and number of arrests (r = -.10, p < .001), and between age at detention
entry and number of prior arrests (r = .10, p < .001). Males (M = 13.98, SD = 1.88) were
significantly younger than females (M = 14.33, SD = 1.66) at first contact (t(1) = 38.51, p
< .001) and had significantly more contacts with the juvenile justice system (M = 3.57,
SD = 2.95, range = 0.0-44.0) than females (M = 2.65, SD = 2.30, range = 0.0-22.0) (t(1)
= 257.72, p < .001). Similarly, Black DAs (M = 14.00, SD = 1.83) were significantly
younger than White DAs (M = 14.00, SD = 1.79) at first contact (t(1) = 39.30, p < .001)
and had significantly more prior arrests (M = 3.41, SD = 2.84, range = 0.0-44.0) than
White DAs (M = 2.87, SD = 2.56, range = 0.0-28.0; t(1) = 84.32, p < .001).
Participants were first detained within the Marion County juvenile detention
center between the years of 1998 and 2011, for an average of 9.72 days (SD = 15.45,
range = 0.0-180.0) in detention. As shown in Table 3.3, detention rates declined over
time, with a notable drop in the number of youth placed in detention after 2006.
Specifically, approximately 6815 (70.5%) adolescents were detained within the 8-year
time frame of 1998 to 2005 (i.e., cohort one), whereas only 2849 (29.5%) adolescents
were detained within the 6-year time frame of 2006 to 2011 (i.e., cohort two). Although
the number of detained youth declined over time, the average length of stay increased
over time (r = .06, p < .001), with the largest mean length of stay occurring between 2008
and 2009 (M = 12.63, SD = 18.29, range = 0.0-141.0). The majority of DAs were
arrested and referred to the Marion County juvenile detention center by the Indianapolis
Police Department (n = 3711, 38.40%), the Marion County Sheriff (n = 2308, 23.9%),
the Indianapolis Metropolitan Police Department (i.e., consolidation of Indianapolis
Police Department and Marion County Sheriff Department in 2007; n = 1082, 11.2%), or
Police Departments of Indianapolis suburbs (e.g., Carmel, Lawrence; n = 1130, 11.3%).
The remaining youth (n = 1170, 12.1%) were referred to detention via school districts,
domestic referrals, or unknown agencies.
As documented in participants’ juvenile justice records, DAs averaged 1.21
charges (SD = 1.22, range = 1.0-104.0) upon detention entry. Table 3.4 displays the
types of criminal charges associated with detention, reasons for detention stays, and
subsequent placements upon release from detention. The most common types of charges

57
included conduct-related charges (n = 4445, 46.0%), property-related charges (n = 2567,
26.6%), and drug/alcohol-related charges (n = 439, 4.5%). With regards to the most
severe charge (1 to 5), there were 2095 (21.7%) DAs with ratings of 5 for violent
felonies, 1114 (11.5%) DAs with ratings of 4 for non-violent felonies, 1450 (15.0%) DAs
with ratings of 3 for violent misdemeanors, 4175 (43.2%) DAs with ratings of 2 for nonviolent misdemeanors, and 511 (5.3%) DAs with ratings of 1 for status offenses or
probation violations. In addition, 319 (3.3%) DAs were not rated, due to unknown
charges. Altogether, 3542 (36.7%) DAs were charged with a violent crime (e.g., Assault,
Rape, Robbery with Deadly Weapon) and therefore classified as violent offenders.
As displayed in Table 2, number of charges was not significantly correlated with
any demographic variables, but higher charge severity was significantly correlated with
male gender (r = .13, p < .001) and younger age at detention entry (r = -.03, p = .002).
Although effect sizes tended to be small, being a violent offender was significantly
related to male gender (r = .04, p < .001), Black race (r = -.03, p < .001), younger age of
first contact with the juvenile justice system (r = -.07, p < .001) and younger age at
detention entry (r = .15, p < .001). Most DAs were placed in detention due to being
outright arrested t (n = 4953, 51.3%) or awaiting action for pending or past charges (n =
3314, 34.3%). Upon release, most DAs were placed on home detention (n = 4687,
48.5%) or released into the care of parents/guardians (n = 2542, 26.3%).
The RAI was administered to DAs within cohort two (i.e., detained during 20062011) to determine risk of recidivism. Of the 2849 adolescents in cohort two, 2568
(90.1%) DAs obtained a mean RAI score of 11.69 (SD = 3.89, range = 0.0-24.0). Males
(M = 12.24, SD = 3.72, range = 0.0-24.0) obtained significantly higher risk scores than
females (M = 448, SD = 4.06, range = 0.0-22.0; t(1) = 128.45, p < .001), and Black DAs
(M = 12.24 SD = 2.39, range = 0.0-24.0) obtained significantly higher risk scores than
White DAs (M = 10.87, SD = 4.31, range = 0.0-22.0; t(1) = 58.48, p < .001). In addition,
older DAs (M = 11.27, SD = 3.98, range = 0.0-22.0) obtained significantly lower risk
scores than mid-age DAs (M = 12.23, SD = 3.82, range = 0.0-24.0; t(1) = 23.21, p
< .001), as well as younger DAs (12.21, SD = 3.66, range = 0.0-24.0; t(1) = 18.62, p
< .001). With regards to level of risk for recidivism, RAI scores classified 201 (7.8%)

58
DAs as low-risk offenders, 1292 (50.3%) as medium-risk offenders, and 1075 (41.9%) as
high-risk offenders. Table 3.5 lists the number of DAs within each of the three RAI risk
levels according to gender, race, and age cohorts. In comparing across demographic
groups, analyses showed significant differences in RAI risk levels for gender (U =
28842.0, p < .001), race (U = 45638.5, p < .001), younger DAs versus older DAs (U =
464593.0, p < .001), and mid-age DAs versus older DAs (U = 244020.5, p = .004).
Table 3.6 displays key findings for criminal activity following release from
detention. A total of 6401 (66.2%) DAs experienced a recidivism event, meaning they
were re-arrested at some point following detention. Prevalence rates for recidivism
differed significantly across gender, race, and age; recidivism was significantly higher
among males than females (χ2 = 169.66, p < .001), Black DAs than White DAs (χ2 =
147.23, p < .001), and younger DAs than mid-age DAs and older DAs (χ2 = 895.94, p
< .001) (see Table 5). With regards to timing, 1953 (20.2%) were re-arrested with 3
months of detention release, 1016 (10.5%) were re-arrested within 4-6 months of
detention release, 801 (8.3%) were re-arrested within 7-24 months of detention release,
and 2631 (27.2%) were re-arrested more than two years after detention release. Similar
to charges upon detention entry (Table 4), the most severe charges associated with
recidivism at 6-month follow-up entailed conduct-related charges (n = 1024, 10.6%) and
property-related charges (n = 747, 7.7%).
In addition to being re-arrested, a total of 5227 (54.1%) DAs had a second
detention stay within approximately 259.63 days (SD = 326.44, range = 0.0-2431.0). As
shown in Table 5, a significantly larger proportion of males versus females (χ2 = 85.65, p
< .001), Black DAs versus White DAs (χ2 = 44.57, p < .001), and younger DAs and midage DAs than older DAs (χ2 = 784.49, p < .001) were re-detained. Most second
detentions (n = 3,993; 90.0%) occurred within two years of first detention and lasted
approximately 11.78 days (SD = 13.29, range = 0.0-180.0) days. Altogether, the sample
experienced a mean of 2.16 (SD = 1.64, range = 1.0-12.0) detentions within two years
and 2.65 (SD = 2.26, range = 0-17) detentions across the entire study time frame. When
restricting analysis to only DAs with multiple detentions (≥2 detentions), the average
number of detentions increased to 4.0 (SD = 2.20, range = 2.0-17.0).

59
A small number of DAs (n = 1538; 15.9%) were committed to the Department of
Corrections (DOC) for long-term prison stays (Table 6). These DOC commitments
occurred within approximately 620.06 days (SD = 447.44, range = 0.0-2305.0) of release
from detention, with a mean length of stay of 359.47 days (SD = 280.91, range = 1.01709.0). Significant correlations were found between being incarcerated and being male
(r = 0.03, p = .005) and younger at detention entry (r = -.19, p < .001) (Table 2). The
proportion of DAs who were incarcerated was not significantly different for gender and
race, but differed significantly for age (χ2 = 324.56, p < .001). As displayed in Table 5,
there were significantly more younger DAs who were placed in prison than middle age
DAs (χ2 = 80.34, p < .001), and more mid-age DAs who were placed in prison than older
DAs (χ2 = 37.72, p < .001). Out of the 1538 participants who were incarcerated, 1267
(82.4%) had one prison stay, 235 (15.3%) had two prison stays, and 36 (2.3%) had 3-4
prison stays. The mean number of commitments for these youth across the entire study
time frame was 1.20 prison stays (SD = .46, range = 1.0-4.0).

3.3 Mental Health and Substance-Related Outcomes
Similar to the Risk Assessment Inventory, the Massachusetts Youth Screening
Instrument, 2nd edition (MAYSI-2) was implemented within Marion County juvenile
detention center in 2006. A total of 2575 out of 2849 (96.6%) DAs within cohort two
completed the MAYSI-2 in approximately 5.38 minutes (SD = 6.35, range = 0.0-93.37).
As shown in Table 3.7, the Angry-Irritable subscale had the highest mean score (M =
4.19 of 9.0, SD = 2.82, range = 0.0-9.0) and the Thought Disturbances subscale had the
lowest mean score (M = 0.71 of 5.0, SD = 0.98, range = 0.0-5.0). A total of 2065
(80.2%) DAs scored in the Caution range and 867 (33.7%) DAs scored in the Warning
range for at least one of the seven subscales. Being female (r = -.20, p < .001), White
race (r = .12, p < .001), and younger age (r = .05, p = .014) was significantly associated
with larger number of scales falling in the caution or warning range. Please see
correlation matrix displayed in Table 3.8 for all associations between demographic
variables and mental health variables. Approximately 1748 (67.9%) DAs screened
positively on the MAYSI-2; there were significantly higher proportions of positive

60
screens among females than males (χ2 = 145.90, p < .001), Black DAs than White DAs
(χ2 = 288.39, p < .001), and younger DAs than mid-age and older DAs (χ2 = 35.59, p
= .001) (Table 3.5).
Table 3.9 displays results for mental disorders and substance-related disorders
among the total sample. Altogether, 7015 (72.6%) DAs met criteria for at least one
disorder, including 2689 (74.7%) females, 4326 (71.4%) males, 2690 (79.9%) White
DAs, 3847 (67.9%) Black DAs, 1867 (77.6%) young DAs, 2869 (73.1%) mid-age DAs,
and 2279 (68.4%) old DAs (Table 3.5). For these DAs, 3571 (50.0% with disorders;
37.0% of sample) adolescents had only mental disorders, 622 (8.9% with disorders; 6.9%
of sample) adolescents had only substance-related disorders, and 2782 (39.6% with
disorders; 28.8% of sample) adolescents were comorbid for both types of disorders. For
the 7015 youths diagnosed with a disorder, 2442 (34.81%) DAs obtained their first
diagnosis (as listed on their medical records) approximately 930.55 days (SD = 798.90,
range = 0.0-2901.0) prior to first detention stay. In contrast, 16 (0.2%) DAs obtained
their first diagnosis during detention within approximately 22.31 days (SD = 25.18, range
= 0.0-80.0) of detention entry and 4557 (65.0%) DAs obtained their first diagnosis within
a mean of 1344.33 days (SD = 1178.21, range = 0.0-4978.0) following release from
detention. For youth diagnosed post-detention, 1838 (40.3%) received a diagnosis within
two years of release from detention.
Participants averaged 3.62 (SD = 4.79, range = 0.0-43.0) unique disorders; 1629
(23.2%) DAs had only one disorder, whereas 1440 (16.3%) DAs had at least eight unique
disorders. Increased number of disorders was significantly related to female gender (r =
-.09, p < .001), White race (r = .18, p < .001), and younger age at detention entry (r
= .09, p < .001) (Table 3.8). In contrast, pre-detention diagnosis was significantly
associated with male gender (r = .20, p < .001), Black race (r = -.05, p < .001) and older
age at detention entry (r = .13, p < .001).
A total of 6353 (65.7%) DAs met criteria for at least one mental disorder (i.e.,
excluding substance-related disorders). With regards to timing of diagnosis, 2315
(36.4%) adolescents obtained a diagnosis approximately 969.16 days (SD = 800.36,
range = 0.0-2901.0) prior to diagnosis, 16 (0.2%) youth obtained a diagnosis within 21.18

61
days (SD = 25.78, range = 0.0-80.9) after detention entry, and 4022 (57.3%) youth
obtained a diagnosis within an average of 1365.96 days (SD = 1202.24, range = 0.04942.0) after release from detention. Less than half of the DAs (n = 1617, 41.7%)
diagnosed post-detention obtained a diagnosis within two years of detention release. As
shown in Table 9, participants averaged 2.90 (SD = 4.08, range = 0.0-35.0) distinct
mental disorders, with 1634 (25.7%) DAs meeting criteria for one mental disorder,
compared to 1673 (26.3%) meeting criteria for at least six mental disorders. Significant
correlations were found for higher number of disorders among female DAs (r = -.09, p
< .001), White DAs (r = .15, p < .001), and younger DAs (r = -.12, p < .001) (Table 3.8).
The most common diagnoses included mood disorders (e.g., major depressive disorder,
bipolar disorder; n = 3872, 60.9%), conduct-related disorders (e.g., conduct disorder,
oppositional defiant disorder; n = 3373, n = 53.1%), and anxiety-related disorders (e.g.,
generalized anxiety disorder, obsessive-compulsive disorder; n = 2468, 38.8%).
A total of 3444 (35.6%) DAs met criteria for at least one substance-related
disorder (Table 3.9). Common disorders included cannabis-related disorders (e.g.,
cannabis abuse disorder, cannabis dependence disorder; n = 2154, 22.3%), drug-related
disorders (e.g., cocaine dependence disorder, opiate dependence disorder, n = 1564,
16.2%), and alcohol-related disorders (e.g., alcohol use disorder, n = 1177; 12.2%).
Regarding timing of substance-related diagnosis, 490 (14.2%) DAs obtained a diagnosis
approximately 317.26 days (SD = 331.18, range = 0.0-1824.0) prior to detention, 5
(0.01%) DAs were diagnosed during detention within 11.80 days (SD =10.18, range =
1.0-23.0) of detention entry, and 2949 (85.62%) DAs received a diagnosis an average of
1515.06 days (SD = 1255.69, range = 0.0-4978.0) following detention release. Across
the sample, DAs averaged 3.62 (SD = 4.79, range = 0.0-43.0) unique substance-related
disorders; 1848 (53.7%) DAs had only one disorder and 395 (11.5%) DAs had at least
four unique disorders. Significant correlations emerged between higher number of
substance-related disorders and female gender (r = -.05, p < .001), White race (r = .17, p
< .001), and younger age (r = .03, p = .005) (Table 3.8).

62
3.4 Prevalence of Pre-Detention Treatment Utilization
Prevalence rates for treatment utilization prior to first detention stay are displayed
in Table 3.10. Overall, 2438 (25.2%) DAs obtained treatment at some point prior to
detention entry, including 2297 (23.8%) DAs who obtained mental health treatment
services, 354 (3.7%) DAs who obtained substance-related treatment services, and 575
(5.9%) DAs who obtained either/both treatment services (i.e., unknown services
pertaining to one of the disorder types or both disorder types). The most common type of
treatment entailed mental health treatment utilization (n = 2297, 23.8%) and the most
common setting for treatment entailed outpatient services (n = 2435, 25.2%). Less than
5% of the sample obtained substance-related services (n = 354, 3.7%), received inpatient
treatment (n = 456, 4.7%), or visited the emergency department for psychiatric reasons (n
= 285, 2.9%) before their first detention stay. As the time frame prior to detention entry
narrowed, prevalence rates for treatment utilization subsequently decreased. Only 918
(9.5%) DAs obtained treatment services within the month prior to detention; again, the
most common type of treatment involved mental health services (n = 797, 8.2%) and
most common setting was outpatient treatment (n = 873, 9.0%).
As shown in Table 3.5, prevalence rates for pre-detention treatment utilization
were significantly different for gender (χ2 = 199.64, p < .001) and between the three age
cohorts (χ2 = 56.26, p < .001). However, race was not significant; the prevalence of predetention treatment among White DAs (n = 841, 25.0%) and Black DAs (n = 1396,
24.6%) was similar. With regards to treatment type, being male was significantly related
to pre-detention mental health treatment utilization (r = .14, p < .001) and pre-detention
substance-related treatment utilization (r = .06, p < .001) (Table 3.11). White race was
significantly related to pre-detention substance-related treatment utilization (r = .07, p
< .001) and pre-detention either/both treatment utilization (r = .09, p < .001). Older age
was significantly related to pre-detention mental health treatment utilization (r = .05, p
< .001), pre-detention substance-related treatment utilization (r = -.13, p < .001, and predetention either/both treatment utilization (r = .07, p < .001).
Once placed in detention, approximately 323 (3.3%) DAs utilized services,
namely outpatient sessions (n = 315, 3.3%) obtained within the detention center (see

63
Table 3.10). A very small number of youth (n = 16, 0.2%) appear to have been
temporarily released from detention to obtain non-outpatient services. As shown in the
correlations matrix of Table 3.11, overall treatment utilization during detention was more
common among males (r = .05, p < .001), although gender was non-significant for
utilization of any specific treatment type. There was a significant relationship between
Black race and use of both/either treatment utilization during detention (r = -.78, p
= .003), as well as significant relationships between age and general treatment utilization
during detention (r = -.03, p = .004) and mental health treatment utilization during
detention (r = -.04, p < .001).

3.5 Aim 1. Prevalence of Post-Detention Treatment Utilization
Total sample. Following release from detention, a total of 6437 (66.6%) DAs in
the sample utilized treatment at least once during the study time frame (Table 3.12). With
regards to treatment type, overall prevalence rates (pr) were 56.9% (n = 5496) for mental
health treatment utilization, 26.6% (n = 2573) for substance-related treatment utilization,
and 17.1% (n = 1652) for either/both types of treatment utilization. It should be noted
that either/both treatment refers to unknown treatment type pertaining to either and/or
both disorder types. In addition, numerous DAs (n = 2440, 25.2%) obtained more than
one type of treatment and were therefore included in more than one estimate of
prevalence rates. Regarding treatment type, overall prevalence rates were 64.5% (n =
6234) for outpatient treatment utilization, 14.5% (n = 1398) for inpatient treatment
utilization, and 19.6% (n =1891) for emergency department (ED) visits. Given the small
number of inpatient and ED visit users, these two treatment settings were collapsed into
one estimate, resulting in an overall prevalence rate of 26.0% (n = 2515) for nonoutpatient treatment utilization. The remaining results pertaining to non-outpatient
services include both inpatient treatment and ED visits combined, unless otherwise
specified.
Over time, as follow-up time periods decreased in length, the prevalence rates for
treatment utilization also decreased (Table 3.12). Prevalence rates for treatment service
utilization among DAs were 36.7% (n = 3269; 36.0% for outpatient; 8.0% for non-

64
outpatient) at 2-year follow-up, 33.3% (n = 3002; 32.6% for outpatient, 6.2% for nonoutpatient) at 18-month follow-up, 29.0% (n = 2659; 28.8% for outpatient, 5.6% for nonoutpatient) at 1-year follow-up, 22.1% (n = 2055; 21.7% for outpatient, 2.7% for nonoutpatient) at 6-month follow-up, and 10.2% (n = 971, 9.6% for outpatient, 0.8% for
non-outpatient) at 1-month follow-up. Regardless of follow-up time point, treatment
utilization for mental health services was notably higher than substance-related services
and either/both services for both outpatient and non-outpatient treatment services.
The number of DAs retained for follow-up analyses decreased over time, due to
attrition. Specifically, a total of 749 (7.7%) participants were dropped from analyses
focused on treatment utilization within two years of release from detention because 5
youths (0.03%; 5 males, 1 White DA, 4 Black DAs) died during the follow-up period,
251 youths (2.6%; 142 males, 92 White DAs, 150 Black DAs) were detained/incarcerated
for ≥18 months during follow-up period, and 493 youths (5.1%; 400 males, 110 White
DAs, 330 Black DAs) lacked adequate follow-up data. The remaining sample of 8915
DAs that served as the primary sample for time-restricted analyses of treatment
utilization within two years of detention release consisted of 5515 (61.9%) males, 3400
(38.1%) females, 3164 (37.9%) White DAs, 5185 (62.1%) Black DAs, 2287 (25.7%)
younger DAs, 3568 (40.0%) mid-age DAs, and 3060 (34.3%) older DAs.

3.5.1 Aim 2. Prevalence of Post-Detention Treatment Utilization by Groups
Demographic groups. To answer the research questions posed in Aim 2, postdetention service utilization was carefully examined across groups of DAs. Table 3.13
displays a correlation matrix of relationships between demographic variables and postdetention treatment variables and Table 3.14 displays the specific prevalence rates for
post-detention treatment utilization within two years of release by demographic groups.
As shown in the matrix, being male was positively associated with most post-detention
treatment utilization outcomes. Specifically, male gender was significantly correlated
with utilization of any treatment, all three treatment types (i.e., mental health, substancerelated, and either/both), higher number of treatment services, and outpatient treatment.
The prevalence (pr) of any treatment utilization within two years of detention release

65
among males (n = 2335, pr = 42.3%) was significantly higher than females (n = 944, pr
= 27.6%; χ2 = 200.22, p < .001). Results revealed significantly higher prevalence rates
for treatment utilization among males compared to females for all three treatment types
and outpatient treatment; however, findings were not significantly different for nonoutpatient treatment utilization (χ 2 = 0.08, p = .782). Regarding race, being White was
positively and significantly associated with utilization of treatment across all three
treatment types and both treatment settings; however, correlations between race and
number of services (regardless of type or setting) were not significant. As shown in
Table 3.14, a higher prevalence of White DAs (n = 1221, pr = 38.6%) compared to
Black DAs (n = 1791, pr = 34.5%) utilized treatment services. Prevalence rates were
significantly higher among White DAs than Black DAs for the three treatment types and
both treatment settings (see Table 3.14 for details).
Age was negatively related to post-detention treatment utilization, in that younger
age was significantly associated with utilization and number of services for any
treatment, mental health treatment, substance-related treatment, outpatient treatment, and
non-outpatient treatment (Table 3.13). Age was not significantly correlated with
both/either treatment services. Prevalence rates across the three cohorts for any treatment
utilization were significant (χ2 = 1054.42, p < .001); there were significantly less older
DAs (n = 903, pr = 29.5%) who utilized treatment than mid-age DAs (n = 1419, pr =
39.8%) and younger DAs (n = 947, pr = 41.4%) (Table 3.14). Prevalence rates for any
treatment utilization among younger DAs versus mid-age DAs failed to be significant (χ2
= 5.94, p = .015). Additionally, findings indicated significant differences between age
groups for mental health treatment utilization (χ2 = 212.76, p < .001), either/both
treatment utilization (χ2 = 26.88, p < .001), and outpatient treatment (χ2 = 114.52, p
< .001), but non-significant differences for substance-related treatment utilization (χ2 =
212.76, p = .147) and non-outpatient treatment utilization (χ2 = 11.20 p = .014). Paired
analyses revealed higher prevalence among mid-age DAs than younger DAs (χ2 = 39.93,
p < .001) and older DAs (χ2 = 18.45, p < .001) for either/both treatment utilization, and
lower prevalence among older DAs than younger DAs (χ2 = 40.38, p < .001) and mid-age
DAs (χ2 = 20.74, p < .001) for outpatient treatment utilization.

66
Mental health groups. As shown in the correlation matrix of Table 3.15, most
relationships between mental health variables and post-detention treatment variables were
positive and significant. Specifically, positive screens on the MAYSI-2 and higher
number of MAYSI-2 subscales in the caution range were significantly associated with
treatment utilization and number of treatment services for all treatment types and settings,
excluding use of substance-related services, number of substance-related services, and
number of either/both treatment services. The prevalence rates for treatment utilization
displayed in Table 3.16 indicate that screening positively on the MAYSI-2 was strongly
related to post-detention treatment utilization. DAs with positive screens on the MAYSI2 had significantly higher prevalence rates than DAs with non-positive screens for any
treatment utilization (χ 2 =788.48, p < .001), treatment use for all three treatment types,
and treatment use for both settings. It should be noted that a clear pattern emerged in the
findings. Regardless of type or setting, DAs with positive screens had significantly
higher prevalence estimates for treatment service use than DAs with non-positive screens,
who in turn had significantly higher prevalence estimates than DAs who did not take the
MAYSI-2 (i.e., detained prior to 2006).
All disorder groups were positively and significantly related to treatment
utilization for the three treatment types, the two treatment settings, and number of
treatment services (Table 3.15). It should be noted, however, that many significant
correlations were in the negative direction for non-conduct mental disorders and
substance-related disorders, in that DAs with diagnoses of mental disorders (other than
conduct-related) or substance-related disorders were associated with the lack of postdetention treatment utilization and fewer number of treatment services. When comparing
DAs by disorder type (Table 3.16), comorbid DAs had a significantly higher prevalence
(n = 1427, pr = 55.6%) for any treatment utilization post-detention than DAs with mental
disorders (n = 1614, pr = 49.5%; χ2 = 40.27, p < .001), who in turn had a higher
prevalence of treatment utilization than DAs with substance-related disorders (n = 228,
pr = 37.0%; χ2 = 235.58, p < .001). Results followed a similar pattern for outpatient
treatment utilization (χ2 = 2003.20, p < .001). When focused on treatment type, DAs
with mental disorders did not utilize any substance-related treatment services and very

67
few (n = 5, pr = 0.2%) utilized either/both treatment services. Similarly, DAs with
substance-related disorders did not utilize mental health treatment services or either/both
services. Interestingly, DAs who were comorbid for both types of disorders had
significantly higher prevalence of non-outpatient treatment utilization (n = 473, pr =
18.4%) than DAs with mental disorders (n = 212, pr = 6.5%; χ2 = 10.32, p < .001) and
substance-related disorders (n = 38, pr = 6.2%; χ2 = 64.53, p < .001) (Table 3.16).
The correlation matrix for Table 3.17 shows the associations between predetention treatment services, during detention treatment services, and post-detention
treatment services. As indicated by strong and positive associations, pre-detention
treatment service utilization proved to be a significant factor in post-detention treatment
utilization. Specifically, pre-detention treatment utilization, regardless of type, setting, or
number, was significantly and positively related to post-detention treatment utilization,
regardless of type, setting, or number. The only non-significant correlations entailed the
association between pre-detention substance-related treatment utilization and number of
post-detention mental health treatment services (r = .01, p = .180), and between number
of pre-detention substance-related services and number of post-detention mental health
treatment services (r = .02, p = .710). As shown in Table 3.16, prevalence rates for
treatment utilization among DAs who had pre-detention services were significantly and
substantially higher than DAs with no prior treatment (χ2 = 1477.71, p < .001). Results
were consistent across the three treatment types and both treatment settings.
Criminal history groups. Regarding the relationships between post-detention
treatment utilization and criminal history variables, results were mixed. As shown via the
correlation matrix of Table 3.18, correlations were all non-significant between treatment
utilization and the following criminal history variables: number of criminal charges,
charge severity, and total number of detentions (following first detention). Age at first
contact with the juvenile justice system was negatively related to treatment utilization, in
that younger age was significantly associated with utilization and number of postdetention treatment services, excluding either/both treatment services and non-outpatient
treatment services. The number of referrals prior to first detention was significantly and
positively related to post-detention treatment utilization and number of services for all

68
treatment types and treatment settings. Similarly, length of stay, being a violent offender,
recidivism, re-detention, and incarceration were all positively associated with postdetention treatment utilization and number of treatment services.
Prevalence rates for treatment utilization across criminal history groups are shown
in Table 3.19. Overall, DAs with more severe criminal history backgrounds tended to
have higher rates of treatment utilization post-detention. Specifically, prevalence rates
were significantly higher for violent offenders than non-violent offenders (χ2 = 24.22, p
< .001), recidivists than non-recidivists (χ2 = 383.37, p < .001), and redetained/incarcerated DAs than non-incarcerated DAs (χ2 = 160.10, p < .001). In
general, DAs with more severe criminal history had higher treatment utilization rates for
all treatment types and settings, with the exception of a significantly lower prevalence of
substance-related treatment utilization among violent offenders than non-violent
offenders (χ2 = 25.71, p < .001). With regards to timing of post-detention criminal
activity and treatment utilization, 520 (16.2%) recidivists were re-arrested before they
obtained treatment, compared to 2687 (82.8%) recidivists who obtained treatment before
being re-arrested (χ2 = 982.84 p < .001). A total of 1088 (33.9%) DAs who were redetained or incarcerated obtained treatment before their placement in detention or prison,
compared to 2119 (66.1%) DAs who obtained treatment before being redetained/incarcerated (χ2 = 205.67, p < .001).
Cohort groups. Results for post-detention treatment utilization showed marked
divergence between DAs in cohort one (i.e., 1998-2005) versus DAs in cohort two (i.e.,
2006-2011). As displayed in Table 3.20, prevalence rates were significantly higher for
cohort two (n = 1334, 57.0%) compared to cohort one (n = 1935, pr = 29.4%; χ2 =
564.21, p < .001) for any treatment utilization within two years post-detention. DAs in
cohort two had higher prevalence rates than DAs in cohort one for all three treatment
types and both treatment settings.

3.6 Aim 1. Quantity of Post-Detention Treatment Services
To address research question two within Aim 1, Table 3.21 displays the number
of treatment services used by the total sample and demographic groups. Adolescents who

69
utilized treatment within two years post-detention (n = 3269) averaged a total of 19.70
(SD = 30.05, range = 1.0-369.0) unique treatment services. These DAs utilized a
significantly higher number of mental health services than either/both services (t(1) =
28.28, p < .001) and substance-related services (t(1) = 27.20, p < .001). Not surprisingly,
the mean number of outpatient sessions (M = 19.87, SD = 30.35, range = 1.0-369.0) was
substantially larger than the mean number of non-outpatient services (M = 2.06, SD =
2.26, range = 1.0-20.0; t(1) = 32.51, p < .001). During the two years following detention
release, participants obtained an average of 1.31 (SD = 1.81, range = 0.4-30.42)
individual outpatient sessions per month. Comparisons across outpatient type indicate
DAs utilized significantly more mental health outpatient sessions than substance-related
services (t(1) = 45.91, p < .001); participants also utilized significantly more either/both
sessions than substance-related services (t(1) = 21.56, p < .001). The difference in mean
sessions between mental health and either/both outpatient services was not significant
(t(1) = 1.71, p = .192).

3.6.1 Aim 2. Quantity of Post-Detention Treatment Services by Groups
Demographic groups. As shown in Table 3.21, male DAs utilized significantly
more substance-related services (t(1) = 8.06, p = .002) and either/both services (t(1) =
6.63, p = .01) than female DAs; Black DAs utilized more substance-related services (t(1)
= 8.86, p = .003) than White DAs. Comparisons between gender and racial groups for
mean number of outpatient sessions per month were not significant. For age, results were
significant for overall treatment services (F(2, 8912) = 64.17, p < .001), mental health
treatment services (F(2, 8912) = 46.25, p < .001), and outpatient services (F (2, 8912) =
62.33, p < .001). Older DAs obtained significantly less treatment services than mid-age
DAs (t(1) = 39.29, p < .001) and younger DAs (t(1) = 17.38, p < .001), less mental health
treatment services than mid-age DAs (t(1) = 102.23, p < .001) and younger DAs (t(1) =
100.49 p < .001), and less outpatient services than mid-age DAs (t(1) = 52.66, p < .001)
and younger DAs (t(1) = 84.02, p < .001). Younger DAs also utilized significantly more
outpatient sessions per month than mid-age DAs (t(1) = 22.55, p < .001) and older DAs

70
(t(1) = 15.74, p < .001), as well as more mental health sessions per month than mid-age
DAs (t(1) = 24.42, p < .001) and older DAs (t(1) = 13.65, p < .001).
Mental health groups. ANOVA tests revealed significant differences in mean
number of services across MAYSI-2 groups for overall treatment services (F(2, 3266) =
12.55, p < .001), mental health treatment services (F(2, 3266) = 6.64, p = .001), and
outpatient services (F(2, 3266) = 12.30, p < .001) (Table 3.22). Results were also
significant for all analyses pertaining to outpatient sessions per month, with the exception
of non-significant findings for substance-related outpatient sessions (F(2, 3266) = 3.22, p
= .041). When overall ANOVA findings were significant, a clear pattern emerged in
which DAs with positive screens on the MAYSI-2 had the highest mean number of
services, DAs with non-positive screens had the second highest mean, and DAs who did
not complete the MAYSI-2 had the lowest mean number of services. Number of postdetention services differed significantly across the three disorder groups for any treatment
utilization (F(2, 3266) = 30.49, p < .001), as well as all three treatment types and two
treatment settings (see Table 3.22). DAs with substance-related disorders obtained
significantly fewer outpatient sessions per month than DAs with mental disorders (t(1) =
39.10, p < .001) and DAs comorbid for both types of disorders (t(1) = 13.85, p < .001).
Results were also significant for mental health outpatient sessions (F(2, 2334 = 28.12, p
< .001) and either/both outpatient sessions per month (F(2, 2334) = 125.44, p < .001).
DAs with a history of pre-detention treatment services averaged significantly higher
number of services than DAs without pre-detention treatment for overall number of
treatment services (t(1) = 114.48 p < .001), mental health treatment services (t(1) = 69.91,
p < .001), both/either treatment services (t(1) = 34.20, p < .001), and outpatient services
(t(1) = 113.34, p < .001). DAs with prior treatment also averaged higher outpatient
sessions per month (t(1) = 168.58, p < .001) and substance-related sessions per month
(t(1) = 22.98, p < .001) than DAs with no prior treatment.
Criminal history groups. The mean number of services used by violent versus
non-violent offenders within two years post-detention failed to significantly differ for any
treatment type or treatment setting (see Table 3.23). The only significant finding
involved higher mental health outpatient sessions per month among non-violent offenders

71
than violent offenders (t(1) = 8.52, p = .004). In contrast, recidivists obtained
significantly more treatment services than non-recidivists in terms of overall treatment
services (t(1) = 26.90, p < .001), mental health treatment services (t(1) = 70.00, p < .001),
and outpatient services (t(1) = 26.36, p < .001). However, the mean number of outpatient
sessions per month did not differ significantly between these groups. Finally, DAs who
were re-detained/incarcerated obtained significantly more overall treatment services (t(1)
= 17.79, p < .001), mental health treatment services (t(1) = 6.92, p < .001), and outpatient
services (t(1) = 17.70, p < .001) than DAs with no additional detentions or incarcerations.
Similar to recidivism groups, the mean number of outpatient sessions per month failed to
differ significantly between detention/incarceration groups.
Cohort groups. As shown in Table 3.24, DAs within cohort two obtained
significantly more treatment services than DAs within cohort one. Specifically, cohort
two DAs obtained higher number of substance-related treatment services (t(1) = 13.99, p
< .001), either/both treatment services (t(1) = 11.23, p < .001), outpatient sessions per
month (t(1) = 12.85, p < .001), mental health outpatient sessions per month (t(1) = 24.85,
p < .001), and either/both outpatient sessions per month (t(1) = 11.36, p = .001) than
cohort one DAs.

3.7 Aim 1. Frequency of Post-Detention Treatment Services
Table 3.25 displays results for research question two within Aim 1 concerning
mean frequency (in days) between treatment services and mean length of stay for
inpatient treatment services. On average, the 3207 DAs who utilized outpatient treatment
within two years of detention release attended treatment every 33.92 days (SD = 54.93,
range = 1.0-680.0). Mean days between different types of outpatient sessions were not
significantly different (F(2, 2786) = 0.50, p = .609), with treatment users experiencing
similar frequency between mental health treatment (M = 33.56, SD = 53.37, range = 1.0680.0), substance-related treatment, (M = 34.43, SD = 56.56, range = 1.0-593.0), and
either/both services (M = 38.08, SD = 72.61, range = 1.0-471.0). Mean time (in days)
between non-outpatient services was much larger than outpatient services. DAs averaged
utilization of non-outpatient treatment services every 117.69 days (SD = 128.13, range =

72
0.0-723.0). Time between non-outpatient substance-related services was significantly
longer than time between mental health services (t(1) = 10.29, p < .001) and either/both
services (t(1) =5.93, p < .001). The average days between mental health services were
also significantly longer than average days between either/both services (t(1) = 8.83, p
< .001). Finally, average length of day (in days) for DAs (n = 373) who utilized inpatient
treatment within two years of release from detention was 11.27 days (SD = 23.36, range
= 1.0-496.0), with no significant differences in lengths of stay across the three treatment
types (F(2, 370) = 0.10, p = .908).

3.7.1 Aim 2. Frequency of Post-Detention Treatment Services by Groups
Demographic groups. Paired t-tests revealed no significant differences in mean
time between services or length of inpatient stay for gender and race (see Table 3.26).
Results remained non-significant across all treatment types and treatment settings.
Similarly, overall ANOVA tests indicated no significant differences between the three
age cohorts regarding mean time between treatment sessions or average length of
inpatient stay, although paired t-tests revealed some significant differences between
specific age cohorts.
Mental health groups. As displayed in Table 3.27, mean days between outpatient
sessions did not differ significantly across MAYSI-2 groups. Results showed nonsignificant differences between groups for time between treatment services and mean
length of stay, although some MAYSI-2 groups were significantly different when directly
compared using t-tests. Because certain disorder groups did not receive treatment (i.e.,
mean days between services = 0), overall ANOVA tests showed significant differences
between disorder groups for average days between mental health outpatient sessions,
substance-related outpatient sessions, and either/both outpatient sessions. However,
paired t-tests between disorder groups that actually obtained services were not significant
for mental health outpatient sessions (t(1) = 0.07, p = .81), substance-related outpatient
sessions (t(1) = 2.21, p = .138), or either/both outpatient sessions (t(1) = .03, p = .87).
Results were non-significant between disorder groups for overall length of stay (F(2,
385) = .76, p = .471), but significant for specific treatment types due to certain disorder

73
groups not obtaining any services (i.e., mean length of stay = 0 days). When restricting
analyses to only disorder groups who utilized inpatient treatment, results were nonsignificant for lengths of stay for mental health inpatient treatment (t(1) = .96, p = .328)
and substance-related inpatient treatment (t(1) = 4.86, p = .037). Finally, as displayed in
Table 3.27, DAs who obtained treatment prior to detention did not significantly differ
from DAs without prior treatment regarding average time between treatment services or
average length of stay.
Criminal history groups. Overall, most findings for days between treatment
services and lengths of stay were non-significant across criminal history groups (see
Table 3.28). Violent offenders and non-violent offenders experienced similar time
between outpatient treatment services and non-outpatient treatment services, as well as
similar lengths of stay for inpatient treatment. Recidivist DAs and non-recidivist DAs
also had similar days between treatment services and inpatient lengths of stay, with the
exception of significantly less days between non-outpatient treatment services for
recidivists versus non-recidivists (t(1) = 13.43, p < .001). Finally, compared to DAs with
no additional detentions/incarcerations, DAs who were re-detained/incarcerated averaged
similar time (in days) between outpatient sessions and lengths of stay for inpatient
treatment. However, re-detained/incarcerated DAs experienced significantly more days
than other DAs between non-outpatient services (t(1) = 96.67, p = < .001).
Cohort groups. As shown in Table 3.29, no significant differences emerged
between cohort one and cohort two regarding mean time (in days) between treatment
services and mean length of inpatient stay. Results were non-significant for all three
treatment types and both treatment settings.

3.8 Aim 1. Intensity of Post-Detention Treatment Services
Results for research question three within Aim 1 are listed in Table 3.30.
Altogether, 2084 (23.4%) DAs utilized only mental health treatment services, 370 (4.2%)
utilized only substance related services, and 58 (0.7%) DAs utilized either/both treatment
services within two years of reentry into the community. The remaining treatment users
obtained a combination of two (n = 575, 6.5%) or three (n = 200, 2.2%) types of

74
treatment. With regards to treatment setting, the majority of treatment users (n = 2554,
28.6%) obtained only outpatient treatment, 70 (0.8%) DAs obtained only non-outpatient
treatment, and the remaining treatment users (n = 653, 7.3%) obtained both outpatient
and non-outpatient services. After coding for intensity of outpatient user, results
classified 700 (7.9%) DAs as low users with 1-2 outpatient sessions, 771 (8.6%) DAs as
low-to-moderate users with 3-7 sessions, 446 (5.1%) DAs as moderate users with 8-12
outpatient sessions, and the remaining 1352 DAs (15.2%) as high or extreme users with ≥
13 outpatient sessions. Regarding intensity of non-outpatient treatment, 449 (5.1%) DAs
were low users with one inpatient stay and/or ED visit, 121 (1.4%) DAs were low-tomoderate users with 2 stays/visits, 75 (0.9%) DAs were moderate users with 3-4
stays/visits, and the remaining 66 DAs (7.4%) were high or extreme users with ≥5
stays/visits.

3.8.1 Aim 2. Intensity of Post-Detention Treatment Services by Groups
Demographic groups. Intensity of treatment utilization differed significantly
across gender, race, and age cohorts. As shown in Table 3.31, males were more intense
treatment users than females; a significantly higher proportion of males obtained each
individual treatment type, two or more types of treatment (e.g., mental health and
substance related), and outpatient treatment. In addition, prevalence rates were
significantly higher for males than females for all outpatient user intensity levels (e.g., 12 sessions, 3-7 sessions). Conversely, results were non-significant regarding proportion
of non-outpatient users, users of two or three treatment settings (e.g., outpatient and ED
visits), and non-outpatient user levels (e.g., 1 stay/visit, 2 stays/visits.). Regarding race,
White DAs were more intense treatment users than Black DAs. The proportion of Black
DAs were significantly smaller than the proportion of White DAs who utilized two
treatment types (i.e., mental health and substance related treatment services; χ2 = 11.15, p
= .001), only non-outpatient treatment services (χ2 = 26.26, p < .001), two treatment
settings (χ2 = 41.44, p = .001), and three treatment settings (χ2 = 27.41, p = .001). When
broken down into user intensity levels, results indicated that the proportion of White DAs
compared to Black DAs was similar for all outpatient user levels, but significantly more

75
White DAs utilized non-outpatient services then Black DAs at all intensity levels.
Regarding age, the proportion of the DAs who utilized only mental health treatment
services (χ2 = 236.54, p < .001), only substance-related treatment services (χ2 = 75.53, p
< .001), and all three treatment services (χ2 = 10.31, p = .006) differed significantly
across age cohorts. Additionally, results were significantly different for treatment
involving outpatient sessions (χ2 = 81.49, p < .001), two settings (χ2 = 9.41, p = .01), and
the following levels of outpatient treatment utilization: low outpatient users (i.e., 1- 2
sessions), and all outpatient user intensity levels except moderate users (i.e., 8-12
sessions). Please see Table 3.31 for full results.
Mental health groups. Significant differences emerged across MAYSI-2 groups
for all individual treatment types, combination of treatment types, and most user intensity
levels (Table 3.32). Specifically, the proportion of DAs with positive screens were
generally larger than other MAYSI-2 groups, with significant differences across groups
for all outpatient user levels and most non-outpatient user levels. The only findings that
failed to be significant involved only non-outpatient treatment (χ2 = 2.95, p = .229) and
extreme non-outpatient treatment users with ≥7 stays/visits (χ2 = 3.67, p = .159). Since
disorder type was strongly tied to type of treatment utilization, some of the analyses
comparing the proportion of treatment users within each treatment type could not be
conducted. With regards to treatment setting, findings showed significant differences
across disorder groups for outpatient treatment utilization (χ2 = 1418.44, p < .001), nonoutpatient treatment utilization (χ2 = 73.48, p < .001), utilization of two treatment settings
(χ2 = 376.55, p < .001), and utilization of three treatment settings (χ2 = 187.19, p < .001).
Results for disorder type were significant for all outpatient user intensity levels and nonoutpatient user levels, with notably higher proportion of comorbid DAs using services
than DAs with one type of disorder. For pre-detention treatment, significant differences
emerged between groups; the prevalence rates for DAs with pre-detention treatment were
significantly higher than DAs with no treatment for all individual treatment types,
combination of treatment types, treatment settings, outpatient user intensity levels, and
user intensity levels for non-outpatient treatment utilization.

76
Criminal history groups. Results for intensity of treatment service utilization
were significantly different across violent offenders groups, recidivist groups, and redetained/ incarcerated groups. As shown in Table 3.33, prevalence rates for DAs who
were violent offenders upon detention entry were significantly higher for mental health
treatment utilization (χ2 = 84.01, p < .001), substance-related treatment utilization (χ2 =
7.04, p < .001), combination of all three treatment types (χ2 = 9.73, p < .001), and
outpatient treatment utilization (χ2 = 21.29, p < .001). For user intensity levels, a
significantly higher proportion of violent offenders than non-violent offenders were
extreme outpatient users (i.e., ≥23 sessions; χ2 = 25.0, p < .001). Significant results
emerged for recidivism, with higher prevalence rates among recidivist DAs than nonrecidivist DAs for all treatment types and treatment settings, except either/both treatment
utilization (χ2 = 0.86, p = .348) and the combination of substance-related and either/both
treatment utilization (χ2 = 3.77, p = .055). Additionally, there were significantly more
recidivist DAs than non-recidivist DAs within all outpatient user intensity levels and
levels of non-outpatient user intensity levels, except for extreme non-outpatient users
(i.e., ≥7 stays/visits; χ2 = .001, p = .99). Finally, findings were significant for redetained/incarcerated DAs versus other DAs for mental health treatment utilization (χ2 =
36.07, p < .001), the combination of mental health and substance-related treatment
utilization (χ2 = 69.03, p < .001), and utilization of all three treatment types (χ2 = 34.68 p
< .001). The proportion of re-detained/incarcerated DAs was significantly higher than
other DAs for outpatient treatment utilization (χ2 = 86.26, p < .001), two treatment
settings (χ2 = 25.23, p < .001), all three treatment settings (χ2 = 17.76, p < .001), as well
as most intensity levels for outpatient and non-outpatient treatment (see Table 3.33).
Cohort groups. Intensity of treatment services for cohort one versus cohort two
are displayed in Table 3.34. Prevalence rates were significantly higher for cohort one
compared to cohort two for users of only outpatient services (χ2 = 564.21, p < .001), users
of both outpatient and inpatient treatment services, all outpatient user intensity levels, and
all non-outpatient user intensity levels except high (i.e., 5-6 visits/stays) and extreme
users (i.e., ≥7 visits/stays).

77
3.9 Aim 1. Treatment Dropouts and Treatment Gaps
Findings pertaining to research question four within Aim 1 are listed in Table
3.35. Out of the 3207 DAs who utilized outpatient treatment within two years postdetention, 721 (22.5%) youth were early dropouts who terminated treatment after 1-3
sessions, including 342 (10.4%) DAs who dropped out of treatment after one session, 230
(7.2%) DAs who dropped out of treatment after two outpatient sessions, and 149 (4.6%)
DAs who dropped out of treatment after three sessions. With regards to treatment type, a
significantly lower percentage of DAs who utilized either/both outpatient treatment
dropped out of treatment (n = 48 of 445, 10.8%) compared to the percentage of DAs who
dropped out of mental health outpatient treatment (n = 512 of 2751, 18.6%; χ2 = 7.49, p
< .001) and the percentage who dropped out of substance-related outpatient treatment (n
= 160 of 833, 19.2%; χ2 = 7.49, p < .001). Potential reasons for treatment dropouts
included 110 (16.4%) DAs who were incarcerated during the same month of dropouts
and 4 (0.6%) DAs who were placed on an inpatient unit during the time of dropouts;
however, there were no clear reasons for the majority of DAs who dropped out of
treatment (n = 605, 84.0%).
For the 2789 DAs who obtained two or more outpatient sessions, approximately
1139 (40.6%) of these adolescents experienced at least one gap between outpatient
sessions (e.g., 46-120 days between sessions of the same treatment type). A total of 700
(25.1%) DAs experienced one gap, 262 (9.4%) DAs experienced 2-3 gaps, and 71 (2.6%)
DAs experienced ≥4 gaps for an average of 1.64 gaps (SD = 1.02, range = 1-7) (see Table
3.35). Potential reasons for gaps between outpatient sessions included 130 (11.5%) DAs
who were incarcerated and 38 (33.6%) DAs who obtained inpatient treatment during the
time of the treatment gaps. Similar to treatment dropouts, the majority of gaps in
outpatient sessions could not be clearly explained (n = 965, 85.2%).

3.9.1 Aim 1. Treatment Dropouts and Treatment Gaps by Groups
Demographic groups. Differences in early termination from outpatient treatment,
as well as gaps between outpatient sessions, across demographic groups are displayed in
Table 3.36. The number of DAs who dropped out of outpatient treatment was not

78
significant for gender (χ2 = .74, p = .209), but was significant for race (χ2 = 8.89, p = .003)
and age cohorts (χ2 = 113.15, p < .001). A pattern emerged in which dropout rates were
larger among Black DAs (n = 425, 23.9%) than White DAs and larger among older DAs
(n = 227, 19.3%) than younger DAs (n = 112, 12.0%). Paired t-tests indicated that the
proportion of dropouts among older DAs was significantly higher than mid-age DAs (χ2 =
94.09, p < .001) and younger DAs (χ2 = 542.28, p < .001); the dropout rate was also
significantly higher among mid-age DAs than younger DAs (χ2 = 212.91, p < .001).
Given the small number of DAs who dropped out of treatment services due to
detention/prison or inpatient stays, results are difficult to interpret. Findings suggest a
significantly higher proportion of females compared to males (χ2 = 15.53, p < .001) may
have dropped out of treatment due to placement in detention or prison.
Regarding gaps in outpatient treatment, findings were non-significant for gender
and race. A significantly larger percentage of younger DAs (n = 402, 47.1%)
experienced at least one gap in outpatient treatment services than mid-age DAs (n = 476,
38.66%; χ2 = 41.28, p < .001) and older DAs (n = 255, 34.05%; χ2 = 451.40, p < .001). In
addition, there were significantly more mid-age DAs with a history of gaps between
outpatient sessions than older DAs (χ2 = 327.84 p < .001). Being incarcerated emerged as
a potentially significant factor to explain outpatient gaps across age cohorts. Specifically,
the proportion of mid-age DAs (n = 67, 14.1%) who were placed in detention/prison
during the same time frame as the treatment gaps was significantly higher than older DAs
(n = 20, 7.8%; χ2 = 0.22, p = .641). The proportion of younger DAs incarcerated (n = 43,
10.7%) when they experienced gaps between outpatient sessions was also significantly
higher than older DAs (χ2 = 28.80 p < .001).
Mental health groups. Dropping out of outpatient treatment services was
significantly related to disorder type and pre-detention treatment utilization, but not
MAYSI-2 groups (Table 3.37). Overall chi-square tests indicated non-significant
differences across the three MAYSI-2 groups (χ2 = 5.46, p = .065), although paired tests
showed significant differences between DAs with positive screens versus DAs who did
not take the MAYSI-2 (χ2 = 3.96, p < .001). Across disorder type groups, the prevalence
of early dropout was significantly higher among DAs with substance-related disorders (n

79
= 12, 47.7%) than DAs with mental disorders (n = 402, 25.0%; χ2 = 69.90, p < .001) and
comorbid DAs (n = 217, 15.7%; χ2 = 188.88, p < .001). A significantly larger proportion
of DAs with mental disorders than comorbid DAs (χ2 = 169. 85, p < .001) also dropped
out of outpatient treatment. When examining potential reasons for treatment dropout, a
significantly higher proportion of comorbid DAs (n = 44, 20.3%) were in
prison/detention at the time of dropout, compared to DAs with mental disorders (n = 47,
11.7%; χ2 = 5.76, p = .007) and substance-related disorders (n = 5, 4.9%; χ2 = 10.89, p
= .001). For pre-detention treatment groups, dropout rates were significantly higher
among adolescents with no history of treatment (n = 474, 27.2%) compared to
adolescents with pre-detention treatment (n = 247, 16.8%; χ2 = 49.18, p < .001); results
for potential reasons to explain outpatient dropouts were all non-significant.
Similar to findings for outpatient dropouts, the prevalence of gaps between
outpatient sessions was significantly different across disorder types and pre-detention
treatment, but not MAYSI-2 groups. As displayed in the bottom portion of Table 3.37,
findings across MAYSI-2 groups for gaps between outpatient sessions were nonsignificant (χ2 = 0.11, p = .946). In contrast, the proportion of DAs with substancerelated disorders that experienced treatment gaps was significantly smaller (n = 41,
24.1%) than DAs with mental disorders (n = 546, 39.4%; χ2 = 101.84, p < .001) and DAs
with both disorders (n = 546, 42.3%; χ2 = 216.15, p < .001). There were no differences
between the prevalence of DAs with mental disorders and comorbid DAs (χ2 = 17.17, p
= .029) for gaps in treatment. Interestingly, the percentage of DAs who experienced gaps
between outpatient sessions was significantly higher among DAs with prior treatment (n
= 636, 46.7%) than DAs without prior treatment (n = 497, 33.5%; χ2 = 46.4, p < .001).
Further, DAs who had pre-detention treatment were significantly more likely to be placed
in prison/detention (χ2 = 9.95, p = .002) and significantly less likely to be placed in
inpatient treatment (χ2 = 12.37, p = .001) during the time of treatment gaps.
Criminal history groups. As indicated in Table 3.38, dropping out of outpatient
treatment was not associated with violent offender status; violent and non-violent
offenders experienced similar rates of early dropout (χ2 = 5.26, p = .022). In contrast,
dropout rates were significantly lower among recidivists (n = 320, 19.2%) than non-

80
recidivists (n = 401, 26.1%; χ2 = 22.04, p < .001) and re-detained/incarcerated DAs (n =
413, 20.2%) than other DAs (n = 308, 26.5%; χ2 = 16.61, p < .001). Early dropouts may
be partially explained by the significantly larger proportion of recidivists than nonrecidivists (χ2 = 37.72, p < .001) and re-incarcerated DAs than not re-incarcerated DAs
(χ2 = 36.8, p < .001) who were placed in prison/detention during the same month of
treatment dropouts. As shown at the bottom of Table 3.38, prevalence rates for gaps
between outpatient sessions were not significantly different for violent offender groups,
recidivist groups, or re-detained/incarcerated groups. However, sub-analyses suggested
that treatment gaps may be partially explained by differences in the number of DAs who
were placed in prison/detention during the time of treatment gaps.
Cohort groups. Dropout rates from outpatient services were similar for cohort
one (n = 291, pr = 22.3%) and cohort two (n = 429, pr = 22.6%; χ2 = .06, p = .830), with
non-significant differences between timing of dropouts (i.e., after 1 session, after 2
sessions) or reasons for dropouts (Table 3.39). Similarly, differences between cohort one
and cohort two were non-significant for gaps between outpatient sessions (χ2 = 1.73, p
= .629), although there were more DAs from cohort one than cohort two who experienced
gaps between mental health outpatient services (χ2 = 8.48, p < .001) compared to more
DAs from cohort two than cohort one who experienced gaps in either/both outpatient
services (χ2 = 12.25, p < .001).
Finally, Tables 3.40-3.41 summarize results discussed thus far and denote
significant differences in post-detention treatment utilization within two years of
detention release across all demographic, mental health, criminal history, and cohort
groups.

3.10

Aim 3. Predictors of Treatment Utilization

Multicollinearity. Prior to running regression analyses, independent predictor
variables were examined for multicollinearity and outlier values. All outlier cases were
eliminated from analyses. Due to a strong correlation between age at detention entry and
age at first contact with juvenile justice system (r = .76, p < .001), age at first contact was
not included as a predictor variable. In addition, values for the variance inflation factors

81
(VIF) indicated significant multicollinearity between several disorder variables and predetention treatment utilization variables; thus, the number of variables included in stage
three was reduced to the following predictors: conduct-related disorder (i.e., conduct
disorder, oppositional defiant disorder; yes/no), non-conduct mental disorder (yes/no),
substance-related disorder (yes/no), number of disorders, pre-detention outpatient
treatment utilization (yes/no) and pre-detention non-outpatient treatment utilization
(yes/no). As shown Table 3.42, the remaining predictors that were included in the
logistic regression analyses satisfied established standards for multicollinearity (i.e.,
Inflation Factor ≤ 5.0; Tolerance ≥ .02).
Summary statistics. Table 3.43 displays the model summary statistics for the six
main hierarchical logistic regression analyses predicting post-detention treatment
utilization within two years of detention release. For the prediction of any treatment
utilization, chi-square tests revealed significant findings for all three stages of the
regression model. The -2 log likelihood value decreased with each stage, while the
Nagelkerke R2 statistic increased, indicating that the addition of predictor variables at
each stage improved the model and increased accuracy (i.e., classification percentage) for
predicting treatment utilization. Results for the other regression models predicting
utilization of different treatment types and treatment settings also yielded goodness of fit
statistics that showed the models improved with the addition of predictors at each
subsequent stage, in that the -2 log likelihood statistic decreased, the Nagelkerke R2
statistic increased, and the classification accuracy percentage increased. Overall, the
addition of mental health predictor variables in stage three appeared to make the most
notable impact on the regression models, as evidenced by large increases in the
Nagelkerke R2 statistics and classification accuracy percentages for the models.
However, the accuracy percentage showed a slow increase for substance-related
treatment utilization, either/both treatment utilization, and non-outpatient treatment
utilization; adding criminal history and/or mental health variables failed to notably
increase classification accuracy, which suggest that demographic variables may be
sufficient to predict post-detention utilization of these treatment services

82
As shown in Table 3.43, the final stage of the regression model predicting any
treatment utilization yielded a Nagelkerke R2 value of .476, meaning the sum of all
predictor variables improved the model by 47.6%, compared to a null model with no
predictors. With an overall classification accuracy of 80.0%, the regression model was
able to accurately predict 80.0% of DAs as treatment users versus non-users within two
years post-detention, based on the predictor variables included in the model. With
regards to different types of treatment, the regression models were able to accurately
classify 82.2% of DAs as users versus non-users of post-detention mental health
treatment services, 89.9% of DAs as users versus non-users of post-detention substancerelated treatment services, and 94.6% of DA as users versus non-users of either/both
treatment services. Similarly, the regression models demonstrated good predictive
accuracy for treatment settings, with 80.0% of DAs correctly classified as users versus
nonusers of post-detention outpatient treatment services and 92.5% of DAs classified as
users versus nonusers of non-outpatient treatment services.
Odds ratios. As noted in the prior paragraph, the overall set of predictors
produced relatively strong models with good classification accuracy for predicting
treatment utilization within two years post-detention. When examining predictors
individually, findings identified several variables that significantly increased or decreased
the likelihood of post-detention treatment. Tables 3.44-3.46 display the odds ratios and
associated statistics for all predictors at the final stage of the regression analyses. For any
treatment utilization, male gender (OR = 1.70, CI = 1.50-1.94), higher number of prior
arrests (OR = 1.09, CI = 1.06-1.12), higher charge severity (OR = 1.09, CI = 1.03-1.16),
being a violent offender (OR = 1.18, CI = 1.04-1.34), having a conduct disorder (OR =
7.18, CI = 2.41-22.97), having a non-conduct mental disorder (OR = 4.88, CI = 4.075.82), having a substance-related disorder (OR = 2.21, CI = 1.93-2.53), obtaining predetention outpatient treatment (OR = 1.02, CI = 1.00-1.21), and obtaining pre-detention
non-outpatient treatment (OR = 1.43, CI = 1.20-1.71) significantly increased the
likelihood of treatment utilization (see Table 3.44 for full results). Race and number of
disorders were non-significant predictors, whereas having a mental and/or substancerelated disorder emerged as the strongest predictors of treatment utilization.

83
As shown in Table 3.44, the following predictors were significantly associated
with an increase in the likelihood of post-detention mental health treatment utilization:
being male (OR = 1.44, CI = 1.26-1.65), higher number of prior arrests (OR = 1.05, CI =
1.02-1.08), having a conduct disorder (OR = 2.08, CI = 1.07-12.08), having a mental
disorder (OR = 1.01, CI = 1.00-2.02), larger number of disorders (OR = 1.05, CI = 1.031.06), and pre-detention outpatient treatment (OR = 1.11, CI = 1.01-1.22). In contrast,
Black race (OR = 0.91, CI = 0.82-1.00), older age (OR = 0.94, CI = 0.90-0.97), and
having a substance-related disorder (OR = 0.63, CI = 0.54-0.72) were significantly
associated with a decreased likelihood of mental health treatment utilization.
For substance-related treatment utilization (Table 3.45), DAs who were male (OR
= 2.07, CI = 1.69-2.54), older at detention entry (OR = 1.33, CI = 1.15-1.59), had more
prior arrests (OR = 1.08, CI = 1.05-1.12), diagnosed with a conduct disorder (OR = 1.41,
CI = 1.09-1.83), diagnosed with a non-conduct mental disorder (OR = 1. 99, CI = 1.482.62), and diagnosed with a substance-related disorder (OR = 2.02, CI = 1.74-2.92) had a
significantly increased likelihood of obtaining substance-related treatment services within
two years post-detention. Interestingly, number of disorders was associated with a
significantly decreased likelihood of substance-related treatment utilization (OR = 0.96,
CI = 0.94-0.98). Finally, for both/either treatment utilization, findings indicated that DAs
who were male (OR = 1.92, CI = 1. 48-2.51), older at detention entry (OR = 1.15, CI =
1.08-1.24), diagnosed with a conduct disorder (OR = 1.02, CI = 1.01-1.92), and
diagnosed with a substance-related disorder (OR = 4.46, CI = 1.14-12.88) were
significantly more likely to obtain both/either treatment services within two years postdetention. Similar to substance-related treatment, pre-detention treatment utilization did
not significantly impact the likelihood of post-detention treatment utilization.
As shown in Table 3.46, the following variables were significantly associated
with an increased likelihood of post-detention outpatient utilization: male gender (OR =
1.48, CI = 1.48-1.92), higher number of prior arrests (OR = 1.09, CI = 1.05-1.12), higher
charge severity (OR = 1.09, CI = 1.02-1.16), being a violent offender (OR = 1.21, CI =
1.06-1.37), having a conduct-related disorder (OR = 7.05, CI = 2.26-13.85), having a
mental disorder (OR = 4.98, CI = 4.14-6.00), having a substance-related disorder (OR

84
= .98, CI = 1.73-2.28), pre-detention outpatient treatment (OR = 1.05, CI = 1.01-1.22),
and pre-detention non-outpatient treatment (OR = 1.26, CI = 1.08-1.48). None of the
predictors significantly decreased the likelihood of outpatient treatment. Finally, higher
number of prior arrests (OR = 1.06, CI = 1.03-1.14), conduct disorder (OR = 5.11, CI =
3.44-7.59), other mental disorder (OR = 1.96, CI = 1.30-2.95), substance-related disorder
(OR = 2.05, CI = 1.67-2.53), number of disorders (OR = 1.14, CI = 1.11-1.20), and predetention non-outpatient treatment (OR = 1.24, CI = 1.09-1.40) were significantly
associated with an increased likelihood of post-detention non-outpatient treatment
utilization within two years post-detention. Black DAs faced a significantly decreased
likelihood of non-outpatient treatment utilization (OR = 0.69, CI = 0.52-0.76).

3.10.1 Aim 3. Predictors of Treatment Utilization by Cohorts
Cohort one. Given that the MAYSI-2 and RAI were implemented in 2006,
additional regression analyses were conducted to examine a possible time effect for
cohort one (i.e., detained during 1998-2005) versus cohort two (i.e., detained during
2006-2011). As shown in Table 3.47, regression models predicting post-detention
treatment use among cohort one showed models were significantly different (p < .01)
then the null model with no predictors. Similar to the main regression analyses for the
total sample, model summary statistics indicated better fitting models with each
subsequent stage; specifically the -2 log likelihood decreased, while the Nagelkerke R2
values and classification percentages increased. In comparing results for the overall
sample versus cohort one, findings indicated slightly better fitting regression models for
cohort one. For example, the Nagelkerke R2 value was .476 and the classification
percentage was 80.0% for predicting treatment utilization within two years post-detention
among the entire sample, whereas the Nagelkerke R2 value was .479 and the
classification percentage was 82.1% for predicting treatment utilization among cohort
one. Overall, findings for cohort one showed improved prediction models, with smaller
log likelihood values, larger Nagelkerke R2 values, and better classification percentages,
with the exception of smaller Nagelkerke R2 statistics for either/both treatment utilization
and non-outpatient treatment utilization.

85
As noted above, the classification rate for treatment utilization among cohort one
was 82.1%, meaning the set of independent predictors entered into the regression model
was able to accurately predict 82.1% of DAs within cohort one as treatment users versus
non-users within two years post-detention. Additionally, the logistic regression models
for cohort one were able to accurately classify 84.4% of cohort one DAs as users versus
non-users of post-detention mental health treatment services, 92.8% of cohort one DAs as
users versus non-users of post-detention substance-related treatment services, and 96.8%
of cohort one DA as users versus non-users of either/both treatment services (Table 3.47).
With regards to treatment setting, 82.2% of cohort one DAs were correctly classified as
users versus non-users of post-detention outpatient treatment services and 93.9% of
cohort one DAs were accurately classified as users versus non-users of non-outpatient
treatment services. Altogether, findings suggest that the total set of independent variables
may be better suited to predict treatment utilization among DAs detained between 1998
and 2005 than DAs detained across the entire study time frame.
Logistic regression results for cohort one identified a slightly different set of
significant predictors of treatment utilization (Table 3.48). With regards to treatment
utilization among cohort one, being male (OR = 1.72, CI = 1.48-2.00), higher number of
prior arrests (OR = 1.13, CI = 1.09-1.18), higher charge severity (OR = 1.09, CI = 1.011.08), being a violent offender (OR = 1.21, CI = 1.04-1.41), having a conduct disorder
(OR = 9.74, CI = 1.14-20.73), having a mental disorder (OR = 5.54, CI = 4.42-6.95),
having a substance-related disorder (OR = 1.77, CI = 1.50-2.09), obtaining pre-detention
outpatient treatment (OR = 1.09, CI = 1.06-1.12), and obtaining pre-detention nonoutpatient treatment (OR = 2.07, CI = 1.37-3.12) were significantly associated with an
increased likelihood of treatment utilization (Table 3.49). Contrary to the main analyses
for the total sample (Table 3.48), age was non-significant and number of disorders (OR =
0.97, CI = 0.95-0.99) emerged as a significant variable associated with decreased
likelihood of treatment utilization. The same predictors were associated with a
significantly increased likelihood of post-detention mental health treatment. However,
number of disorders was no longer significant, while age (OR = 0.90, CI = 0.86-0.94) and

86
having a substance-related disorder (OR = 0.65, CI = 0.55-0.79) emerged as variables
associated with a significantly decreased likelihood of mental health treatment utilization.
As shown in Table 3.50, the following predictors were associated with an
increased likelihood of substance-related treatment utilization among cohort one: male
gender (OR = 1.92, CI = 1.49-2.48), older age (OR = 1.23, CI = 1.14-1.32), number of
prior arrests (OR = 1.17, CI = 1.10-1.24), conduct disorder (OR = 2.00, CI = 1.42-2.79),
substance-related disorder (OR = 4.67, CI = 1.59-7.90), and pre-detention non-outpatient
treatment (OR = 1.81, CI = 1.32-2.50). Variables associated with a decreased likelihood
of substance-related treatment services included being a violent offender (OR = 0.71, CI
= 0.55-0.91), non-conduct mental disorder (OR = 0.44, CI = 0.32-0.62), and larger
number of disorders (OR = 0.96, CI = 0.93-0.99). For both/either treatment utilization,
findings indicated that DAs who were male (OR = 1.87, CI = 1.30-2.67), older at
detention entry (OR = 1.18, CI = 1.07-1.31), diagnosed with a conduct disorder (OR =
3.02, CI = 1.01-5.23), obtained pre-detention non-outpatient treatment (OR = 1.49, CI =
1.08-2.06), and obtained pre-detention outpatient treatment (OR = 1.01, CI = 1.00-1.02)
were significantly more likely to obtain post-detention treatment services. Similar to
other treatment types, race was not significantly associated with increased (or decreased)
likelihood of post-detention treatment utilization.
As shown in Table 3.48, findings for post-detention outpatient treatment among
cohort one were similar to findings from the main regression analyses for the total
sample. Specifically, male gender (OR = 1.73, CI = 1.56-1.98), higher number of prior
arrests (OR = 1.13, CI = 1.08-1.18), higher charge severity (OR = 1.09, CI = 1.01-1.17),
being a violent offender (OR = 1.20, CI = 1.03-1.40), having a conduct-related disorder
(OR = 14.57, CI = 3.69-25.93), having a mental disorder (OR = 5.80, CI = 4.60-7.31),
having a substance-related disorder (OR = 1.62, CI = 1.37-4.91), pre-detention outpatient
treatment (OR = 1.07, CI = 1.01-1.10), and pre-detention non-outpatient treatment (OR =
1.93, CI = 1.31-2.84) were significantly associated with an increased likelihood of
outpatient treatment within two years post-detention. Older age at detention entry (OR =
0.94, CI = 0.91-0.98) and larger number of disorders (OR = 0.98, CI = 0.95-0.99) were
associated with a decreased likelihood of outpatient treatment. Compared to the total

87
sample, the following three variables were identified as newly significant predictors of
outpatient treatment among cohort one: age, number of disorders, and non-conduct
mental disorder (Table 3.48). Finally, the following variables were associated with an
increased likelihood of post-detention non-outpatient treatment utilization: number of
prior arrests (OR = 1.09, CI = 1.03-1.16), conduct disorder (OR = 6.80, CI = 4.16-11.12),
other mental disorder (OR = 2.47, CI = 1.50-4.07), substance-related disorder (OR = 1.69,
CI = 1.31-2.20), number of disorders (OR = 1.13, CI = 1.09-1.15), and pre-detention nonoutpatient treatment (OR = 2.03, CI = 1.52-2.79) (see Table 3.51). Being Black (OR =
0.68, CI = 0.54-0.86) and length of stay (OR = 0.99, CI = 0.98-1.00) were associated with
decreased likelihood of non-outpatient treatment utilization. It should be noted that length
of stay and non-conduct disorder were both significant predictors for cohort one, but nonsignificant for the total sample (see Table 3.48).
Cohort two. Table 3.52 displays the model summary statistics for logistic
regression analyses for cohort two, with the addition of Risk Assessment Inventory (RAI)
scores as a predictor in stage two and positive screen on Massachusetts Youth Screening
Instrument-2nd Edition (MAYSI-2) as a predictor in stage three. Similar to results from
the other regression analyses, goodness of fit indicators showed a better fitting model
with each stage, in that the -2 log likelihood statistics decreased, the Nagelkerke R2
values increased, and the classification accuracy percentage increased with the addition
of predictors. Based on comparisons between findings for the overall sample, cohort one,
and cohort two, adding RAI and MAYSI-2 failed to make a noticeable improvement in
predicting treatment utilization post-detention. With the inclusion of these two
predictors, the Nagelkerke R2 value was .360 and the classification percentage was 80.0%
for treatment utilization within two years post-detention. As shown in Tables 3.43 and
3.47, the Nagelkerke R2 values were .476 and .479 and the classification percentages
were 80.0% and 82.1% for the total sample and cohort one. Thus, the total set of
predictors was slightly less accurate in predicting treatment utilization within two years
post-detention for DAs in cohort one compared to cohort two. With regards to
classification accuracy for the other treatment types and treatment settings, the regression
models for cohort two were able to accurately classify 80.3% of DAs as users versus non-

88
users of post-detention mental health treatment, 84.5% of DAs as users versus non-users
of post-detention substance-related treatment, 88.7% of DAs as users versus non-users of
post-detention either/both treatment, 78.9% of DAs as users versus non-users of postdetention outpatient treatment, and 93.9% of as users versus non-users of post-detention
non-outpatient treatment.
Regression analyses for treatment utilization among cohort two yielded somewhat
similar findings as cohort one; however, several new predictors emerged as significant
and several predictors impacted treatment in the opposite direction (i.e., decreased
likelihood rather than increased likelihood) (Table 3.48). For any treatment utilization
within two years post-detention, length of detention stay (OR = 1.01, CI = 1.00-1.20),
conduct disorder (OR = 5.74, CI = 3.80-8.66), non-conduct mental disorder (OR = 2.86,
CI = 1.91-4.27), substance-related disorder (OR = 2.20, CI = 2.37-4.32), and number of
disorders (OR = 1.29, CI = 1.20-1.39) were associated with a significantly increased
likelihood of treatment (Table 3.53). In contrast to findings for the overall sample (Table
3.48), older age (OR = 0.88, CI = 0.90-0.95), number of charges (OR = 0.79, CI = 0.640.99), and pre-detention non-outpatient treatment (OR = 0.68, CI = 0.55-0.85) were
associated with decreased likelihood of post-detention treatment utilization. As shown in
Table 3.53, significant variables for predicting post-detention mental health treatment
utilization included the following: conduct disorder (OR = 7.37, CI = 1.03-11.98), nonconduct mental disorder (OR = 4.07, CI = 1.02-9.09), and higher number of disorders
(OR = 1.38, CI = 1.28- 1.48). DAs who were older at detention entry (OR = 0.70, CI =
0.71-0.85), had substance-related disorders (OR = 0.45, CI = 0.33-0.62), and obtained
pre-detention non-outpatient treatment (OR = 0.77, CI = 0.62-0.95) were less likely to
obtain mental health treatment services. Contrary to prior analyses for the total sample
and cohort one (Table 3.48), gender, prior arrests, and pre-detention outpatient treatment
failed to remain significant predictors; the direction for pre-detention non-outpatient also
changed from increased to decreased likelihood of treatment services.
Regarding substance-related treatment utilization among cohort two (Table 3.54),
male gender (OR = 1.68, CI = 1.06-2.65), older age (OR = 1.20, CI = 1.08-1.34), and
substance-related disorder (OR = 3.49, CI = 2.94-4.17) were associated with an increased

89
likelihood of treatment, whereas number of charges (OR = 0.71, CI = 0.51-0.98), conduct
disorder (OR = 0.36, CI = 0.21-0.63), non-conduct mental disorder (OR = 0.43, CI =
0.34-.079), and pre-detention non-outpatient treatment (OR = 0.74, CI = 0.63-0.96) were
associated with a decreased likelihood of treatment utilization. As shown in Table 48,
number of charges emerged as a newly significant finding for cohort two, whereas prior
arrests, violent offender, and number of disorders failed to remain significant predictors,
which they were for the total sample and cohort one. Regarding both/either treatment
utilization (Table 3.54), findings indicated that DAs who were male (OR = 1.92, CI =
1.16-3.13), older at detention entry (OR = 1.25, CI = 1.11-1.41), had higher number of
prior arrests (OR = 1.06, CI = 1.01-1.11), diagnosed with a conduct disorder (OR = 2.01,
CI = 1.05-2.23), diagnosed with a substance-related disorder (OR = 5.91, CI = 6.5316.93), and had higher number of disorders (OR = 1.10, CI = 1.01-1.15) were more likely
to utilize services; DAs who obtained pre-detention outpatient treatment (OR = 0.99, CI =
0.98-1.00) or pre-detention non-outpatient treatment (OR = 0.78, CI = 0.63-0.97) were
less likely to utilize either/both services within two years post-detention.
As shown in Table 3.55, significant predictors associated with increased
likelihood for outpatient treatment among cohort two included length of stay (OR = 1.01,
CI = 1.00-1.20), conduct disorder (OR = 5.74, CI = 3.80-8.66), non-conduct mental
disorder (OR = 2.86, CI = 1.91-4.27), substance-related disorder (OR = 3.20, CI = 2.374.32), number of disorders (OR = 1.29, CI = 1.19-1.39); predictors associated with
decreased likelihood included older age (OR = 0.88, CI = 0.80-0.95), number of charges
(OR = 0.79, CI = 0.64-0.99), and pre-detention non-outpatient treatment (OR = 0.68, CI =
0.55-0.85). Interestingly, gender, prior arrests, and pre-detention outpatient treatment
were non-significant for cohort two, although they were significant predictors for the
total sample and cohort one (Table 3.48). Finally, screening positively on the MAYSI-2
(OR = 2.56, CI = 1.61-4.09), having a conduct disorder (OR = 3.94, CI = 1.87-4.33),
having a substance-related disorder (OR = 2.36, CI = 1.61-3.46), and larger number of
disorders (OR = 1.24, CI = 1.29-1.31) were associated with a significantly increased
likelihood of non-outpatient treatment utilization within two years post-detention for
cohort two (Table 3.54). Black DAs were significantly less likely to obtain non-

90
outpatient treatment services than White DAs (OR = 0.62, CI = 0.43-0.89). Altogether,
RAI scores and positive screens on the MAYSI-2 failed to serve as significant predictors
of treatment utilization, except for non-outpatient treatment services.

3.11 Aim 4. Patterns of Treatment Utilization: Time to Treatment
Life table. To address research questions 1-2 within Aim 4, a series of survival
analyses were conducted. The life table shown in Table 3.56 displays the main results for
connection to care (i.e., time until first treatment utilization), divided into 3-month
intervals for the two years following detention release years and 6-month intervals for the
remaining time frame of the study. At the beginning of the study (i.e., date of detention
release), a total of 9664 DAs were eligible for treatment. As shown in the life table, a
total of 1582 (16.4%) DAs utilized treatment within the first 3-month time period,
resulting in approximately 83.6% of the sample surviving as non-users (i.e., denoted as
cumulative proportion surviving in the life table). The probability function was .055 (SD
= .001), meaning participants had a 5.5% chance of utilizing treatment during the first
time interval. The hazard ratio equaled 0.06 (SD = .00), which provides an estimate of
treatment utilization by the end of the time interval, assuming a participant has survived
to the start of the time interval (i.e., has not been censored—dropped out—or already
obtained treatment). After one year post-detention, a total of 6820 DAs remained as nonusers and 2843 DAs utilized their first treatment service; approximately 68% of non-user
DAs were still surviving to utilize their first treatment service. The probability of
treatment during that time interval (i.e., 9-12 months) equaled .008 (SD = .001) and the
hazard rate equaled .01 (SD = .00).
At the time of the final time interval, which occurred at 14-year follow-up, a total
of 216 DAs remained to utilize treatment and 3 DAs obtained their first treatment service
(Table 3.56). After taking into account all the participants who utilized treatment and
participants who were lost to attrition/censoring, approximately 22.0% of the sample
remained to utilize treatment at the end of the study time frame. As shown across the
time intervals of the life table, as the length of time from detention release (i.e., day 0)
increased, the number of treatment users per time interval decreased. However, there

91
were still new treatment users within each individual time interval. Overall,
approximately 1/3 of the sample utilized their first treatment service within 18 months of
release from detention and another 1/3 of the sample utilized treatment between 18
months and the end of the study time frame; the remaining 1/3 of the sample survived as
non-users of treatment services.

3.11.1 Aim 4. Time to Treatment by Groups
Table 3.57 shows the results for the series of Kaplan-Meier survival analyses
conducted to examine differences in time to first treatment across groups of DAs. The
table displays the percentage of treatment users versus non-users (i.e., censored cases),
median (Mdn) days to first treatment utilization, and chi-square tests indicating whether
survival curves depicting patterns of treatment utilization differed between groups. To
aid in interpretation, please refer to Figures 3.1-3.23, which display the survival curves
between comparison groups. Regarding treatment type, the curves for time (in days) to
first substance-related treatment utilization and either/both treatment utilization crossed
over time (Figure 3.2); however, results indicated that the survival curves for all three
treatment types significantly differed at early follow-up (Wilcoxon = 86.65, p < .001),
middle follow-up (Tarone-Ware = 82.882, p < .001), and long-term follow-up of the
study time frame (Log Rank = 68.74, p < .001) (Table 3.57). Paired statistical tests
showed that each type of treatment differed significantly from the other type, with DAs
obtaining mental health treatment services significantly earlier than either/both services
(Wilcoxon = 20.45, p < .001; Tarone-Ware = 23.40, p < .001 Log Rank = 24.70, p
< .001) and substance-related services (Wilcoxon = 100.90, p < .001; Tarone-Ware =
89.10, p < .001; Log Rank = 61.21, p < .001). Regarding treatment setting, treatment
users obtained outpatient treatment services within a median of 439.0 days (CI = 405.11472.89), compared to non-outpatient treatment services within approximately 1528.0
days (CI = 1430.95-1625.05). As shown in Figure 3.3, the survival curve for time to first
non-outpatient treatment diverged significantly from the survival curve for time to first
outpatient treatment service, with significant differences over the entire study time frame
(Table 3.57).

92
Demographic groups. Results for time (in days) to first treatment were
significantly different for gender, race, and age cohorts (Table 3.57; Figures 3.4-3.6).
Specifically, male DAs utilized treatment significantly earlier than female DAs (Figure
3.4), resulting in significantly different survival curves (Wilcoxon = 507.334, p < .001;
Tarone-Ware = 540.08, p < .001; Log Rank = 454.37, p < .001) (Table 3.57). Black DAs
also utilized treatment significantly earlier than White DAs (Figure 3.5), with
significantly different survival curves over time (Wilcoxon = 21.08, p < .001; TaroneWare = 23.60, p < .001; Log Rank = 20.23, p < .001). Time to first treatment was
significantly different between the three age groups, despite some interactions between
the survival curves of age cohorts (Figure 3.6). Mid-age DAs utilized treatment within a
significantly shorter number of days than younger DAs (Wilcoxon = 11.05, p < .001;
Tarone-Ware = 24.15, p < .001, Log Rank = 39.80, p < .001) and older DAs (Wilcoxon =
12.54, p < .001; Tarone-Ware = 19.61, p < .001; Log Rank = 21.98, p < .001). Older
DAs also had a significantly different curve than younger DAs (Wilcoxon = 1.78, p
= .009; Tarone-Ware = 5.71, p < .001; Log Rank = 9.83, p < .001).
Mental health groups. As displayed in Table 3.57, patterns for time (in days) to
first treatment differed significantly across MAYSI-2 groups (Wilcoxon = 1868.98, p
< .001; Tarone-Ware = 2053.05, p < .001, Log Rank = 2213.78, p < .001). As shown in
Figure 3.7, DAs who never completed the MAYSI-2 required a significantly larger
number of days to obtain treatment post-detention than DAs with positive screens on the
MAYSI-2 (Wilcoxon = 1619.75, p < .001; Tarone-Ware = 756.57, p < .001, Log Rank =
1854.06, p < .001) and DAs with non-positive screens on the MAYSI-2 (Wilcoxon =
700.49, p < .001; Tarone-Ware = 756.57, p < .001, Log Rank = 805.91, p < .001). Such
results should be interpreted with caution, however, given that youth who did not
complete the MAYSI-2 were detained during the early portion of the study (i.e., 19982005) and had much longer follow-up periods to utilize treatment, compared to youth
who completed the MAYSI-2. Thus, the length of follow-up in which DAs were able to
obtain their first treatment differed greatly between the MAYSI-2 screened groups (i.e.,
cohort two) and the non-screened group (i.e., cohort one). When restricting analyses to
DAs within cohort two, DAs with positive screens and DAs with non-positive screens on

93
the MAYSI-2 had similarly shaped survival curves that were significantly different in the
long-term (Log Rank = 2.70, p = .01), but not significantly different in the short-term
(Wilcoxon = 5.18, p = .023) or middle of the study time frame (Tarone-Ware = 4.47, p
= .034).
Patterns for time to first treatment utilization were significantly different between
disorder groups throughout the entire time frame of the study (Log rank = 81.91, p
< .001; Wilcoxon = 109.18, p < .001; Tarone-Ware = 102.69, p < .001) (Table 3.57). As
shown in Figure 8, the survival curve for DAs with substance-related disorders was
divergent from the curves for DAs with mental disorders and DAs with both disorders
(Figure 3.8). DAs with only substance-related disorders tended to obtain post-detention
treatment services significantly later than DAs with mental disorders (Wilcoxon = 81.98,
p < .001; Tarone-Ware = 71.15 p < .001, Log Rank = 47.58, p < .001) and those with
both disorders (Wilcoxon = 110.38, p < .001; Tarone-Ware = 105.30, p < .001, Log Rank
= 84.28, p < .001). Further, the curve for the mental disorder group differed significantly
from the curve for the comorbid group (Wilcoxon = 8.15, p < .001; Tarone-Ware = 9.64,
p < .001, Log Rank = 13.33, p < .001). Finally, the Kaplan-Meier survival analyses
showed significant differences in time to first treatment between pre-detention treatment
groups. As displayed in Figure 3.9, DAs who obtained pre-detention treatment utilized
treatment services within a significantly shorter number of days after detention release
than DAs who did not obtain pre-detention treatment (Log rank = 3493.60, p < .001;
Wilcoxon = 3604.13, p < .001; Tarone-Ware = 3617.47, p < .001).
Criminal history groups. As shown in Table 3.57, DAs with more serious
criminal history tended to utilize treatment within a significantly smaller number of days
upon community reentry than other DAs. The survival curve (Figure 3.10) for violent
offenders differed significantly from the survival curve for non-violent offenders across
all study time points (Wilcoxon= 28.82, p < .001; Tarone-Ware = 33.32, p < .001; Log
Rank = 33.92, p < .001). Violent offenders obtained treatment services within
approximately 454.0 days (CI = 493.21-502.97), whereas non-violent offenders did not
obtain treatment services until approximately 580.0 days (CI = 405.02-502.97) postdetention. Similarly, recidivists utilized treatment much sooner than non-recidivists, with

94
significantly different survival curves for the two recidivism groups (Figure 3.11) across
all time points (Wilcoxon= 170.48, p < .001; Tarone-Ware = 182.07, p < .001; Log Rank
= 153.46, p < .001). In contrast, results were non-significant for re-detained/incarcerated
groups; DAs from both groups obtained post-detention treatment within a similar number
of days and survival curves (Figure 3.12) were non-significantly different throughout the
study time frame (see Table 3.57 for full results).

3.11.2 Time to Treatment: Cox Regression Analyses
Table 3.58 displays the results for the Cox proportional hazards regression
analyses predicting risk of first treatment utilization at any time during study follow-up.
Results indicated that the regression models were significantly different then null models
with no predictors, with significant chi-square tests (p ≤ .01) at stage one, stage two, and
stage three of all six main models predicting risk of treatment utilization. Adding
predictors at each stage significantly improved the model, as evidenced by decreased -2
log likelihood values and significant chi-square results for the change statistics (i.e.,
results for addition of predictors, compared to previous stage).
The Cox regression analyses identified several variables with significant hazard
ratios for predicting treatment utilization. As shown in Table 3.59, number of prior
arrests (HR = 1.04, CI = 1.02-1.06), charge severity (HR = 1.04, CI = 1.01-1.07), being a
violent offender (HR = 1.11, CI = 1.05-1.18), being re-detained/incarcerated (HR = 1.19,
CI = 1.12-1.26), having a conduct disorder (HR = 3.03, CI = 2.3-3.24), having a nonconduct mental disorder (HR = 3.92, CI = 3.62-4.23), having a substance-related disorder
(HR = 1.95, CI = 1.83-2.07), and obtaining pre-detention outpatient treatment (HR =
1.51, CI = 1.40-1.64) were all associated with increased risk of utilizing treatment
services at any time during the 14 years of the study time frame. The majority of these
variables also emerged as significant predictors of mental health treatment utilization
post-detention. Specifically, number of prior arrests (HR = 1.03, CI = 1.02-1.05), charge
severity (HR = 1.05, CI = 1.01-1.07), violent offender (HR = 1.09, CI = 1.03-1.15), redetained/incarcerated (HR = 1.19, CI = 1.11-1.27), conduct disorder (HR = 3.46, CI =
3.22-3.72), non-conduct disorder (HR = 9.97, CI = 8.96-11.10), number of disorders (HR

95
= 1.02, CI = 1.01-1.03), pre-detention outpatient treatment (HR = 1.35, CI = 1.24-1.47),
and pre-detention non-outpatient treatment (HR = 1.12, CI = 1.01-1.42 were all
associated with significantly increased risk of treatment services, whereas older age (HR
= 0.93, CI = 0.92-0.95) yielded a significantly decreased risk of mental health treatment
utilization post-detention.
As shown in Table 3.60, variables associated with significantly increased risk of
substance-related treatment at any time post-detention included male gender (HR = 1.22,
CI = 1.11-1.33), larger number of prior arrests (HR = 1.03, CI = 1.02-1.05), getting redetained/incarcerated (HR = 1.20, CI = 1.09-1.31), having a conduct disorder (HR = 1.99,
CI = 1.79-2.01), greater number of disorders (HR = 1.02, CI = 1.01-1.04), and predetention outpatient treatment utilization (HR = 2.34, CI = 2.05-2.65); predictors
associated with decreased risk of services included larger number of charges (HR = 0.94,
CI = 0.88-0.99), having a non-conduct mental disorder (HR = 0.77, CI = 0.70-0.86), and
pre-detention non-outpatient treatment utilization (HR = 0.72, CI = 0.60-0.85).
Interestingly, having a substance-related disorder failed to emerge as a significant
variable of substance-related treatment. For either/both treatment utilization at any time
post-detention, male gender (HR = 1.21, CI = 1.08-1.35), number of prior arrests (HR =
1.05, CI = 1.03-1.06), violent offender (HR = 1.14, CI = 1.02-1.29), redetained/incarcerated (HR = 1.13, CI = 1.00-1.27), conduct disorder (HR = 2.77, CI =
2.43-3.17), non-conduct mental disorder (HR = 4.22, CI = 3.38-5.27), number of
disorders (HR = 1.05, CI = 1.04-1.07), and pre-detention outpatient treatment (HR = 2.53,
CI = 2.15-2.97) were linked to an increased risk of treatment services, whereas predetention non-outpatient treatment had a significant hazard ratio indicating decreased risk
of either/both treatment services (HR = 0.76, CI = 0.63-0.92).
Results for Cox regression analyses predicting first post-detention outpatient
treatment utilization and first post-detention non-outpatient treatment utilization are listed
in Table 3.61. Regarding utilization of outpatient treatment services at any time,
significant variables included age (HR = 0.95, CI = 0.93-0.96), number of pre-detention
arrests (HR = 1.04, CI = 1.03-1.05), charge severity (HR = 1.05, CI = 1.01-1.06), violent
offender (HR = 1.12, CI = 1.06-1.19), re-detained/ incarcerated (HR = 1.18, CI = 1.11-

96
1.25), conduct disorder (HR = 3.12, CI = 2.91-3.34), non-conduct mental disorder (HR =
4.02, CI = 3.71-4.36), substance-related disorder (HR = 1,79, CI = 1.69-1.91), number of
disorders (HR = 1.01, CI = 1.00-1.02), and pre-detention outpatient treatment (HR = 1.54,
CI = 1.42-1.67). Regarding utilization of non-outpatient treatment services at any time
post-detention, the following variables were associated with increased risk of treatment
services: number of prior arrests (HR = 1.03, CI = 1.01-1.05), conduct disorder (HR =
2.06, CI = 1.81-2.34), non-conduct mental disorder (HR = 5.73, CI = 4.77-6.88),
substance-related disorder (HR = 3.32, CI = 2.93-3.76), number of disorders (HR = 1.07,
CI = 1.06-1.08), pre-detention outpatient treatment (HR = 1.48, CI = 1.24-1.77), and predetention non-outpatient treatment (HR = 1.34, CI = 1.09-1.66). In contrast, Black race
(HR = 0.75, CI = 0.67-0.83) and age (HR = 0.93, CI = 0.90-0.96) were associated with
significantly decreased likelihood of non-outpatient treatment services.

3.11.3 Time to Treatment by Cohorts
Cohort one. When restricting the Cox regression analyses to only DAs from
cohort one, results were generally similar to findings for the total sample. As displayed
in Table 3.62, the model summary statistics for the six regression models indicated that
models were significantly different than null models and that entering predictor variables
at each stage produced significantly better fitting models. Compared to the total sample
(Table 3.63), the same predictors significantly impacted the risk of treatment utilization
at any time during the study time frame. Specifically, number of prior arrests (HR = 1.05,
CI = 1.03-1.06), charge severity (HR = 1.05, CI = 1.02-1.08), violent offender (HR =
1.17, CI = 1.10- 1.25), re-detained/incarcerated (HR = 1.14, CI = 1.07-1.22), conduct
disorder (HR = 3.44, CI = 3.18-3.72), non-conduct mental disorder (HR = 5.01, CI =
4.58-5.48), substance-related disorder (HR = 1.88, CI = 1.75-2.02), and pre-detention
outpatient treatment (HR = 2.10, CI = 1.89-2.34) were associated with increased risk of
post-detention treatment utilization; older age at detention entry (HR = 0.96, CI = 0.940.98) was associated with decreased risk of treatment utilization (Table 3.64). For
utilization of mental health treatment services among cohort one at any time upon
community reentry, larger number of prior arrests (HR = 1.04, CI = 1.03-1.06), higher

97
charge severity (HR = 1.04, CI = 1.01-1.08), being a violent offender (HR = 1.15, CI =
1.08-1.24), getting re-detained/incarcerated (HR = 1.15, CI = 1.07-1.24), having a
conduct disorder (HR = 3.83, CI = 3.53-4.16), having a non-conduct mental disorder (HR
= 13.57, CI = 11.98-15.36), larger number of disorders (HR = 1.01, CI = 1.00-1.02), and
obtaining pre-detention outpatient treatment (HR = 1.92, CI = 1.72-2.14) were associated
with increased risk, whereas age was associated with decreased risk (HR = 0.95, CI =
0.93-0.97) of mental health treatment utilization. In contrast to findings for the total
sample, pre-detention non-outpatient treatment failed to remain a significant variable
associated with increased risk of treatment utilization among cohort one (Table 3.63).
As shown in Table 3.65, significant variables associated with increased risk for
substance-related treatment at any time post-detention included male gender (HR = 1.16,
CI = 1.05-1.29), number of prior arrests (HR = 1.05, CI = 1.02-1.08), violent offender
(HR = 1.12, CI = 1.01-1.24), re-detained/incarcerated (HR = 1.14, CI = 1.02-1.27),
conduct disorder (HR = 2.43, CI = 2.15-2.75), substance-related disorder (HR = 1.27, CI
= 1.20-3.74), number of disorders (HR = 1.02, CI = 1.01-1.03), and pre-detention
outpatient treatment (HR = 3.91, CI = 3.28-4.65). DAs with mental disorders (HR = 0.72,
CI = 0.64-0.81) and pre-detention non-outpatient treatment (HR = 0.55, CI = 0.41-0.74)
faced a decreased risk of substance-related treatment utilization post-detention.
Interestingly, being a violent offender and having a substance-related disorder emerged as
significant predictors for cohort one, whereas these variables were not significant for the
total sample; in addition, number of charges was no longer significant for cohort one
(Table 3.63). Significant variables associated with increased risk of either/both treatment
utilization among cohort one included male gender (HR = 1.16, CI = 1.02-1.32), number
of prior arrests (HR = 1.06, CI = 1.03- 1.10), violent offender (HR = 1.29, CI = 1.131.48), re-detained/ incarcerated (HR = 1.06, CI = 1.03-1.10), conduct disorder (HR =
2.69, CI = 2.31-3.14), non-conduct mental disorder (HR = 7.62, CI = 5.43- 10.70),
number of disorders (HR = 1.05, CI = 1.03-1.06), and pre-detention outpatient treatment
(HR = 3.58, CI = 2.83-4.53) (Table 3.65). The utilization of pre-detention non-outpatient
treatment failed to remain a significant variable for cohort one, whereas this predictor
was significant for analyses for the total sample (Table 3.63).

98
As displayed in Table 3.66, results for utilization of outpatient treatment services
at any time post-detention among cohort one DAs revealed that number of prior arrests
(HR = 1.05, CI = 1.03-1.07), charge severity (HR = 1.04, CI = 1.01-1.07), violent
offender (HR = 1.17, CI = 1.09-1.25), re-detained/ incarcerated (HR = 1.14, CI = 1.061.22), conduct disorder (HR = 3.53, CI = 3.26-3.83), mental disorder (HR = 5.23, CI =
4.76-5.74), substance-related disorder (HR = 1.69, CI = 1.57-1.82), and pre-detention
outpatient treatment (HR = 2.19, CI = 1.96-2.43) were significant variables associated
with increased risk of outpatient treatment utilization. In contrast, age (HR = 0.96, CI =
0.94-0.97) and pre-detention non-outpatient treatment (HR = 0.81, CI = 0.67-0.98) were
associated with decreased risk of outpatient treatment utilization. Compared to the total
sample (Table 3.63), pre-detention non-outpatient treatment was a new significant
predictor for cohort one, whereas number of disorders was no longer significant (as it was
for the total sample). Finally, with regards to non-outpatient treatment utilization at any
time among cohort one (Table 66), the following variables significantly increased the risk
of treatment services: prior arrests (HR = 1.05, CI = 1.02-1.09), conduct disorder (HR =
2.20, CI = 1.91-2.52), non-conduct mental disorder (HR = 6.53, CI = 5.35-7.37),
substance-related disorder (HR = 3.25 CI = 2.83-3.73), number of disorders (HR = 1.06,
CI = 1.05-1.07), and pre-detention outpatient treatment (HR = 2.26, CI = 1.83-2.81);
Black race (HR = 0.80, CI = 0.70-0.87) and age (HR = 0.96, CI = 0.93-0.99) were
associated with significantly decreased risk of non-outpatient treatment services.
Although pre-detention non-outpatient treatment was a significant variable for regression
analyses based on the total sample, this variable had no impact on the risk of postdetention non-outpatient treatment services for DAs in cohort one (Table 63).
Cohort two. Tables 3.67-3.70 show the results for the Cox proportional hazards
regression analyses among DAs within cohort two, with the addition of RAI scores and
MAYSI-2 positive screen as predictor variables. Consistent with analyses discussed
earlier, overall model summary statistics and change statistics at each stage were
significant and showed increasingly better fitting models at each stage (Table 3.67).
Significant variables associated with increased risk for treatment utilization at any time
post-detention for cohort two including being re-detained/incarcerated (HR = 1.33, CI =

99
1.14-1.54), having a conduct disorder (HR = 1.93, CI = 1.62-3 2.30), having a nonconduct mental disorder (HR = 2.10, CI = 1.71-2.57), having a substance-related disorder
(HR = 2.11, CI = 1.83-2.43), and larger number of disorders (HR = 1.06, CI = 1.05-1.08)
(Table 3.68). Variables associated with decreased risk for post-detention treatment
utilization included Black race (HR = 0.82, CI = 0.71-0.94), older age (HR = 0.88, CI =
0.84-0.93), and pre-detention outpatient treatment (HR = 0.74, CI = 0.62-0.88). As
displayed in Table 3.63, most of the criminal history variables (e.g., prior arrests, charge
severity, and violent offender) that significantly impacted treatment utilization for the
total sample were no longer significant; race emerged as a new significant predictor for
cohort two. As noted in Table 3.68, hazard ratios associated with increased risk for
mental health treatment utilization among cohort two were re-detained/incarcerated (HR
= 1.33, CI = 1.12-1.58), conduct disorder (HR = 2.76, CI = 2.25-3.40), non-conduct
mental disorder (HR = 4.24, CI = 3.21-5.61), and number of disorders (HR = 1.07, CI =
1.06-1.09). Black race (HR = 0.80, CI = 0.68-0.93), older age (HR = 0.85, CI = 0.810.90), and pre-detention outpatient treatment (HR = 0.63, CI = 0.52-0.77) were associated
with decreased risk of post-detention mental health treatment services. RAI scores and
MAYSI-2 positive screens were not significantly related to treatment utilization. Similar
to findings for any treatment utilization, most criminal history variables that were
significant among the total sample and cohort two failed to remain significant for
predicting mental health treatment utilization, while Black race emerged as a newly
significant finding associated with decreased risk of post-detention treatment services.
With regards to substance-related treatment at any time among cohort two, only
four variables emerged as significant (Table 3.69). Male gender (HR = 1.52, CI = 1.092.12), re-detained/ incarcerated (HR = 1.32, CI = 1.04-1.68), and number of disorders
(HR = 1.06, CI = 1.03-1.09) were linked to significantly increased risk of post-detention
treatment services, whereas pre-detention non-outpatient treatment (HR = 0.60, CI =
0.44-0.83) was associated with a decreased risk of post-detention substance-related
treatment services. In contrast to findings for the total sample and cohort one, most of the
criminal history variables (e.g., prior arrests, charge severity) and disorder variables (e.g.,
mental disorder, substance-related disorder) failed to significantly impact risk for

100
treatment utilization (Table 3.63). Regarding either/both treatment utilization at any time
post-detention among cohort two, male gender (HR = 1.61, CI = 1.12-2.31), redetained/incarcerated (HR = 4.54, CI = 1.16-2.05), conduct disorder (HR = 2.62, CI =
1.74-3.96), non-conduct mental disorder (HR = 1.83, CI = 1.23- 2.72), and number of
disorders (HR = 1.11, CI = 1.08-1.13) were associated with increased risk of treatment
services (Table 3.69). Black race (HR = 0.72, CI = 0.56-0.93), older age (HR = 0.89, CI
= 0.81-0.98), and pre-detention outpatient treatment (HR = 0.70, CI = 0.52-0.96) were
associated with decreased risk of either/both services during the study time frame.
Compared to the total sample, Black race and age became newly significant variables,
while prior arrests and violent offender became non-significant variables (Table 63).
As shown in Table 3.70, Cox regression analyses for outpatient treatment
utilization at any time among DAs in cohort two identified re-detained/incarcerated (HR
= l.30, CI = 1.12-1.52), conduct disorder (HR = 1.96, CI = 1.64-2.33), non-conduct
mental disorder (HR = 2.07, CI = 1.96-2.54), substance-related disorder (HR = 2.03, CI =
1.77-2.35), and number of disorders (HR = 1.06, CI = 1.05-1.08) as variables associated
with increased risk of post-detention outpatient treatment. In contrast, Black race (HR =
0.86, CI = 0.74-0.99), age (HR = 0.88, CI = 0.84-0.92) and pre-detention outpatient
treatment (HR = 0.74, CI = 0.63-0.88) emerged as variables associated with decreased
risk of treatment utilization. With the exception of re-detained/incarcerated, criminal
history variables that significantly impacted treatment utilization for the total sample and
cohort one were no longer significant (Table 3.63). In addition, Black race and predetention outpatient treatment emerged as significant predictors of outpatient treatment
services. Lastly, significant variables associated with an increased likelihood for nonoutpatient treatment utilization at any time among DAs in cohort two included prior
arrests (HR = 1.07, CI = 1.03-1.10), positive screen on the MAYSI-2 (HR = 1.68, CI =
1.04-2.70), non-conduct mental disorder (HR = 3.88, CI = 1.79-8.41), substance-related
disorder (HR = 3.23, CI = 2.09-4.98), and number of disorders (HR = 1.11, CI = 1.081.14) (Table 3.70). Male gender (HR = 0.61, CI = 0.41-0.90), Black race (HR = 0.42, CI
= 0.29-0.60), and older age (HR = 0.78, CI = 0.68-0.89) were all associated with
decreased likelihood of non-outpatient treatment services. As noted in Table 3.63,

101
utilization of pre-detention treatment services failed to remain significant for cohort two.
Of note, positive screen on the MAYSI-2 emerged as a significant variable for the first
time and was associated with increased risk of non-outpatient treatment utilization.

3.12 Aim 4. Patterns of Treatment Utilization: Time to Termination
Life table. Table 3.71 and Figure 3.13 displays the results for time (in days)
involved in continuous outpatient treatment services until treatment termination, starting
with the date of first outpatient session and ending with date of final continuous (i.e., no
gaps) outpatient treatment session. As shown in the life table, a total of 6234 DAs
obtained at least one outpatient treatment session. By the end of the first time interval at
3-months follow-up, a total of 1656 DAs terminated treatment services, leaving
approximately 73% of the sample surviving to continue utilization of outpatient
treatment. The probability of dropping out within the first three months of detention
release was approximately 8.9% (SD = .002) and the hazard rate equaled .10 (SD = .00).
At one year post-detention, a total of 3390 DAs remained involved in outpatient
treatment, meaning approximately 2844 DAs terminated treatment and approximately
50% of the sample remained engaged in outpatient treatment. Altogether, the majority of
DAs terminated outpatient treatment within two years of detention release. Treatment
termination occurred during all time intervals, with the final treatment termination
occurring within nine years of release from detention.
Kaplan-Meier survival analyses. Table 3.72 displays the main results from the
Kaplan-Meier survival analyses for days involved in continuous outpatient treatment until
terminating treatment. Timing of termination differed significantly across treatment type,
with DAs terminating substance-related services much sooner than mental health services
and either/both services (Figure 3.14). The survival curve for terminating mental health
treatment differed significantly from the curves for terminating substance-related
outpatient treatment (Wilcoxon= 56.58, p < .001; Tarone-Ware = 56.88, p < .001; Log
Rank = 53.84, p < .001) and terminating either/both outpatient treatment (Wilcoxon=
25.15, p < .001; Tarone-Ware = 26.06, p < .001; Log Rank = 21.09, p < .001) (Table
3.72). Survival curves for substance-related outpatient treatment and either/both

102
outpatient treatment were also significantly different over time (Wilcoxon= 78.77, p
< .001; Tarone-Ware = 82.61, p < .001; Log Rank = 79.99, p < .001). Altogether, the
survival curves for the three treatment types (Figure 3.14) did not intersect and remained
significantly different throughout the entire time frame of the study (Wilcoxon= 96.18, p
< .001; Tarone-Ware = 97.40 p < .001; Log Rank = 89.23, p < .001).

3.12.1 Time to Termination by Groups
Demographic groups. Consistent with survival analyses results for time to first
treatment services, results were significant for time to treatment termination across
gender, race, and age cohorts (see Table 3.72 and Figures 3.15-3.17). Survival curves
(Figure 3.15) differed significantly between males versus females in the short-term (Log
Rank = 144.72, p < .001), middle, (Wilcoxon = 135.66, p < .001) and long-term (TaroneWare = 165.85, p < .001) of the study time frame (Table 3.72); females utilized
outpatient services for a significantly longer number of days (Mdn = 746.0 days, CI =
671.90-820.07) than males (Mdn = 356.0 days, CI = 328.30-383.71). Regarding race,
White DAs remained involved in outpatient treatment for approximately 639.0 days (CI =
574.98-702.02), which was significantly longer than the approximately 369.0 days (CI =
335.30-402.70) that Black DAs remained involved in outpatient treatment. The survival
curves (Figure 3.16) for White DAs versus Black DAs were notably distinct and
remained significantly different over time (Wilcoxon= 70.25, p < .001; Tarone-Ware =
72.37, p < .001; Log Rank = 54.44, p < .001). Regarding age, survival curves (Figure
3.17) for time involved in outpatient treatment significantly differed across the three age
cohorts (Wilcoxon= 151.75, p < .001; Tarone-Ware = 128.77, p < .001; Log Rank =
71.99, p < .001). A clear pattern emerged in which older DAs terminated outpatient
treatment earlier than mid-age DAs (Wilcoxon = 27.99, p < .001; Tarone-Ware = 21.97, p
< .001, Log Rank = 12.55, p < .001), who in turn terminated outpatient treatment earlier
than younger DAs (Wilcoxon = 69.91, p < .001; Tarone-Ware = 61.18, p < .001; Log
Rank = 33.71, p < .001).
Mental health groups. Kaplan-Meier survival analyses produced significant
results for MAYSI-2 groups, disorder groups, and pre-detention groups regarding time

103
until treatment termination (Table 3.72). Regarding MAYSI-2 groups, DAs who did not
complete the MAYSI-2 remained in outpatient services longest, with a median of
approximately 552.0 days (CI = 710.56-855.44), compared to approximately 244.0 days
(CI = 201.74-286.26) for DAs with non-positive screens and approximately 398.0 days
(CI = 361.34-434.66) for DAs with positive screens on the MAYSI-2. As displayed in
Figure 3.18, the survival curves across the three groups intersected, but these curves were
significantly different from each other in the short-term (Wilcoxon = 84.88, p < .001),
middle (Tarone-Ware = 154.39, p < .001), and long-term (Log Rank = 155.15, p < .001).
As noted earlier, since the follow-up periods differed drastically for these groups (i.e.,
cohort one versus cohort two), results should be interpreted with caution. When
comparing just DAs who completed the MAYSI-2 (and were all detained post-2005),
adolescents with positive screens remained involved in outpatient treatment services
significantly longer than adolescents with non-positive screens throughout the entire
study time frame (Wilcoxon= 29.50, p < .001; Tarone-Ware = 30.01, p < .001; Log Rank
= 26.64, p < .001) (Table 3.72).
Survival curves for the three disorder groups (Figure 3.19) were significantly
different from each other over time (Wilcoxon = 1197.91, p < .001; Tarone-Ware =
1185.34, p < .001; Log Rank = 1044.45, p < .001) (Table 3.72). Specifically, comorbid
DAs remained engaged in outpatient treatment longer than DAs with mental disorders
(Wilcoxon= 555.53, p < .001; Tarone-Ware = 513.78, p < .001; Log Rank = 395.44, p
< .001) and DAs with substance-related disorders (Wilcoxon= 1262.40, p < .001; TaroneWare = 1241.91, p < .001; Log Rank = 1034.70, p < .001). In addition, DAs with mental
disorders remained involved in outpatient treatment significantly longer than DAs with
substance-related disorders (Wilcoxon= 269.80, p < .001; Tarone-Ware = 296.3, p
< .001; Log Rank = 295.80, p < .001). Regarding pre-detention treatment groups,
survival curves (Figure 3.20) were significantly different throughout the 14 years of the
study (Wilcoxon= 4.72, p < .001; Tarone-Ware = 11.32, p < .001; Log Rank = 331.85, p
< .001). As noted in Table 3.72, DAs with pre-detention treatment remained involved in
outpatient treatment for a significantly larger number of days than DAs with no history of
pre-detention treatment.

104
Criminal history groups. Similar to findings for time to first treatment utilization,
survival analyses revealed significant differences in time to terminate from treatment
across violent offender groups, recidivist groups, and re-detained/incarcerated groups
(Table 3.72). The survival curves (Figure 3.21) differed significantly between violent
offenders and nonviolent offenders (Wilcoxon= 12.85, p < .001; Tarone-Ware = 17.34, p
< .001; Log Rank = 20.77, p < .001), with violent offenders remaining in outpatient
treatment for a median of approximately 519.0 days (CI = 394.72-467.28) compared to
non-violent offenders remaining in treatment for approximately 431.0 days (CI = 462.41575.59) (Table 3.72). Similarly, recidivist youth utilized outpatient treatment for
significantly more days than non-recidivist youth (Mdnrecidivist = 544.0 days vs. Mdnnon =
342.0 days) and re-detained/incarcerated DAs utilized outpatient treatment for
significantly more days than other DAs (Mdnre-detained = 532.0 days vs. Mdnnot = 343.0
days). The survival curves (Figure 3.22) for the two recidivism groups were significantly
different (Wilcoxon= 42.11, p < .001; Tarone-Ware = 34.4, p < .001; Log Rank = 22.70,
p < .001), as were the survival curves (Figure 3.23) for the two re-detained/incarcerated
groups (Wilcoxon= 37.74, p < .001; Tarone-Ware = 31.90, p < .001; Log Rank = 21.08, p
< .001) (see Table 3.72 for more details).

3.12.2 Time to Termination: Cox Regression Analyses
Table 3.73 displays the model summary statistics and change statistics for the Cox
proportional hazards regression analyses predicting time until terminating outpatient
treatment. Results from the chi-square tests indicate that overall models were
significantly different than null models; the addition of predictors at each stage reduced
the -2 log likelihood and significantly improved the models. With regards to terminating
from outpatient treatment at any time post-detention (Table 3.74), the following variables
were associated with increased risk of termination: male gender (HR = 1.34, CI = 1.271.42), older age (HR = 1.06, CI = 1.04-1.08), higher number of prior arrests (HR = 1.04,
CI = 1.03-1.05), more severe charge (HR = 1.04, CI = 1.01-1.07), conduct disorder (HR =
1.19, CI = 1.12-1.28), pre-detention non-outpatient treatment (HR = 1.51, CI = 1.40-163),
and pre-detention outpatient treatment (HR = 1.26, CI = 1.13-1.40). In addition, DAs

105
who were violent offenders (HR = 0.94, CI = 0.89-0.99), diagnosed with non-conduct
mental disorders (HR = 0.59, CI = 0.55-0.65), diagnosed with substance-related disorders
(HR = 0.90, CI = 0.85-0.95), and diagnosed with more disorders (HR = 0.94, CI =0.910.97) faced a decreased risk of terminating from post-detention outpatient treatment.
Results were similar for mental health outpatient treatment, in that male gender (HR =
1.25, CI = 1.17-1.34), age (HR = 1.05, CI = 1.03-1.07), number of prior arrests (HR =
1.05, CI = 1.03-1.06), conduct disorder (HR = 1.15, CI = 1.06-1.25), pre-detention
outpatient treatment (HR = 1.32, CI = 1.17-1.50), and pre-detention non-outpatient
treatment (HR = 1.32, CI = 1.17-1.50) were associated with increased risk of treatment
termination (Table 3.74). Larger number of disorders (HR = 0.94 CI = 0.93-0.96) and
non-conduct mental disorder (HR = 0.42, CI = 0.37-0.45) yielded decreased risk of
terminating mental health outpatient treatment during the study time frame.
As shown in Table 3.75, significant variables associated with termination from
substance-related outpatient treatment at any time post-detention included male gender
(HR = 1.73, CI = 1.51-1.97), older age (HR = 1.12, CI = 1.08-1.16), number of prior
arrests (HR = 1.04, CI = 1.03-1.05), number of charges (HR = 1.08, CI = 1.02-1.15),
charge severity (HR = 1.34, CI = 1.11-1.62), and pre-detention outpatient treatment (HR
= 1.13, CI = 1.11-1.62). DAs with substance-related disorders (HR = 0.94, CI = 0.920.96) and greater number of disorders (HR = 0.94, CI = 0.91-0.97) had decreased risk of
terminated substance-related outpatient treatment. Finally, the following variables were
associated with significantly increased risk of termination from either/both outpatient
treatment at any time post-detention: older age (HR = 1.13, CI = 1.07-1.20), larger
number of prior arrests (HR = 1.03, CI = 1.01-1.06), and pre-detention outpatient
treatment (HR = 1.89, CI = 1.47-2.42). Having a non-conduct mental disorder (HR =
0.40, CI = 0.28-0.55) and larger number of disorders (HR = 0.94, CI = 0.93-0.96) were
associated with decreased risk of termination from post-detention either/both outpatient
treatment.

106
3.12.3 Time to Termination by Cohorts
Cohort one. As shown on Table 3.76, Cox regression analyses for termination
from outpatient treatment among cohort one produced regression models that were
significantly different than null models. With the exception of either/both outpatient
treatment, the addition of predictor variables at each stage generated significantly better
fitting models, as indicated by decreased -2 log likelihood values and significant change
statistics. With regards to either/both outpatient treatment, the regression model for stage
three, with all predictors entered into the analyses was the only significant model.
Compared to findings based on the total sample, the Cox regression models
predicting treatment termination among cohort one yielded slightly different results (see
Table 3.77 for direct comparisons). For termination from any outpatient treatment among
cohort one, the following variables were associated with an increased risk of termination
at any time post-detention: male gender (HR = 1.27, CI = 1.19-1.36), higher number of
prior arrests (HR = 1.04, CI = 1.02-1.06), conduct disorder (HR = 1.16, CI = 1.07-1.25),
and pre-detention outpatient treatment (HR = 1.20, CI = 1.08-1.34) (Table 3.78).
Variables associated with decreased risk of terminating outpatient treatment included
being a violent offender (HR = 0.93, CI = 0.87-0.99), having a non-conduct mental
disorder (HR = 0.62, CI = 0.56-0.68), having a substance-related disorder (HR = 0.91, CI
= 0.85-0.98), and larger number of disorders (HR = 0.95, CI = 0.94-0.96). As shown in
Table 3.77, results were similar for the total sample and cohort one, with the exception
that age, charge severity, and pre-detention non-outpatient treatment were no longer
significant predictors of termination among DAs in cohort one.
With regards to terminating mental health outpatient treatment for cohort one,
male gender (HR = .19, CI = 1.09-1.29), larger number of prior arrests (HR = 1.35, CI =
1.01-1.55), conduct disorder (HR = 1.11, CI = 1.01-1.22), and pre-detention outpatient
treatment (HR = 1.25, CI = 1.11-1.41) emerged as significant variables associated with
increased risk of termination. In contrast, non-conduct mental disorder (HR = 0.40, CI =
0.35-0.47) and higher number of disorders (HR = 0.94, CI = 0.93-0.96) were associated
with decreased risk of termination from post-detention mental health outpatient
treatment. When restricting analyses to only DAs within cohort one, age and pre-

107
detention non-outpatient treatment were not significant variables associated with
terminating mental health treatment, whereas these variables were significant predictors
for the total sample (Table 3.77).
As shown in Table 3.79, significant variables associated with increased risk of
terminating substance-related outpatient treatment among cohort one included male
gender (HR = 1.60, CI = 1.38-1.85), age (HR = 1.07, CI = 1.02-1.12), and larger number
of prior arrests (HR = 1.07, CI = 1.03-1.11). DAs with substance-related disorders (HR =
0.90, CI = 0.84-0.97) and more disorders (HR = 0.95, CI = 0.93-0.98) faced significantly
decreased risk of termination from post-detention substance-related outpatient treatment.
Contrary to findings for the total sample, number of charges, charge severity, conduct
disorder, non-conduct mental disorder, and pre-detention outpatient treatment failed to
remain significant variables for cohort one with regards to risk of termination from
substance-related outpatient treatment services (Table 3.77). For termination from
either/both post-detention outpatient treatment among cohort one, age (HR = 12.10, CI =
1.02-1.18) was associated with increased risk, while length of detention stay (HR = 0.98,
CI = 0.97-0.99), mental disorder (HR = 0.29, CI = 0.17-0.51) and number of disorders
(HR = 0.95, CI = 0.93-0.97) were associated with decreased risk of terminating services
(Table 3.79). As noted in Table 3.77, compared to findings for the total sample, length of
stay emerged as a newly significant variable and prior arrests failed to remain significant
in impacting the risk of termination from either/both outpatient treatment among DAs in
cohort one.
Cohort two. Table 3.80 displays the model summary statistics and change
statistics for the Cox regression analyses for treatment termination among cohort two,
with the addition of RAI scores and MAYSI-2 positive screen (yes/no) as independent
predictors. Results from the chi-square tests indicate that the four regression models
were significantly different than null models with no predictors. The addition of
predictors at stage one and stage three significantly improved the models; however, the
entry of criminal history variables at stage two (including RAI scores) failed to generate
better fitting regression models. The following variables were associated with
significantly increased risk of termination from post-detention outpatient treatment

108
among cohort two: age (HR = 1.37, CI = 1.31-1.42), larger number of prior arrests (HR =
1.02, CI = 1.00-1.03), charge severity (HR = 1.09, CI = 1.03-1.16), pre-detention nonoutpatient treatment (HR = 1.19, CI = 1.03-1.27), and pre-detention outpatient treatment
(HR = 1.31, CI = 1.13-1.53). In contrast, DAs diagnosed with non-conduct mental
disorders (HR = 0.53, CI = 0.45-0.63), substance-related disorders (HR = 0.68, CI = 0.610.71), and more disorders (HR = 0.91, CI = 0.90-0.93) faced a decreased risk of
terminating post-detention outpatient treatment. As shown in Table 3.77, male gender,
violent offender, and conduct disorder were all significant variables associated with
treatment termination for the total sample and cohort one, but these variables were nonsignificant for cohort two.
Regarding mental health outpatient treatment among DAs in cohort two, variables
associated with increased risk of termination included male gender (HR = 1.19, CI =
1.01-1.40), age (HR = 1.63, CI = 1.25-1.36), number of prior arrests (HR = 1.02, CI =
1.00-1.04), pre-detention non-outpatient treatment (HR = 1.37, CI = 1.14-1.64), and predetention outpatient treatment (HR = 1.43, CI = 1.18- 1.68) (Table 3.81). Significant
variables associated with decreased risk of termination included RAI scores (HR = 0.98
CI = 0.97-0.99), non-conduct mental disorder (HR = 0.45, CI = 0.35-0.58), substancerelated disorder (HR = 0.67, CI = 0.58-0.78), and larger number of disorders (HR = 0.92,
CI = 0.90-0.94). Compared to prior analyses for the total sample and cohort one,
substance-related disorders emerged as a significant variable for cohort two, whereas
conduct disorder was no longer associated with risk of terminating mental health
outpatient treatment (Table 3.77). It should be noted that RAI scores emerged as a
significant variable; this marked the only regression analysis in which RAI scores were
significant.
As shown in Table 3.82, four variables were significantly associated with
terminating post-detention substance-related outpatient treatment among DAs in cohort
two. Specifically, male gender (HR = 1.56, CI = 1.04-2.33) and older age (HR = 1.47, CI
= 1.34-1.62) were linked to an increased risk, whereas substance-related disorder (HR =
0.78, CI = 0.44-0.93) and number of disorders (HR = 0.89, CI = 0.85-0.94) were linked to
a decreased risk of termination. When considering results based on the total sample

109
(Table 3.77), charge number, charge severity, conduct disorder, non-conduct disorder,
and pre-detention outpatient treatment were significant variables for the total sample, but
non-significant variables for cohort two. Finally, with regards termination of postdetention either/both outpatient treatment among cohort two, older age (HR = 1.52, CI =
1.33-1.73) and length of stay (HR = 1.02, CI = 1.01-1.03) were associated with
significantly higher risk of termination. Mental disorder (HR = 0.52, CI = 0.32-0.86) and
larger number of disorders (HR = 0.90, CI = 0.86-0.94) were associated with lower risk
of terminating from either/both outpatient treatment services (Table 3.82). Compared to
findings for the total sample and cohort one (Table 3.77), charge number and predetention outpatient treatment were no longer significant variables and length of stay
produced significant hazard ratios in the opposite direction for cohort one versus cohort
two. Finally, screening positively on the MAYSI-2 failed to significantly impact the risk
of termination from post-detention outpatient treatment during the study time frame.

110

CHAPTER 4. DISCUSSION

4.1 Discussion
This longitudinal study represents one of the only projects to carefully examine
mental health and substance-related treatment data within the detained adolescent (DA)
population over an extended period of time (Paskar, 2008; Teplin et al., 2012) in order to
better understand treatment utilization patterns within this population, address gaps
within the literature, and identify areas for future research. By sampling a large number
of DAs for a total of 14 years, the study was able to achieve its primary purpose and four
study aims involving a cross-sectional exploration of post-detention treatment utilization,
examination of differences in post-detention treatment utilization among DA groups,
identification of predictors of post-detention treatment utilization, and investigation of
longitudinal patterns of post-detention service utilization over time. Study findings mark
an important contribution to the literature by helping to frame and better understand the
relationships between high rates of mental health and substance use problems (Fazel et
al., 2008; Shufelt & Cocozza, 2006; Wasserman et al., 2002), low rates of service
utilization (Hussey et al., 2008; Paskar, 2008; Trupin et al., 2004), and high rates of
recidivism among DAs (Abram et al., 2008; Grunwald et al., 2010). Key findings and
recommendations for future research are discussed below.

4.2 Criminal History Outcomes
First, several important findings emerged related to criminal history. Consistent
with other research of detained adolescents (Sedlak, 2009; Sedlak & Bruce, 2010), the
majority of participants in the sample had contact with the juvenile justice system prior to
being placed in detention and most were arrested for conduct-related charges, propertyrelated charges, or substance-related charges. More importantly, the DAs in this study

111
had negative long-term outcomes that are common among juvenile justice-involved
youth. Specifically, roughly two out of three DAs were re-arrested following release
from detention, with about 20% of DAs recidivating within three months of reentry into
the community and about 40% of DAs recidivating within two years of reentry into the
community. In addition, approximately 54% of the sample had a second detention stay
and 16% of the sample was incarcerated in prison. Again, the majority of DAs
experienced these outcomes within two years of release from detention. As found in
prior research (Grunwald et al., 2010; Harrison, 2001), such findings highlight the
revolving door problem, in which DAs are stuck in an ongoing cycle of juvenile justice
system involvement by being repeatedly arrested, detained, released, re-arrested, redetained, and sometimes placed in the adult prison population. Being stuck in the
revolving door of the juvenile justice system is problematic because it reduces the
likelihood of obtaining evidence-based mental health treatment services and interferes
with successful reintegration into the community (Bullis et al., 2002; Calley, 2012;
Samuel, 2015). Findings from this study therefore highlight the need for additional
research that examines ways to reduce and ideally resolve the revolving door problem
faced by many juvenile justice-involved youth.
Studies of detained adolescents have shown that males and racial/ethnic
minorities tend to be disproportionately targeted and involved in the juvenile justice
system (Bostwick, & Ashley, 2009; Rozie-Battle, 2002); the current study also found
evidence of such gender and racial disparities. With regards to gender, the initial sample
of detained adolescents contained significantly more males than females (i.e., about
62.7% versus 37.3%) and significantly more males than females were re-arrested and redetained following release from detention. These findings may reflect true differences, in
that males engage in more frequent and/or serious criminal behavior and are therefore
more heavily involved in the juvenile justice system (Office of Juvenile Justice and
Delinquency Prevention, 2013; Veysey, 2003). This would explain the fact that males in
the sample averaged more pre-detention contacts, higher number of criminal charges,
more severe criminal charges, higher RAI scores, and higher prevalence of violent
offenses than females in the sample. Alternatively, it is possible that gender differences

112
stem from preferential treatment of females over males by authority figures, who are
more likely to punish, arrest, and/or detain males than females (Bostwick, & Ashley,
2009; Cauffman et al., 2005).
Interestingly, gender differences did not emerge in this study for rates of
incarceration in prison facilities. Although this finding is unexpected and contrary to
other research studies (Barrett, Katsiyannis, & Zhang, 2006; Bostwick & Ashley, 2009),
it is important to note that commitment to DOC prison facilities was fairly rare among the
sample (i.e., about 16%) and likely reserved for adolescents who committed extremely
serious crimes (e.g., murder, homicide, rape) (Barrett et al., 2006). Thus, being
incarcerated may be strongly tied to severity of crime, rather than demographic factors
like gender. Additional support for this possibility is that rates of incarceration failed to
differ significantly between Black DAs and White DAs, despite clear racial differences
(discussed below) for almost all other criminal history variables.
Similar to findings for gender, the percentage of Black DAs was notably higher
than White DAs in the initial detention sample, as well as among recidivists and redetained youth. Black DAs averaged younger age of first contact with the juvenile
justice system, more arrests prior to detention, higher prevalence of violent offenses, and
higher RAI scores than White DAs. Although not discussed in the Results section, these
racial differences in criminal activity remain significant when considering race and
gender together. That is, when restricting analyses by gender, results continue to show
more severe criminal activity (e.g., higher severity of charges, more prior arrests) and
more serious criminal history outcomes (e.g., recidivism, re-detention) among Black
females compared to White females, as well as Black males compared to White males.
Altogether, findings may be interpreted as reflecting true racial differences, with
minority youth being more likely than White youth to engage in criminal activity.
Alternatively, many researchers have suggested that racial differences are most likely
caused by bias and discrimination among law enforcement officers who
disproportionately arrest Black youth in the community and among court officials who
disproportionately place Black youth in detention centers or prison facilities (Maschi et
al., 2008; Towberman, 1994; Wolf, 2008). This bias against Black DAs is fairly apparent

113
when considering that study findings applied to both Black males and Black females,
thereby eliminating the argument made by some researchers that racial disparities seem to
largely impact Black males, rather than Black females (Bostwick & Ashley, 2009).
Moreover, according to the United States Census Bureau, the demographic distribution of
adolescents in Marion County was 69.1% White and 25.5% Black in 2000 and 61.8%
White and 27.5% Black in 2010, whereas the distribution of the sample during the same
time frame was 34.8% White and 58.6% Black (and 6.5% Other). The marked
divergence between the racial distributions of the general population and the detained
adolescent population in this study provides strong evidence of prejudice and unfair
treatment against Black youth within the juvenile justice system, which is consistent with
prior research (Hoytt, Schiraldi, Smith, & Zeidenberg, 2001; Isaacs, 1992; Wolf, 2008).

4.3 Mental Health Outcomes
Besides criminal history outcomes, several important findings emerged pertaining
to the mental health status of detained adolescents. First and foremost, findings showed
high rates of mental health and substance-related concerns within the sample.
Approximately 68% of DAs who completed the Massachusetts Youth Screening
Instrument-2nd Edition (MAYSI-2) screened positively on the measure, meaning they
were experiencing significant behavioral health concerns during their detention stay that
likely required additional assessment and treatment (Grisso et al., 2001; Vincent et al.,
2008). Furthermore, almost 73% of participants were diagnosed with a mental disorder
and/or substance-related disorder, 65.7% of DAs met criteria for at least one mental
disorder, 35.6% of DAs met criteria for at least one substance-related disorder, and
39.6% of DAs were comorbid for both types of disorders. To put such rates into context,
consider that the prevalence of such disorders among the general adolescent population is
only about 15%-20% (Robertson et al., 2004; Hoeve et al., 2013), which is markedly
lower than the rates found in the current study, as well as other research (Fazel et al.,
2008; Hussey et al., 2008; Paskar, 2008; Sedlak & Bruce, 2010). Moreover, the
prevalence of substance-related disorders, mood disorders (40.1%), and conduct disorders
(34.9%) were actually lower in this study than prior epidemiological studies of detained

114
adolescents (Fazel et al., 2008; Shelton, 2001; Teplin et al., 2006). Overall, current
findings not only highlight the substantial mental health concerns found among DAs, but
also add to the growing literature indicating that the detained adolescent population is a
vulnerable and high-risk population in desperate need of appropriate mental health and
substance-related treatment services (as discussed in more detail below).
In addition to prominent behavioral health needs, findings also revealed
interesting patterns between demographic variables and mental health variables. With
regards to gender, females tended to endorse higher overall symptomology and more
severe mental health concerns than males, whereas males tended to have more severe
substance-related concerns than females. For example, a significantly higher proportions
of females than males screened positively on the MAYSI-2 (about 81.2% versus 66.8%)
and met criteria for a mental disorder (about 38.9% versus 35.8%); a higher proportion of
males than females met criteria for a substance-related disorder (about 8.1% versus
4.8%). Further, being female was significantly correlated with higher MAYSI-2 scores,
higher number of mental disorders, and comorbidity, while being male was significantly
correlated with having a substance-related disorder and higher number of substancerelated disorders. Altogether, current findings fit well with other research suggesting
more significant mental health concerns in female DAs compared to male DAs
(Robertson et al., 2004; Veysey, 2003; Vincent et al., 2008). As mentioned earlier, the
differential treatment of males and females within the legal system may account for such
findings. Specifically, females are less likely to be arrested and detained (Bostwick, &
Ashley, 2009; Office of Juvenile Justice and Delinquency Prevention, 2013), so female
DAs tend to be the most problematic and deviant females involved in the juvenile justice
system, who may also have the most severe mental health problems (Hussey et al., 2008;
Veysey, 2003).
Of the few studies that have examined mental disorders and substance-related
disorders separately (Aarons et al., 2004; Domalanta, Risser, Roberts, & Risser, 2003;
Paskar, 2008), results have largely followed the pattern found in this study (i.e., higher
prevalence of mental disorders in females and higher prevalence of substance-related
disorders in males). However, this distinction is not widely discussed and represents an

115
important finding with implications for connecting DAs to appropriate treatment services.
Specifically, it may be more efficient and cost-effective for staff and providers within the
juvenile justice system to refer and/or recommend female DAs for mental health
interventions and refer and/or recommend male DAs for substance-related interventions.
Given the limited data currently available, however, additional research is needed to
investigate the efficacy of such a system.
Interestingly, study findings revealed a pattern of divergent results for mental
health concerns versus substance-related concerns for age, much like findings for gender.
Analogous to findings for females versus males, younger DAs (age 11-13) tended to have
more prominent mental health problems, with significantly more younger and mid-aged
DAs than older DAs screening positively on the MAYSI-2 (approximately 71.8% versus
71.1% versus 66.5%) and meeting criteria for any disorder (approximately 77.6% versus
73.1% versus 68.4%). Being younger upon detention entry (like being female) was
significantly associated with having any disorder, having a conduct disorder, and larger
number of mental disorders. Conversely, being older upon detention entry (like being
male) was associated with having a substance-related disorder and larger number of
substance-related disorders.
To my knowledge, research has yet to fully examine differences in behavioral
health status across age cohorts, thereby making it difficult to explain current findings. It
is reasonable to suppose that DAs who are younger during their first detention stay are
likely to be serious offenders who represent the most deviant youth involved in the
juvenile justice system, much like detained females. These youth may also be more
likely to have serious mental health concerns, particularly conduct-related issues (Kates
et al., 2014; Sedlak & Bruce, 2010), thereby resulting in higher prevalence of mental
disorders in younger DAs than older DAs (Aalsma et al., 2012b). The higher prevalence
of substance-related disorders among older DAs than younger DAs may be explained by
the onset of substance use. Typically, adolescents do not start experimenting heavily
with substances until ages 14-15 years (Aarons et al., 2004; Kessler, Berglund, Demler,
Jin, Merikangas, & Walters, 2005). Thus, younger DAs, who are only 11-13 years at

116
detention entry, are less likely to have experimented and/or developed substance-related
disorders than their older counterparts, who are 16-18 years at detention entry.
With regards to differences in mental health outcomes for race, findings clearly
showed more prominent mental health concerns among White DAs than Black DAs.
Specifically, a significantly higher percentage of White DAs compared to Black DAs
screened positively on the MAYSI-2 (about 75.7% vs. 67.6%), met criteria for a mental
disorder (about 73.9% vs. 60.8%), met criteria for a substance-related disorder (about
43.5% versus 30.8%), and met criteria for comorbidity (about 37.5% vs. 23.4%). Being
White was also significantly correlated with higher MAYSI-2 scores, larger number of
mental disorders, and larger number of substance-related disorders. In general, findings
are consistent with the literature, in which White DAs tend to have more severe
symptomology than Black DAs and are more likely to meet criteria for a mental disorder,
as well as multiple disorders (Soenksen et al., 2015; Teplin et al., 2012; Vincent et al.,
2008). However, findings from a large meta-analysis based on approximately 70,000
youth from 283 juvenile justice facilities indicated that mental health concerns among
DAs may not actually differ by race (Vincent et al., 2008); rather, racial differences may
be due to White DAs being more likely than minority DAs to report their symptoms
and/or providers being biased against minority DAs, resulting in higher likelihood of
White DAs receiving psychiatric diagnoses. Based on findings from their meta-analysis,
Vincent and colleagues (2008) noted, “One cannot conclude that there are consistent,
cross-site racial differences in the presence of symptoms among youths in the juvenile
justice system (pp. 289).” Inconsistencies and discrepancies across racial groups may
therefore reflect differences in the behaviors of DAs and/or providers, rather than true
differences in mental health concerns and treatment needs.

4.4 Aim 1. Post-Detention Treatment Utilization
Given that the current literature provides inconsistent and varied findings
regarding treatment utilization among DAs, one of the primary aims of the study involved
a cross-sectional exploration of post-detention treatment utilization among DAs. This
exploration yielded important findings related to this population’s use of mental health

117
and substance-related treatment service upon community reentry. First and foremost,
results revealed post-detention prevalence rates of approximately 66.6% for any
treatment utilization, as well as 56.9% for mental health treatment utilization, 26.6% for
substance-related treatment utilization, and 17.1% for either/both treatment utilization
(i.e., unclear type of service). As a comparison, the meta-analysis I conducted (White et
al., under review), based on 32 studies, yielded post-detention prevalence rates of 34.3%
for mental health treatment, 29.0% for substance-related treatment, and 27.7% for
either/both treatment. Thus, compared to the results of the meta-analysis, current
estimates for any treatment service utilization and mental health treatment utilization
were notably higher than expected, whereas estimates for substance-related treatment
utilization and both/either treatment utilization were fairly consistent with the literature.
It should be noted, however, that the longitudinal studies included in the metaanalysis (Bullis et al., 2004; Hussey et al., 2007; Teplin et al., 2005; Trupin et al., 2004)
tracked DAs for one to three years after detention release, whereas the current study
tracked participants for one to fourteen years. As a more accurate comparison to prior
work, prevalence estimates for treatment utilization within two years of detention release
for this sample were 36.7% for any treatment utilization, 31.4% for mental health
treatment utilization, 10.4% for substance related treatment utilization, and 5.4% for
either/both treatment utilization. Using these estimates, current prevalence rates for
mental health treatment services appear to be similar to other studies (Paskar, 2008;
Teplin et al., 2005), with the unexpected finding of notably low utilization of substancerelated services.
With regards to limited utilization of substance-related treatment, study findings
revealed significantly higher utilization of mental health services than substance-related
services for both outpatient and non-outpatient treatment services. Specifically, about
55.5% of DAs utilized mental health outpatient treatment and 18.6% utilized nonoutpatient treatment, compared to only 23.5% of DAs who utilized substance-related
outpatient treatment and 10.1% who utilized non-outpatient treatment during the entire
study time frame. Findings that more adolescents received mental health services than
substance-related services is somewhat surprising, since substance use programs are the

118
most commonly offered services in juvenile justice facilities and judges are more likely
to mandate court-ordered substance-related treatment than mental health treatment (Desai
et al., 2006). In addition, substance use is strongly associated with increased risk of
chronic and/or violent delinquency and higher rates of recidivism (Bath et al., 2013;
Colins, Vermeiren, Vreugdenhil, Van den Brink, Doreleijers, & Broekaert, 2011), which
should theoretically prompt efforts within the juvenile justice system to ensure DAs
obtain substance-related treatment upon reentry into the community.
Several factors may account for current findings. First, the prevalence of mental
disorders among DAs was higher than substance-related disorders (about 65.7% versus
35.6%), so high rates of mental health service utilization may simply be related to more
DAs in need of such services. However, this does not fully explain findings, since only
37.0% of all DAs with substance-related disorders and only 26.3% of comorbid DAs
obtained services within two years of detention release. At the same time, over 90% of
DAs with substance-related disorders eventually obtained treatment services by the end
of the study time frame. Given that adolescents are often resistant to tackling substancerelated concerns (Becan, Knight, Crawley, Joe, & Flynn, 2014), the delayed timing for
treatment utilization may reflect the longer time it takes youth to become ready and
willing to participate in substance-related treatment. Alternatively, findings may be
explained by the fact that effective mental health treatments (e.g. multisystemic therapy,
family functional therapy) typically address comorbid issues, often producing positive
outcomes related to both mental health and substance-related symptomology (Darnell &
Schuler, 2015; Lipsey, 2011; Sexton, & Turner, 2010). Thus, staff within the juvenile
justice system may be intentionally referring and/or connecting DAs to mental health
services in the community, rather than substance-related services, because such services
are more comprehensive and effective treatment options (Hoeve et al., 2013). Similarly,
DAs participating in mental health interventions may experience improvements in
substance-related concerns (Darnell & Schuler, 2015; Henggeler & Schoenwald, 2011),
thereby eliminating the need for substance-related services. Any or all of these reasons
may account for the larger prevalence rates for mental health treatment utilization than
substance-related treatment utilization at different follow-up periods of the study.

119
4.5 Aim 4. Patterns of Post-Detention Treatment Utilization
Several important findings emerged pertaining to the pattern in which DAs were
connected to treatment services following release from detention. Not surprisingly, the
prevalence of treatment utilization increased as length of time from detention increased,
since DAs had more opportunities to obtain treatment services. Somewhat surprisingly,
however, was that the rate in which DAs were connected to their first treatment service
was not consistent over time; about 25% of all treatment users (16.6% of sample)
obtained their first treatment service within three months of release from detention and
about 44% of treatment users (29.4% of sample) obtained their first treatment service
within one year of release from detention. The remaining treatment users obtained their
first treatment at increasingly larger distances from detention release, including DAs who
utilized their first service approximately 13½ to 14 years following first stay in detention.
The fact that a sizable percentage of DAs obtained treatment services shortly after
detention release suggests that providers within the Marion County juvenile detention
center are putting forth good efforts to identify and connect DAs to treatment services in
the community. This appears to be particularly true for the time period following
implementation of the MAYSI-2, since the prevalence of treatment utilization within one
year of detention release increased from 20.9% pre-implementation of the MAYSI-2 to
50.1% post-implementation of the MAYSI-2. As documented in other research, mental
health screeners like the MAYSI-2 can promote improved awareness among staff
members of adolescents’ mental health needs, enhanced communication between staff
and DAs, increased efforts by staff to obtain mental health services for DAs, and more
efficient use of treatment resources (Aalsma et al., 2014; Williams & Grisso, 2011). It is
likely that the implementation of the MAYSI-2 in 2006 prompted similar improvements
in the functioning of the Marion County juvenile detention center, which resulted in
marked improvements in post-detention treatment utilization among DAs.
While the MAYSI-2 may have improved service delivery, it is important to
consider that about 70% of DAs who completed the MAYSI-2 screened positively and
about 72.6% of DAs met criteria for a disorder, but only 29.0% of the sample obtained
services within one year and only 36.7% obtained services within two years of release

120
from detention. Based on such findings, several important points can be made regarding
the healthcare delivery system for DAs in this study. First, as documented in
participants’ medical records, providers within the community appear to have been
generally successful with screening and identifying DAs with mental and/or substancerelated disorders, since prevalence rates were quite consistent with other studies of DAs
(Fazel et al., 2008; Shufelt & Cocozza, 2006). Additionally, the implementation of the
MAYSI-2 in 2006 appears to have been quite effective, since nearly 97% of all
adolescents admitted into the detention center post-implementation completed the
screener and nearly 68% screened positively, a rate that is similar to other research
(Aalsma et al., 2014; Vincent et al., 2008). In addition, the prevalence and amount of
services post-detention significantly increased post-implementation of the MASYI-2
(although there were other factors that will be discussed in subsequent sections).
Altogether, findings suggest that the juvenile justice facility and healthcare system
involved in this study were generally able to identify detained youth with behavioral
health needs. The proliferation of research and awareness-raising efforts that have taken
place since the late 1990s to highlight the mental health and substance-related concerns
of DAs (Pajer et al., 2007; Yazzie, 2011) likely served as a key factor in the satisfactory
identification of DAs with needs in this study.
Although DAs were correctly identified as having mental health and substancerelated needs, findings nevertheless indicated problems within the juvenile justice system
and healthcare delivery system, in that the identification of DAs did not consistently
result in connection to treatment and/or retention in treatment services. Specifically,
many DAs with disorders failed to obtain treatment within a reasonable amount of time
upon community reentry. Only 38.2% of adolescents with mental disorders, 26.9% of
adolescents with substance-related disorders and 46.9% of comorbid adolescents utilized
treatment within one year of detention release; rates increased to 49.5%, 37.0%, and
55.6% within two years. Such findings suggest gaps in the juvenile justice system for
ensuring connection to care for DAs with behavioral health concerns.
In trying to account for the gaps in the delivery care system, it is possible that
some participants who were diagnosed post-detention may not have presented with

121
serious mental health concerns during their first detention; hence, they were not referred
for post-detention services. When focused on only DAs who were diagnosed prior to
detention and likely presented with mental health concerns during their detention stays,
post-detention treatment utilization among these DAs was about 65.6% within one year
of detention release. Although this prevalence rate is certainly higher than overall rates,
findings still indicate gaps within the juvenile justice system. Staff should have known
that these DAs were in need of treatment, as diagnoses are listed in their records, but
something happened (e.g., lack of referral, refusal to attend treatment) within the delivery
system that interfered with these adolescents obtaining treatment. At best, study findings
indicate that nearly one in three DAs with documented mental health and/or substancerelated concerns were not connected to services within one year of reentry into the
community, which is unfortunate since treatment services have the potential to alleviate
mental health problems and reduce recidivism (Henggeler & Sheidow, 2012; Lipsey,
2009). Numerous researchers and advocates have critiqued the justice system for
inefficient and/or inconsistent methods of referring and connecting DAs to post-detention
treatment services (Braverman & Murray, 2011; Towberman, 1994). Although such
critiques have prompted policy changes to improve service provision (Hoytt et al., 2001),
such as the implementation of mental health screeners like the MAYSI-2, results from
this study indicate continued limitations within the mental healthcare delivery system for
serving the detained adolescent population.

4.6 Aim 1. Quantity of Post-Detention Treatment Services
As mentioned in the Introduction section, the few studies that have examined
mental health treatment among DAs have primarily examined prevalence rates and/or
effectiveness of mental health interventions for detained youth. Little is known about the
actual characteristics of treatment services used by this population. In this regard, the
study produced several new and notable findings worthy of mention. First, DAs
participated in approximately 20 outpatient sessions over a 2-year period, which equaled
an average of 1.31 sessions per month. Approximately 22.5% of these users dropped out
of treatment after 1-3 sessions and 40.6% experienced at least one significant gap of 46-

122
120 days between sessions. Since no study (that I have found) has specifically examined
quantity and/or gaps in services for detained adolescents, these outcomes are difficult to
put into context, although the dropout rate appears to be lower than the 30%-60% dropout
rate found among the general adolescent population (Kazdin & Mazurick, 1994; Oruche,
et al., 2014).
As recommended on the Blueprints for Healthy Youth Development website
(2015), a registry of evidence-based youth development programs designed to promote
the health and well-being of children and adolescents, effective treatments for high-risk
youth (i.e., detained adolescents) require at least 12 to as many as 30 intensive weekly
treatment sessions involving both the adolescent and his/her family (Darnell & Schuler,
2015). When considering the results for total number of outpatient sessions, dropouts,
and gaps between services, approximately 33.5% of all outpatient users utilized services
within two years of detention release that satisfy the basic treatment recommendations
established by Blueprints for Health Youth Development (2015). Further, it is likely that
the one-third of youth who obtained the minimum number and frequency of services may
not have participated in evidence-based treatment services, like multisystemic therapy or
family functional therapy. Thus, it is likely that few DAs obtained treatment that
qualifies as evidence-based (Darnell &Schuler, 2015; Lipsey, 2009); rather, the majority
of treatment users probably terminated treatment early or obtained too few services to
experience significant treatment benefits. Not surprisingly, high rates of recidivism and
re-incarceration were found within the sample; such outcomes may be related to the lowquality and limited quantity of treatment services that DAs obtained.
Given that most DAs likely failed to obtain sufficient, evidence-based treatment,
it is important to consider possible reasons and treatment barriers that may have
prevented DAs from obtaining adequate services. Unfortunately, the majority of
treatment dropouts and gaps between treatment services occurred for no clear reason;
findings could not be explained by placements in detention/prison or inpatient treatment
facilities. One of the most commonly cited explanations within the literature for low
treatment utilization among DAs involves lack of insurance coverage, since many DAs
experience temporary suspensions or lapses in insurance coverage that disrupt availability

123
and continuity of treatment (Aalsma, Blythe, Harezlak, & Rosenman, 2012a; Cuellar et
al., 2005; Gupta et al., 2005). However, all DAs in the study were actively insured in
Medicaid programs at the time of detention release and should have been able to obtain
mental health and/or substance-related services in the community. It is possible that
some participants may have lost insurance coverage at some point during the study, given
the long time frame of the study, coupled with the low socioeconomic status of most DAs
(Sedlak, 2009). Such loss of insurance may partially explain treatment terminations
and/or gaps in care, but the absence of insurance data over time makes it nearly
impossible to draw conclusions about the impact of insurance on treatment patterns for
DAs in this study. More longitudinal research is needed to examine the relationship
between insurance status and patterns of treatment services over time.
Besides insurance, other potential factors that may have impacted treatment
utilization among DAs include lack of referrals to treatment services (as discussed
earlier), poor knowledge about how or where to obtain services, transportation
limitations, lack of financial means to obtain treatment, or residential move outside of
provider network (Abram et al., 2008; Chapman et al., 2006; Kazdin & Mazurick, 1994;
Luk et al., 2001). In addition, interviews with detained youth have shown that DAs often
feel stigmatized for having a mental and/or substance-related disorder, hold negative
attitudes about seeking treatment, and/or inaccurately believe their problems will improve
without the aid of treatment (Abram et al., 2008), thereby preventing them from seeking
post-detention treatment. DAs also report significant difficulties related to trying to
reintegrate back into the community while simultaneously trying to remain actively
involved in treatment (Moore et al, 2013). Many endorse interest in services, but tend to
terminate treatment services for a variety of reasons, including distrust of treatment
providers, feeling disrespected or unfairly judged by treatment providers (especially for
relapses), being unable to find services that fit their busy schedules with school, work,
and taking care of family members, feeling powerless in overly-controlled treatment
programs (e.g., strict substance use rules, no contact with former friends), and lack of
support system/family involvement in services (Moore et al., 2013). Unfortunately,
evidence suggests that adolescents who drop out of treatment are much less likely to seek

124
help for mental health problems in the future and more likely to experience poor
outcomes (Luk et al., 2001; Oruche et al., 2014); thus, it is imperative that efforts be
made to encourage treatment-seeking among DAs, as well as continued engagement in
treatment services.
In response to the challenges that DAs experience in obtaining treatment services,
juvenile probation officers can help DAs address treatment barriers by serving as a source
of information or “gateway providers” (Holloway et al., 2013; Stiffman et al., 2004)
regarding treatment options and available resources; officers can also regularly
communicate with treatment providers in the community to monitor DAs’ progress and
adjust treatment services and/or probation plans accordingly (Howell, 2003; Stiffman et
al., 2004; Wasserman et al., 2008). To illustrate the importance of probation officers, the
Idaho Department of Juvenile Corrections recently redesigned their healthcare delivery
system for juvenile justice-involved youth by establishing stronger linkages and lines of
communication between mental health providers already in the community, county
probation officers, and members of the court (Riley, 2014). The new system yielded a
new management and service planning structure, in which probation officers and judges
took active parts in treatment planning and resource decisions for DAs. As a result, the
Idaho Department of Juvenile Corrections was able to successfully work with a limited
budget to implement state-wide substance use disorder treatment services for their
juvenile justice population. The following quote by Riley (2014) emphasizes the
importance of collaboration between different individuals involved in the juvenile justice
system as key to a good service delivery system: “Partnership with the judicial branch
and with other agencies managing substance use disorder appropriations for target
population has been both advantageous and necessary to maximize the resources
dedicated to service delivery. A partnership with county probation department provides
the best opportunity to identify and authorize appropriate levels of substance use disorder
services for juveniles who are on probation or otherwise served in the county justice
system” (pp. 39).

125
4.7 Aim 2. Group Differences: Demographic Groups
Upon comparing differences in treatment utilization across groups of DAs, study
findings clearly showed that demographic factors play a substantial role in post-detention
treatment utilization among DAs. Whereas prior research has found higher rates of
treatment referrals, treatment utilization, and number of treatment services among female
DAs than male DAs (Hussey et al., 2008; Kataoka et al., 2001; Lopez-Williams et al.,
2006), current findings failed to substantiate this pattern of service utilization. Instead,
prevalence rates were significantly higher among male DAs compared to female DAs
with regards to treatment services prior to detention, during detention, and after
detention. Male DAs obtained a larger number of post-detention substance-related
services and either/both treatment services within two years of detention release and were
connected to community-based treatment within a shorter number of days than their
female counterparts. Male gender was also associated with an increased likelihood of
post-detention utilization of any treatment services, mental health services, substancerelated services, either/both services, and outpatient services. Although the majority of
findings indicated higher treatment utilization among males than females, a few
exceptions were found; male gender was associated with decreased risk of non-outpatient
treatment among cohort two DAs. Male DAs also tended to terminate outpatient services
earlier than females, so that male gender was associated with increased risk of
terminating outpatient treatment, mental health outpatient treatment, and substancerelated outpatient treatment. Further, no differences were found between males and
females on total number of services, number of outpatient sessions per month, length of
inpatient stays, time between treatment services, dropout rates, and gaps in outpatient
sessions.
Altogether, findings are unexpected and inconsistent with other studies that
typically show higher rates of treatment services among females than males in the
juvenile justice system (Kataoka et al., 2001; Teplin et al., 2005; Veysey, 2003). It is
possible that results reflect new, accurate findings regarding gender differences in
treatment services over an extended period of time, since prior studies have generally
focused on treatment utilization prior to detention and/or within a few years of detention

126
release (Dalton et al., 2009; Domalanta et al., 2003; Hussey et al., 2008; Shelton, 2005).
Alternatively, results may stem from timing of diagnosis; males were significantly more
likely to have a diagnosis upon detention entry, whereas females were more likely to be
diagnosed post-detention. Thus, providers within the detention center may have been
more aware of male DAs in need of treatment services and therefore more likely to
provide referrals and/or planning for transition to community services (Aalsma et al.,
2014; Trupin et al., 2004; Williams et al., 2008). However, this explanation does not
account for the fact that females in the sample endorsed significantly higher mental
health needs on the MAYSI-2 and had higher rates of mental disorders (regardless of
timing in relation to detention), so they should have theoretically obtained referrals and
connections to care in the community (Hoeve et al., 2013; Kates et al., 2014).
Although difficult to determine the exact reason for the unanticipated genderrelated findings, the sampling frame used for the study may explain current results.
Specifically, approximately 9800 potential participants were eliminated from the sample
for lacking adequate medical records and insurance data (e.g., private insurance or no
insurance). The eliminated sample consisted of 19% female and 81% male, compared to
the study sample of 37% female and 63% male. Thus, a large number of males were
excluded from the study; these males lacked Medicaid insurance and were probably
unlikely to utilize treatment (Aalsma et al, 2012b). The current sample may therefore
include a disproportionate number of males who are more likely to use treatment than
normal male DAs, resulting in biased and unexpected results that overestimate treatment
utilization among male DAs. Given that this is one of only studies to find higher service
utilization among male DAs, additional studies are certainly needed to replicate and
confirm findings. Finally, it should be noted that female DAs tended to participate in
services for longer periods of time and were less likely to terminate from outpatient
services than male DAs, which suggest that the number of females involved in treatment
may have been lower than males, but these females tended to remain in treatment
services and were probably more likely to benefit from treatment services (Kazdin &
Mazurick, 1994; Luk et al., 2001).

127
Consistent with prior research (Dalton et al., 2009; Novins et al., 1999; Samuel,
2015), White DAs in this study obtained significantly more treatment services than their
Black counterparts. Being White was significantly associated with receiving more predetention services, as well as post-detention services. Prevalence rates were significantly
higher for White DAs compared to Black DAs for all three treatment types, both
treatment settings, and most non-outpatient user intensity levels (except extreme users).
On average, White DAs obtained significantly more substance-related treatment services
and were less likely to drop out of outpatient treatment. Survival analyses for the entire
study time frame revealed that Black youth tended to be connected to services in a shorter
number of days than White youth, but also terminated outpatient treatment earlier than
White youth. This fits well with other research indicating that minority youth are more
likely to drop out of treatment and terminate services prior to treatment completion
(Johnson et al., 2004; Kazdin & Mazurick, 1994).
As addition racial differences, regression analyses revealed that Black race was
significantly associated with decreased likelihood of mental health treatment services and
non-outpatient treatment services within two years post-detention, as well as a decreased
risk of non-outpatient treatment services at any time. Interestingly, race was not a
significant variable for any other Cox regression analyses, except for decreased risk
among cohort two regarding utilization of any treatment services, mental health services,
either/both treatment services, and outpatient services. Despite clear differences in
treatment utilization across racial groups, Black DAs and White DAs experienced similar
mean number of services, average time between treatment services, number of outpatient
services per month, gaps between outpatient sessions, outpatient intensity levels, and
length of inpatient stay.
Overall, study findings fit well with previous research and my meta-analysis
regarding the greater utilization of treatment services among White DAs than minority
DAs (Dalton et al., 2009; Rawal et al., 2004; Shelton, 2005). Since White DAs endorsed
higher needs on the MAYSI-2, averaged more unique mental diagnoses, and boasted
higher rates of mental disorders and comorbidity, it is possible that the behavioral
healthcare delivery system is simply responding to more significant concerns among

128
White DAs (Shufelt & Cocozza, 2006; Teplin et al., 2012), thereby resulting in higher
treatment utilization among White DAs. This may account for the higher prevalence of
non-outpatient services among White DAs and decreased likelihood of Black DAs
obtaining non-outpatient services post-detention, since DAs with more serious mental
health needs (i.e., White DAs) would require more intensive services like long-term
inpatient treatment. However, Black DAs had higher rates of substance-related disorders
and averaged more substance-related disorders than White DAs. If treatment services
were truly being equally distributed based on need and regardless of race, results would
be expected to follow a pattern in which White DAs obtained more mental health and
comorbid treatment services, whereas Black DAs obtained more substance-related
treatment services. Clearly, results failed to follow this pattern and suggest a possible
bias against Black adolescents.
Current findings point to treatment disparities favoring White adolescents that are
likely a combination of various factors, such as differences in treatment availability,
treatment-seeking behaviors, and attitudes about treatment services among White DAs
versus Black DAs (Brannan & Heflinger, 2005). Evidence suggests that minority
adolescents are often less aware and less willing to endorse mental health problems, less
likely to trust mental health providers, less likely to seek treatment, more likely to drop
out of treatment services, and/or less likely to believe in the effectiveness of treatment
interventions (Abram et al., 2008; Cauffman, 2004; Johnson et al., 2004; Lo et al., 2003).
Thus, the Black DAs in the current study may have been experiencing similar mental
health concerns as the White DAs, but were less likely to obtain treatment for any (or all)
of the reasons just mentioned. In addition to racial differences in adolescent behavior,
racial biases among staff and providers can result in differential treatment regarding who
is identified, referred, and/or connected to services (Herz, 2001; Towberman, 1994), so
that minority DAs are less likely than White DAs to be referred to services, placed in
mental health treatment, or assisted with managing treatment barriers (Dalton et al., 2009;
Hoeve et al., 2013). Moreover, another outcome of racial biases is that Black adolescents
are more likely to be re-arrested and/or re-incarcerated than White adolescents (Office of

129
Juvenile Justice and Delinquency Prevention, 2013; Rozie-Battle, 2002), which limits the
opportunities for Black DAs to obtain treatment in the community (Herz, 2001).
Just as treatment utilization differed significantly across gender groups and racial
groups, findings for treatment utilization across the three age cohorts were quite
disparate. The majority of post-detention treatment utilization followed a pattern in
which treatment services were most prevalent among younger DA (11-13 years),
followed by mid-aged DAs (14-15 years), and lastly older DAs (16-18 years).
Prevalence rates followed this pattern for post-detention treatment services, mental health
services, either/both services, and outpatient services within two years of release from
detention. On average, younger DAs obtained the highest number of treatment services;
mid-age DAs obtained the middle number of services and older DAs obtained the
smallest number of such services. Interestingly, dropout rates were significantly higher
among older DAs than younger DAs, whereas gaps between outpatient services were
more common among younger DAs than older DAs. This relationship makes sense when
considering that younger DAs were more likely to be involved in outpatient treatment
and remained in outpatient treatment for a significantly longer period of time than midage DAs and older DAs, so they had more opportunities to experience gaps between
outpatient sessions. Not surprisingly, older age was linked to an increased risk of
termination from all three types of outpatient treatment.
Findings highlight the unique relationship between age and treatment utilization,
with decreasing likelihood of treatment services as age upon entry into detention
increases. Although this pattern was fairly clear across findings, it should be noted that
results were non-significant across the three age cohorts for time between treatment
services or average length of inpatient stays. Mid-age DAs were actually connected to
post-detention services in a shorter number of days than younger DAs and older DAs. In
addition, age was associated with odds ratios in the opposite direction for the logistic
regression analyses, in that older age was associated with significantly decreased
likelihood of post-detention treatment services, mental health services, and outpatient
services, but increased likelihood of substance-related services and either/both treatment

130
services within two years post-detention. However, survival analysis for the entire study
time frame showed that older age was linked to decreased risk of treatment services.
Altogether, the pattern of results showing decreasing prevalence and quantity of
treatment services as age increases is compatible with other research studies that have
examined the role of age in treatment utilization among detained youth (Aalsma et al.,
2012b; Lopez-Williams et al., 2006). Several factors may account for the negative
relationship between age and treatment utilization. First, the longitudinal nature of this
study may be partially driving results. DAs who were younger at first detention may
have remained in the dataset longer (i.e., avoided attrition related to death, moving, going
into adult prison) than older DAs, so they had more opportunities to obtain treatment and
data about treatment utilization may be more complete for the younger cohort than the
older cohort. Second, as detailed in the beginning of the Discussion section, youth who
are younger at first detention are likely to be more serious criminal offenders with
significant mental health symptomology (Aalsma et al., 2012a; Kates et al., 2014). Thus,
higher treatment utilization by younger DAs may stem from higher needs among this
cohort than other age cohorts. Third, research has suggested possible bias among
juvenile justice staff and providers, who may be more likely to identify and/or refer
younger DAs to community-based services due to beliefs that younger adolescents have
more time or motivation than older adolescents to try to reintegrate back into the
community (Stiffman et al., 2004). Unfortunately, as shown by current findings, provider
bias against older adolescents can promote a two-tiered approach within the juvenile
justice system, in which younger offenders are placed on a treatment-focused,
rehabilitation track whereas older offenders are placed on a punishment-focused,
incarceration track (Herz, 2001; Ricks & Louden, 2014).

4.8 Aim 2. Group Differences: Mental Health Groups
Besides demographic variables, mental health variables also significantly impact
post-detention treatment utilization among DAs. With regards to differences between
Massachusetts Youth Screening Instrument-2nd Edition (MAYSI-2) groups, comparing
DAs who completed the screener versus did not complete the screener is analogous to

131
comparing cohort one to cohort two, which was already addressed in an earlier section.
For DAs who responded to the MAYSI-2, findings showed significantly higher
prevalence rates of treatment utilization among DAs with positive screens versus DAs
with non-positive screens for all three treatment types, both treatment settings, and all
non-outpatient user intensity levels. In addition, the positive screen group obtained
significantly more total treatment services, outpatient sessions, and remained involved in
treatment for significantly more days than DAs with non-positive screens. It should be
noted that DAs with positive screens and DAs with non-positive screen utilized services
rather quickly upon release from detention. Specifically, DAs with positive screens
utilized treatment within 93 days and DAs with non-positive screens utilized treatment
within 111 days of detention release; such length of time was markedly shorter than any
other DA group and significantly shorter than the approximately 992 days it took for DAs
from cohort one who did not take the MAYSI-2 to utilize services. Findings provide
additional support for the existence of a timing effect, since the two MAYSI-2 groups
that utilized services within a short timespan were in cohort two, during which efforts
were made to increase connections to community-based care (Aalsma et al., 2012b;
Aalsma et al., 2012b).
Given the results for treatment utilization across MAYSI-2 groups, it seems
reasonable to presume that DAs who screened positively on the mental health screener
utilized more post-detention treatment services than other youths for several reasons.
First, DAs who endorsed behavioral health concerns were probably more aware of their
treatment needs and more likely to be interested in services to address these needs
(Braverman & Murray, 2011; Lo et al., 2003). Second, due to the implementation of the
mental health screening and referral program at the Marion County juvenile detention
center, staff were able to review MAYSI-2 results, identify youth with positive screens,
and plan appropriate services for these identified adolescents (Aalsma et al., 2012a;
Wasserman et al., 2003). The combination of both these factors likely explains the
higher prevalence and quantity of treatment services utilized among DAs with positive
screens than non-positive, as well as shorter time frames for obtaining treatment upon
reentry into the community.

132
Although screening positively on the MAYSI-2 was associated with higher
prevalence of treatment utilization, results revealed an unanticipated pattern in which
DAs with positive screens obtained more substance-related treatment services and DAs
with non-positive screens obtained more mental health treatment services. For example,
the positive screen group obtained greater number of substance-related outpatient
sessions per month and longer substance-related inpatient stays, whereas the non-positive
screen group obtained a greater number of mental health treatment services, mental
health outpatient sessions per month, and longer mental health inpatient stays. DAs who
screened positively experienced fewer mean days between substance-related outpatient
sessions, whereas DAs with non-positive screens experienced fewer days between mental
health outpatient sessions.
The pattern of greater utilization of substance-related services among the positive
screen group and greater utilization of mental health services among the non-positive
screen group is a new finding that has not been discussed within the literature (that I have
found). Since there are multiple ways to screen positively on the MAYSI-2 via
elevations on different symptom domains, it is not possible to determine whether results
reflect an appropriate match between mental health needs, substance-related needs, and
treatment utilization, i.e., DAs with positive screens were more likely to have substance
use disorders and therefore more likely to obtain substance-related services. Instead, the
best conclusion that can be currently offered is that the implementation of the MAYSI-2
prompted increased treatment utilization among all youth, with generally larger
utilization among DAs with positive screens than non-positive screens. Such increased
utilization represents improvements within the juvenile justice system and healthcare
delivery system (Hendrix, Doebbeling, & Aalsma, 2012; Aalsma et al., 2014), although
the screener does not appear to be being utilized as efficiently as possible for matching
different types of needs (mental health versus substance-related) to corresponding types
of services.
Treatment utilization prior to first detention stay emerged as a significant variable
related to post-detention treatment services upon reentry into the community. Predetention treatment utilization was not only strongly correlated with post-detention

133
treatment utilization, but the prevalence rates for post-detention treatment utilization were
about 3 times larger among DAs with a history of prior treatment than other DAs. For
example, rates were 72.1% for post-detention treatment and 18.0% for post-detention
non-outpatient treatment among DAs who had pre-detention treatment versus 25.8% for
post-detention treatment and 4.8% for non-outpatient treatment among DAs without predetention treatment. Further, DAs with prior treatment obtained significantly more
treatment services, mental health services, either/both services, outpatient services, and
outpatient sessions per month within two years post-detention than their counterparts.
Research examining patterns of treatment-seeking behaviors among the detained
adolescent population is limited (Samuel, 2015; Wasserman et al., 2003), so current
findings are important in showing that DAs engage in similar treatment utilization
behaviors as the general adolescent population. Specifically, studies of treatment-seeking
behaviors and behavioral health services utilization among adolescents have found strong
evidence for continuity of care and associations between past and future treatment (Cauce
& Srebnik, 2003). Evidence suggests that adolescents are more likely to seek mental
health treatment or remain in treatment when they have a history of treatment,
particularly if prior treatment services were helpful, easy to access, and/or affordable
(Abram et al., 2008; Kates et al., 2014). The participants in the current sample showed a
similar pattern; DAs with pre-detention treatment obtained greater number of services,
had higher prevalence rates and lower dropout rates, remained involved in outpatient
treatment longer, and utilized services within a mean of 46 days post-detention, compared
to 1047 days for DAs without pre-detention treatment.
Several factors are likely contributing to the strong association between predetention treatment and post-detention treatment. First, youth with more serious mental
health concerns are more likely to obtain treatment at an early age (Kessler et al., 2005).
Thus, DAs with a history of treatment services are likely to be younger at detention entry,
with prominent mental health and/or substance-related problems that require continued
treatment following release from detention post-detention. Second, adolescents’ history
of treatment services is listed on their records, so staff within the juvenile justice system
should be aware of DAs with prior treatment and make appropriate arrangements for

134
continuity of care (Hendrix et al., 2012; Wasserman et al., 2003). Third, DAs with predetention treatment may be more willing to utilize treatment post-detention, due to
having a strong therapeutic alliance with a provider, wanting to continue seeing a
provider, possessing buy-in or belief in treatment services, experiencing good outcomes
in past treatment, and/or being comfortable and knowledgeable of the therapeutic
treatment process (Orunche et al., 20014; Samuel, 2015). It should be noted that
approximately 87.8% of the DAs with prior treatment obtained these treatment services
within the two years prior to being admitted into detention, so these youth likely
remember their providers, are familiar with treatment, know where to seek treatment,
etc., which eliminates many treatment barriers that may prevent DAs from seeking
treatment after release from detention. Finally, the ease of service planning may account
for the link between pre- and post-detention treatment; it may be easier for staff within
the juvenile justice system to coordinate care with DAs’ current (or past providers) in the
community and make arrangements for these youth to continue services upon release
from detention versus independently searching and finding providers for youth with no
history of prior treatment (Lyons et al., 2003; Riley, 2014).

4.9 Aim 2. Group Differences: Criminal History Groups
Findings revealed that DAs with more serious criminal history background tended
to utilize more treatment services than other DAs. For example, being a violent offender
was significantly correlated with pre-detention treatment, during-detention treatment, and
post-detention treatment utilization. Prevalence rates for violent offenders were
significantly higher than non-violent offenders for any post-detention treatment services,
mental health treatment services, and outpatient treatment services. Violent offenders
were also connected to services in the community significantly sooner and remained
involved in outpatient services longer than non-violent offenders. Moreover, being a
violent offender was associated with an increased risk of post-detention treatment
utilization, as well as decreased risk of terminating from general outpatient treatment.
More serious offenders, such as violent offenders, tend to engage in more
problematic behaviors that ensure heightened attention within the juvenile justice system

135
and the community (Barrett et al., 2006; Bath et al., 2013). As a result, providers are
likely to have increased awareness of such DAs and may make increased efforts to
generate referrals and/or provide services to ensure rehabilitation for these troubled youth
(Hoeve et al., 2013; Kataoka et al., 2001). In addition, mental health problems and
criminal history tend to be highly related (Colins et al., 2013; Maschi et al., 2008). Not
surprisingly, being a violent offender was correlated with having a conduct disorder, nonconduct mental disorder, and substance-related disorder, so being a violent offender may
represent a proxy variable of mental health needs and therefore account for the higher
prevalence of treatment services for these youth.
Much like violent offenders, adolescents who were rearrested within two years of
detention experienced higher treatment utilization than other adolescents. Recidivism
was not only significantly correlated with pre-detention treatment utilization, but also
associated with post-detention treatment utilization for the three treatment types and two
treatment settings. Prevalence rates among recidivist DAs were about 2 times greater
than rates for non-recidivist DAs. Moreover, recidivists averaged significantly more
general treatment services, mental health treatment services, and outpatient sessions;
there were significantly more recidivist DAs within all outpatient user intensity levels and
non-outpatient user intensity levels (except extreme users). With regards to long-term
outcomes, survival analyses showed that recidivists obtained services within a
significantly shorter number of days and remained involved in outpatient services for a
longer number of days than non-recidivist DAs. This likely explains why dropout rates
were smaller for recidivists than non-recidivists, since they obtained larger number of
services and stayed involved in services longer than non-recidivists.
Comparisons between the incarceration groups mirrored the results for recidivism
groups, with higher treatment services among adolescents who were re-detained and/or
incarcerated in prison. Being re-detained/incarcerated was significantly associated with
treatment utilization prior to detention, during detention, and after detention. Prevalence
rates were significantly higher for re-detained/incarcerated DAs for all three treatment
types and both treatment settings, although rates were not as divergent as recidivism
groups. In contrast to other findings, the survival curves for incarceration groups failed

136
to significantly differ, meaning that DAs for both groups obtained services within similar
periods of time; however, re-detained/incarcerated DAs remained involved in outpatient
treatment significantly longer than other DAs. Analogous to findings for recidivism,
youth who were re-detained/incarcerated were less likely to drop out of outpatient
treatment, but more likely to be able to explain dropouts via placements in
detention/prison.
Taken together, results for recidivism and re-detained/incarcerated highlight
several key findings. Because the specific date of recidivism was not available for this
study, this variable was not examined as a potential predictor of treatment utilization;
dates for detention/prison were available and included in the survival analyses. Being redetained/incarcerated prior to treatment increased the likelihood of any post-detention
treatment services, mental health treatment services, substance-related treatment services,
either/both treatment services, and outpatient services. Such findings correspond to prior
work (Kataoka et al., 2001; Kates et al., 2014; Lopez-Williams, et al., 2006) and seem
reasonable since increased involvement in the juvenile justice system equates to more
contacts with the system to be correctly identified as having behavioral health concerns,
referred for community services, and/or connected to appropriate treatment programs.
Additionally, providers in the system and in the community may be more motivated to
help DAs who are stuck in the revolving door of the juvenile justice via recidivism and/or
re-detentions by getting them engaged in treatment services (Becan et al., 2014).
A second key finding is that approximately 82.8% of recidivist DAs obtained
treatment prior to re-arrest and 66.1% of incarcerated DAs obtained treatment prior to
being placed in detention/prison. Given that the primary goals for many mental health
interventions for detained youth involves reduced delinquency, the treatment services
DAs received do not appear to have been particularly effective in promoting
rehabilitation and preventing recidivism. Such results are not entirely surprising since the
majority of DAs obtained services that fell below standards for evidence-based treatment
(as discussed earlier). At the same time, however, it is possible that treatment services
delayed the onset and/or intensity of recidivism, which is considered a positive treatment
outcome (Cuellar et al., 2006; Schwalbe et al., 2012). While possible, it is more likely

137
that the strong association between mental health treatment and criminal activity stems
from the severity of DAs’ behavioral health needs (Aalsma et al., 20015; Kates et al.,
2014). DAs with serious mental and/or substance-related issues are more likely to obtain
treatment, and these DAs may also be more likely to experience a recidivism event or
repeated detention upon reentry into the community (Aalsma et al., 2015; Barrett et al.,
2006). Thus, underlying mental health concerns may represent a third variable that is
driving current findings. Additional data pertaining to recidivism events (e.g., date,
severity, charge count) and treatment services (e.g., treatment model, quality of services,
treatment) are needed to adequately examine these issues and determine the effectiveness
of current treatment services.

4.10 Aim 3. Predictors of Treatment Utilization
Besides violent offenses, recidivism, and incarceration, several other criminal
history variables emerged as significant predictors of treatment utilization among DAs.
Most notably, higher number of arrests prior to detention was associated with increased
likelihood of post-detention treatment services, mental health services, substance related
services, either/both services, outpatient services, and non-outpatient services at 2-year
follow-up, as well as any time during the study time frame. Higher number of prior
arrests was also linked to significantly increased risk of dropping out from any outpatient
treatment, including all three types of treatment services. Additionally, higher charge
severity (1 to 5) yielded an increased likelihood of any post-detention treatment services,
mental health treatment services, and outpatient treatment services; more serious charge
severity also increased the risk of terminating outpatient treatment. Overall, results
correspond well to prior studies linking increased criminal history to greater treatment
utilization (Barrett et al., 2006; Johnson et al., 2004; Ryan et al., 2013). As outlined in
previous paragraphs, DAs with more significant criminal backgrounds (i.e. larger number
of prior arrests and/or more serious charges) may be we more likely to have behavioral
health concerns, particularly impaired emotional regulation, impulsivity, poor impulse
control, and conduct problems (Kates et al., 2014), and may therefore be more likely to
obtain treatment services upon release into the community.

138
Interestingly, comparing cohort one to cohort two yielded some unexpected
findings related to the relationship between criminal history and treatment utilization.
First, the number of prior arrests failed to predict treatment utilization for cohort two,
with the exception of non-outpatient treatment and outpatient dropout. Second, charge
severity was significant for cohort two, but failed to predict post-detention treatment
services for cohort one. Third, length of stay emerged as a significant variable associated
with decreased risk of termination from either/both outpatient treatment for cohort one,
but increased risk of termination from the same treatment for cohort two. Longer length
of stay also produced a decreased likelihood of post-detention treatment for cohort one
compared to increased likelihood of post-detention outpatient treatment for cohort two.
Clearly, the time effect is impacting results, so that findings are different between
cohorts. The fact that most criminal history variables were significant for the total
sample and cohort one, but became non-significant (or less significant) for cohort two,
suggests a trend in the right direction because criminal history variables are making less
of an impact on whether DAs obtain services. At the same time, behavioral health needs
and diagnoses are becoming more important factors in treatment utilization among DAs.

4.11 Time Effect
During data analysis, evidence emerged suggesting a possible time effect that
occurred during the study time frame. Research revealed that several key events occurred
between 2005 and 2006 that impacted the functioning of the Marion County juvenile
detention center (i.e., site of this study). Specifically, Marilyn Moore was appointed to
serve as head judge for the Marion County Superior Court, Juvenile Division in March
2005 (Maguire, 2012). In contrast to Judge James Payne, who previously served as the
head judge for 20 years, Judge Moore pushed for more emphasis on the behavioral health
needs and employment possibilities for juvenile justice-involved youth (Bercovitz, 2009).
Thus, under her leadership, the Marion County Superior Court invited the Anne E. Casey
Foundation to implement the Juvenile Detention Alternatives Initiative (JDAI) in 2006
within Marion County (Bercovitz, 2009). The JDAI, which is a nationwide program
designed to reduce juvenile detention and increase the use of alternatives non-detention

139
programs (Bercovitz, 2009), resulted in a reduction in the number of adolescents detained
in Marion County’s juvenile detention center (i.e., census reduction from 144 beds to 112
beds) and the establishment of a mental health screening and referral program (Aalsma et
al., 2012a; Aalsma et al., 2012b). The program required that all adolescents complete the
MAYSI-2 upon detention entry and established procedures to guide staff in
understanding screener results and taking appropriate actions (e.g., referrals, request
comprehensive psychological evaluation, create transition plan) (Aalsma et al., 2012b).
Given that the events discussed likely impacted study outcomes and introduced a serious
time effect (Roe & Korn, 1993), I conducted sub-analyses comparing findings for DAs
detained pre-implementation of the screening and referral program (i.e., cohort one) to
findings for DAs detained post-implementation (i.e., cohort two).
Since an examination of this time effect was not initially proposed as a primary
study aim, discussion of findings related to the time effect will be brief. It appears the
events that occurred during 2005 and 2006 greatly impacted treatment utilization
outcomes (Aalsma et al., 2012b), as evidenced by markedly divergent results between
cohorts. Prevalence rates were roughly 2-3 times larger among cohort two DAs
compared to cohort one DAs, with significantly more cohort two DAs obtaining mental
health services, substance-related services, either/both services, outpatient services, nonoutpatient services, all outpatient user intensity levels, and most non-outpatient user
intensity levels. Moreover, predictors of treatment utilization differed between cohorts.
Many criminal history variables (e.g., number of prior arrests, violent offender) that were
significant for cohort one became non-significant (or less significant) for cohort two,
while many mental health variables (e.g., conduct disorder, substance-related disorder)
remained significant for cohort one.
Altogether, the significant differences that emerged between cohorts support the
conclusion that a time effect was present during the follow-up period of the study, with
unique events that prompted greater utilization of treatment services among adolescents
detained in more recent years. Findings suggest a positive trend over time, in which the
prevalence of treatment utilization is growing and mental health needs are becoming
increasingly important factors (rather than criminal history) in predicting post-detention

140
treatment utilization (Lyons et al., 2001; Vincent et al., 2008). This trend is not perfect,
however, given that male gender, Black race, and older age were associated with
decreased likelihood of any treatment services, mental health treatment services,
either/both treatment services, and outpatient services among DAs in cohort two; such
disparities significantly limit the opportunities for certain groups of DAs to obtain
treatment services (Aarons et al., 2004; Herz, 2001; Hoytt et al., 2001).

4.12 Study Limitations
Several limitations associated with this study should be acknowledged. First, the
sample for the study consisted of adolescents from the detention center in Marion
County, Indiana. Results may not generalize to the overall detained adolescent
population, since adolescents from long-term prison facilities and facilities outside
Marion County were not part of the sample. However, the sample size was large, a
sizable minority of adolescents (about 10%-20%) in the sample experienced at least one
stay in prison, and the demographic makeup of sample is consistent with census data
pertaining to the overall detained adolescent population (Sedlak & Bruce, 2010); thus,
issues related to generalizability are minimal.
Second, although the Recmatch software program was able to link approximately
83% of the files from the Marion County Juvenile Justice system (QUEST) with
Regenstrief Medical Record System (RMRS), there is no guarantee that all adolescents’
records were accurately matched to the correct individual. Multiple steps were taken to
enhance the matching process, including the use of numerous participant identifiers (e.g.,
SSN, name, DOB) and a secondary software program (i.e., RecFarm) that specifically
analyzed pairs of files to identify true matches and eliminate duplicate (or non-matching)
files. However, it is likely that some participants’ files were either incorrectly matched
and/or discarded, resulting in missing and/or biased data.
One way to address this issue involved eliminating participants with files
containing significant amounts of missing data (i.e., either all mental health information
or all criminal history information). Unfortunately, this created a limitation involving the
sampling frame, particularly the exclusion of a significant number of adolescents. Prior

141
to analysis, approximately 2,126 adolescents were eliminated from the sample due to
either an inability to link records from both databases or blank records from one (or both)
of the databases. Further, approximately 7,734 adolescents with private insurance or no
insurance were dropped from the sample due to a lack of treatment utilization data, and
approximately 749 adolescents were lost to attrition (e.g., death, lack follow-up) at 2-year
follow-up. The adolescents included in the study may be systematically different than the
adolescents who were excluded from the study, but it is not possible to determine the
nature of such bias. Given the low socioeconomic status within the DA population
(Sedlak, 2009) and evidence that lack of insurance can serve as a treatment barrier
preventing adolescents from obtaining treatment (Abrams, et al., 2008), it is possible that
current study findings may have been different with the inclusion of non-insured DAs.
Future research that assesses the relationship between insurance status and treatment
utilization among the DA population is needed to replicate current study findings.
Finally, the large number of statistical tests conducted for this study introduces
potential problems with alpha inflation and heightened risk of Type 1 errors (Altman,
2000). As mentioned in the Introduction section, this study represents one of the first and
only longitudinal research studies to thoroughly assess treatment utilization patterns
among DAs. Thus, the study was exploratory in nature, guided by four primary study
aims with multiple analyses to fully examine the data and better understand the topic
area. To avoid missing potentially important findings, an alpha correction was not used.
Rather, the statistical significance level for most analyses was set at p ≤ .01, with the
expectation that Type 1 errors may occur for this exploratory study. I recognize this as a
potential limitation, with the recommendation that additional research aim to replicate
and confirm findings.

4.13 Implications and Recommendations for Future Research
The study revealed unique patterns of service utilization over time, marked by
low prevalence rates (in comparison to disorder rates), early termination of treatment
services, disparities across certain groups of DAs, and differences between utilization of
mental health services versus substance-related services. As one of the first and only

142
longitudinal studies to examine service utilization among a large sample of detained
adolescents, study findings serve as a crucial foundation for future research. In light of
such findings, I present the following recommendations to researchers and public policy
makers within the mental healthcare field and the juvenile justice system.
1. Juvenile justice facilities are not required to maintain accreditation with National
Commission on Correctional Health Care guidelines and therefore tend to provide
limited, low-quality treatment services (Braverman & Murray, 2011). Legislation
is needed that requires all facilities to maintain national standards of behavioral
healthcare. Furthermore, an independent group of evaluators should regularly
monitor and evaluate services within correctional facilities, to ensure facilities are
providing quality treatment. Ideally, such efforts should improve the quantity, as
well as quality, of mental health and substance-related services within facilities
(Mulvey et al., 2007).
2. Epidemiology research has largely focused on the prevalence of mental health and
substance-related concerns among DAs (Fazel et al., 2008; Shufelt & Cocozza,
2006 Teplin et al., 2002). While such research has greatly enhanced
understanding of common symptoms and diagnoses among DAs, researchers need
to take the next step in determining how to best address these problems (Johnson
et al., 2004). Juvenile justice facilities are encouraged to enact programs for
consistent identification of mental health needs for all detained youth (Aalsma et
al., 2014; Desai et al., 2006). Specifically, facilities should institute formal
mental health screenings during intake, if not doing so already. Results from
these screenings should be used to identify treatment needs and prompt
appropriate referrals and treatment planning for mental health and substancerelated services in the community (Hendrix et al., 2012; Wasserman et al., 2003).
3. Professionals within the academic community and juvenile justice field need to
implement policies and procedures to improve the behavioral healthcare delivery
system for DAs to ensure that detained youth obtain treatment referrals,
connections to care, resources to manage treatment barriers, support with
remaining in treatment (rather than terminating early), as well as appropriate

143
evidence-based services. Given that many DAs tend to become stuck in the
revolving door of the juvenile justice system, with multiple stays in correctional
facilities, an ongoing monitoring process is crucial for helping DAs maintain
continuity of care, address treatment obstacles, and achieve positive treatment
outcomes (Bullis et al., 2004; Wasserman et al., 2003).
4. Different parties involved in the juvenile justice system, such as probation
officers, judges, detention center staff, police officers, and community providers,
should take part in service planning for DAs with mental health needs (Holloway
et al., 2013; Stiffman et al., 2004). A system based upon regular communication
and coordination of crucial information and resources related to treatment
services, such as the one implemented by the Idaho Department of Juvenile
Corrections, is recommended (Riley, 2014).
5. Few evidence-based mental health interventions and substance-related
interventions exist for treating the detained youth population (Desai et al., 2006;
Hoeve et al., 2013). Researchers are advised to advance the development,
implementation, and dissemination of evidenced-based treatments that not only
address the mental health concerns of DAs, but also produce reductions in
recidivism (Braverman & Murray, 2011; Schwalbe et al., 2012). Such
interventions should be affordable, convenient, and widely-available, given that
the DA population tends to have low socioeconomic status and poor access to
treatment services (Rogers et al., 2009; Sedlak & Bruce, 2010).
6. As part of an effective healthcare delivery system for DAs, all referrals, treatment
planning, and actual treatment services must be provided equally across groups
(Dalton et al., 2009; Herz, 2001). Public policy makers are advised to implement
mechanisms that ensure all DAs have an equal opportunity to utilize high-quality
mental healthcare services, regardless of demographic status, insurance status,
mental health status, or criminal history status. The program outlined by Lyons
and colleagues (2003) provides a good example of intensive efforts to connect
DAs to appropriate mental health services. For this program, a designated mental
health liaison worked with staff and providers within the juvenile justice system

144
and mental healthcare system to coordinate care for DAs with behavioral health
concerns. The program was incredibly successful, boasting a post-detention
service utilization rate of 75%. It is recommended that the juvenile justice system
adopt similar programs nationally, as a first step toward addressing and increasing
mental health and substance-related treatment utilization among the detained
adolescent population.

8

REFERENCES

145

REFERENCES

Aalsma, M. C., Blythe, M. J., Harezlak, J., & Rosenman, M. B. (2012a). Insurance status
of urban detained adolescents. Journal of Correctional Healthcare, 18(4), 274284.
Aalsma, M. C., Schwartz, K., & Perkins, A. (2014). A state-wide collaboration to initiate
mental health screening and assess services for detained youths in Indiana.
Research and Practice, 104(10),e82-e 88.
Aalsma, M. C., Tong, Y., Lane, K., Katz, B., & Rosenman, M. B. (2012b). Use of
outpatient care by juvenile detainees upon community reentry: Effects of mental
health screening and referral. Psychiatric Services.
Aalsma, M. C., White, L. M., Lau, K. S., Perkins, A., Monahan, P., & Grisso, T. (2015).
Behavioral health care needs, detention-based care, and criminal recidivism at
community reentry from juvenile detention: A multisite survival curve analysis.
American Journal of Public Health, 10, e1-e7.
Aarons, G. A., Brown, S. A., Garland, A. F., & Hough, R. L. (2004). Race/ethnic
disparities and correlates of substance-abuse service utilization and juvenile
justice involvement among adolescent substance use disorders. Journal of
Ethnicity in Substance Abuse, 3(1), 47-64.
Abram, K. M., Teplin, L. A., McClelland, G. M., & Dulcan, M. K. (2003). Comorbid
psychiatric disorders in youth in juvenile detention. Archives of General
Psychiatry, 60(11), 1097.
Abram, K. M., Paskar, L. D., Washburn, J. J., & Teplin, L. A. (2008). Perceived barriers
to mental health services among youths in detention. Journal of the American
Academy of Child and Adolescent Psychiatry, 47(3), 301-308.

146
Allison, P. (2013). What’s the best R-squared for logistic regression. Statistical Horizons
[website]. Retrieved from http://statisticalhorizons.com/r2logistic. Last accessed
on June 22, 2015.
Altman, D. (2000). A review of experimentwise Type I error: Implications for univariate
post hoc and for multivariate testing. Retrieved from ERIC Clearinghouse. Last
accessed on October 4, 2015.
Barrett, B., Byford, S., Chitsabesan, P., & Kenning, C. (2006). Mental health provision
for young offenders: Service use and cost. The British Journal of Psychiatry,
188(6), 541-546.
Bath, E., Clark, L. O., & Low, J. Y. (2013). Treatment Issues for Youths with Substance
Abuse in Juvenile Detention. Clinical Handbook of Adolescent Addiction, 405415.
Becan, J. E., Knight, D. K., Crawley, M. S., Joe, G. W., & Flynn, P. M. (2014).
Effectiveness of the treatment readiness and induction program for increasing
adolescent motivation for change. The Journal of Substance Abuse Treatment,
50, 38-49.
Bercovitz, J. (2009, October 31). Juvenile Detention Alternative Initiative in Indiana.
Indiana Court Times. Retrieved from
http://indianacourts.us/times/2009/10/juvenile-detention-alternatives-initiative-inindiana/. Last accessed on June 17, 2015.
Bostwick, L., & Ashley, J. (2009). Examining At-Risk and Delinquent Girls in Illinois.
Illinois: Illinois Criminal Justice Information Authority.
Bouliotis, G., & Billingham, L. (2011). Crossing survival curves: Alternatives to the logrank test. Trials, 12(1), A137-138.
Brannan, A. M., & Heflinger, C. A. (2005). Child behavioral health service use and
caregiver strain: Comparison of managed care and fee-for-service Medicaid
systems. Mental Health Services Research, 7(4), 197-211.
Braverman, P. K., & Murray, P. J. (2011). Health Care for Youth in the Juvenile Justice
System. Pediatrics, 128(6), 1219-1235.

147

Bullis, M., Yovanoff, P., Mueller, G., & Havel, E. (2002). Life on the" outs"—
Examination of the facility-to-community transition of incarcerated youth.
Exceptional Children, 69(1), 7-22.
Bullis, M., Yovanoff, P., & Havel, E. (2004). The importance of getting started right:
Further examination of the facility-to-Community transition of formerly
incarcerated youth. The Journal of Special Education, 38(2), 80-94.
Bureau of Justice Statistics (2015). The justice system. What is the sequence of events in
the criminal justice system? [Government website]. Retrieved from:
http://www.bjs.gov/content/justsys.cfm#juvenile. Last accessed on June 17, 2015.
Calley, N.G. (2012). Juvenile offender recidivism: An examination of risk factors.
Journal of Child Sexual Abuse, 21, 257-272.
Cauce, A. M., & Srebnik, D. (2003). Before Treatment: Adolescent Mental Health HelpSeeking. Prevention Researcher, 10(4), 6-9.
Cauffman, E. (2004). A statewide screening of mental health symptoms among juvenile
offenders in detention. Journal of the American Academy of Child and Adolescent
Psychiatry, 43(4), 430-439.
Cauffman, E., Scholle, S. H., Mulvey, E., & Kelleher, K. J. (2005). Predicting first time
involvement in the juvenile justice system along emotionally disturbed use
receiving mental health services. Psychological Services, 2(1), 28-38.
Center for Disease Control and Prevention (2013). Classification of diseases, functioning,
and disability. [Government Website]. Retrieved from
http://www.cdc.gov/nchs/icd.htm. Last accessed on May 8, 2015.
Cloyes, K. G., Wong, B., Latimer, S., & Abarca, J. (2010). Time to prison return for
offenders with serious mental illness released from prison. Criminal Justice and
Behavior, 37,(2), 175-187.
Colins, O. F., Boonmann, C., Veenstra, J., van Domburgh, L., Buffing, F., Doreleijers, T.
H., & Vermeiren, R. M. (2013). Mental health problems and recidivism among
detained male adolescents from various ethnic origins. European Child and
Adolescent Psychiatry, 22 (8), 486-490.

148

Colins, O., Vermeiren, R., Vreugdenhil, C., van den Brink, W., Doreleijers, T., &
Broekaert, E. (2011). Psychiatric disorders in detained male adolescents: a
systematic literature review. Canadian Journal of Psychiatry. Revue Canadienne
De Psychiatrie, 55(4), 255-263.
Corning, A. F., & Malofeeva, E. V. (2004). The application of survival analysis to the
study of psychotherapy termination. Journal of Counseling Psychology, 54(3),
354-367.
Cottle, C. C., Lee, R. J., & Heilbrun, K. (2001). The prediction of criminal recidivism in
juveniles: A meta-analysis. Criminal Justice and Behavior, 28(3), 367-394.
Cuellar, A. E., Kelleher, K. J., Rolls, J. A., & Pajer, K. (2005). Medicaid insurance policy
for youths involved in the criminal justice system. Journal Information, 95(10),
1707-1714.
Cuellar, A., McReynolds, L. S., & Wasserman, G. A. (2006). A cure for crime: Can
mental health treatment diversion reduce crime among youth? Journal of Policy
Analysis and Management, 25(1), 197-214.
Curtin, M. (2002). Lesbian and bisexual girls in the juvenile justice system. Child &
Adolescent Social Work Journal, 19(4), 285-301
Dalton, R. F., Evans, L. J., Cruise, K. R., Feinstein, R. A., & Kendrick, R. F. (2009).
Race Differences in Mental Health Service Access in a Secure Male Juvenile
Justice Facility. Journal of Offender Rehabilitation, 48(3), 194-209.
Darnell, A. J., & Schuler, M. S. (2015). Quasi-experimental study of family functional
therapy effectiveness for juvenile justice aftercare in the racially and ethnically
diverse community sample. Children and Youth Services Review, 50, 75-82.
Dembo, R., Pacheco, K., Schmeidler, J., Ramirez-Garmica, G., Guida, J., & Rahman, A.
(1997). A further study of gender differences in service needs among youths
entering a juvenile assessment center. Journal of Child & Adolescent Substance
Abuse, 7(4), 49-77.
Desai, R. A., Goulet, J. L., Robbins, J., Chapman, J. F., Migdole, S. J., & Hoge, M. A.
(2006). Mental health care in juvenile detention facilities: A review. Journal of
the American Academy of Psychiatry and the Law Online, 34(2), 204-214

149

Domalanta, D. D., Risser, W. L., Roberts, R. E., & Risser, J. (2003). Prevalence of
depression and other psychiatric disorders among incarcerated youths. Journal of
the American Academy of Child & Adolescent Psychiatry, 42(4), 477-484.
Erdfelder, E., Faul, F., & Buchner, A. (1996). GPOWER: A general power analysis
program. Behavior Research Methods, Instruments & Computers, 28, 1-11.
Fazel S, Doll H, Långström, N. (2008). Mental disorders among adolescents in juvenile
detention and correctional facilities: a systematic review and metaregression
analysis of 25 surveys. Journal of the American Academy of Child and Adolescent
Psychiatry, 47(9), 1010-1019.
Gallagher, C. A., & Dobrin, A. (2007). Can juvenile justice detention facilities meet the
call of the American Academy of Pediatrics and National Commission on
Correctional Health Care? A national analysis of current practices.
Pediatrics, 119(4), e991-e1001.
Getahun, D., Jacobsen, S. J., Fassett, M. J., Chen, W., Demissie, K., Rhoads, G. G.
(2013). Recent trends in childhood attention-deficit/hyperactivity disorder.
Journal of the American Medical Association, 167(3), 282-288.
Gordon, J. A., Diehl, R. L., & Anderson, L. (2012). Does ADHD Matter? Examining
Attention Deficit and Hyperactivity Disorder on the Likelihood of Recidivism
among detained youth. Journal of Offender Rehabilitation, 51(8), 487-518.
Grannis, S. J., Overhage, J. M., Hui, S., & McDonald, C. J. (2003). Analysis of a
probabilistic record linkage technique without human review. In AMIA Annual
Symposium Proceedings (Vol. 2003, p. 259). American Medical Informatics
Association.
Grisso, T., Barnum, R., Fletcher, K. E., Cauffman, E., & Peuschold, D. (2001).
Massachusetts Youth Screening Instrument for mental health needs of juvenile
justice youths. Journal of the American Academy of Child and Adolescent
Psychiatry, 40(5), 541-548.
Grisso, T., & Barnum, R. (1998). Massachusetts youth screening instrument (MAYSI):
Preliminary manual and technical report. University of Massachusetts Medical
School, Worcester, MA, b15.

150

Grunwald, H.E, Lockwood B., Harris, P.W., & Mennis, J. (2010). Influences of
neighborhood context, individual history and parenting behavior on recidivism
among juvenile offenders. Journal of Youth and Adolescence, 39(9), 1067-1079.
Gupta, R. A., Kelleher, K. J., Pajer, K., Stevens, J., & Cuellar, A. (2005). Delinquent
youth in corrections: Medicaid and reentry into the community. Pediatrics,
115(4), 1077-1083.
Harrison, L. D. (2001). The revolving prison door for drug-involved offenders:
Challenges and opportunities. Crime & Delinquency, 47(3), 462-485.
Henggeler, S., & Schoenwald, S. J. (2011). Evidence-based interventions for juvenile
offenders and juvenile justice policies that support them. Social Policy Report,
25(1), 1-20.
Henggeler, S. W., & Sheidow, A. J. (2012). Empirically supported-family based
treatments for conduct disorder and delinquency in adolescents. Journal of
Marital and Family Therapy, 38(1), 30-58.
Herz, D. C. (2001). Understanding the Use of Mental Health Placements by the Juvenile
Justice System. Journal of Emotional and Behavioral Disorders, 9(3), 172-181.
Hendrix, K. S., Doebbeling, C. C., & Aalsma, M. C. (2012). Psychological and
neuropsychological assessment in the juvenile justice system: Recommendations
for protocols. Criminal Justice Studies, 25 (3), 239-249.
Holloway, E. D., Brown, J. R., Suman, P. D., & Aalsma, M. C. (2013). A qualitative
examination of juvenile probation officers as gateway providers to mental
healthcare. Criminal Justice Policy Review, 24(3), 370-392.
Hoeve, M., McReynolds, L. S., & Wasserman, G. A. (2013). Service referral for juvenile
justice youths: Associations with psychiatric disorder and recidivism.
Administration and Policy in Mental Health and Mental Health Services
Research, 1-11.
Hoskins, M. W. (2008, May 14). Marion County: A model for juvenile detention reforms.
The Indiana Lawyer.com. Retrieved from
http://www.theindianalawyer.com/marion-county-a-model-for-juvenile-detentionreforms/PARAMS/article/24024. Last accessed on June 29, 2015.

151

Howell, J.C. (2003). Preventing and reducing juvenile delinquency: A comprehensive
framework. Thousand Oaks, CA: Sage.
Hussey, D. L., Drinkard, A. M., & Flannery, D. J. (2007). Comorbid substance use and
mental disorders among offending youth. Journal of Social Work Practice in the
Addictions, 7(1-2), 117-138.
Hussey, D. L., Drinkard, A. M., Falletta, L., & Flannery, D. J. (2008). Understanding
clinical complexity in delinquent youth: Comorbidities, service utilization, cost,
and outcomes. Journal of Psychoactive Drugs, 40(1), 85-95.
Hoytt, E. H., Schiraldi, V., Smith, B. V., & Zeidenberg, J. (2001). Reducing racial
disparities in juvenile detention. Annie E. Casey Foundation.
Isaacs, M. R. (1992). The impact of community violence on African American children
and families: Collaborative approaches to prevention and intervention: workshop
summary. The Impact of Community Violence on African American Children And
Families: Collaborative Approaches To Prevention And Intervention: Workshop
Summary.
Johnson, T. P., Cho, Y. I., Fendrich, M., Graf, I., Kelly-Wilson, L., & Pickup, L. (2004).
Treatment need and utilization among youth entering the juvenile corrections
system. Journal of Substance Abuse Treatment, 26(2), 117-128.
Jones, D. L., Heflinger, C. A., & Saunders, R. C. (2007). The ecology of adolescent
substance abuse service utilization. American Journal of Community Psychology,
40(3-4), 345-358.
Kataoka, S.H., Zima B.T., Dupre D.A., Moreno K.A., Yang X., & McCracken, J. (2001).
Mental health problems and service use among female juvenile offenders: their
relationship to criminal history. Journal of the American Academy of Child and
Adolescent Psychiatry, 40(5), 549-555.
Kates, E., Gerber, E.B., & Casey, S. (2014). Prior service utilization in detained youth
with mental health needs. Administration Policy and Mental Health, 41,86-92.
Kazdin, A. E., & Mazurick, J. L. (1994). Dropping out of child psychotherapy:
Distinguishing early and late drop-outs over the course of treatment. Journal of
Consulting and Clinical Psychology, 62(5), 1069-1077.

152

Kessler, R. C., Berglund, P., Demler, O., Jin, R., Merikangas, K. R., & Walters, E. E.
(2005). Lifetime prevalence and age-of-onset distributions of DSM-IV disorders
in the National Comorbidity Survey Replication. Archives of General
Psychiatry, 62(6), 593-602.
Latessa, E., Lovins, B., Makarios, M. (2013). Validation of the Indiana risk assessment
system final report. Unpublished technical report. Retrieved from
http://www.in.gov/judiciary/cadp/files/prob-risk-iras-final.pdf. Last accessed on
June 17, 2015.
Liddle, H. A., Dakof, G. A., Henderson, C., & Rowe, C. (2011). Implementation
outcomes of multidimensional family therapy-detention to community: A
reintegration program for drug-using juvenile detainees. International Journal of
Offender Therapy And Comparative Criminology, 55(4), 587-604.
Lipsey, M. W. (2011). Effective interventions for juvenile offenders. Using metaanalysis to bridge from research to practice. In T. Bliesener, A. Beelmann, & M.
Stemmler. (Eds.). Antisocial behavior and crime: Contributions of development
and evaluation research to prevention and intervention. (pp.181-198).
Cambridge,, MA. Hogrefe Publishing.
Lipsey, M. W. (2009). The primary factors that characterize effective interventions with
juvenile offenders: A meta-analytic overview. Victims and Offenders, 4(2), 124147.
Lo, C., Howell, R., & Cheng, T. (2003). Disparities in whites’ versus blacks’ self-rated
health: Social status, health-care services, and health behaviors. Journal of
Communication and Health, 38, 727-733.
Lopez-Williams, A., Vander Stoep, A. V., Kuo, E., & Stewart, D. G. (2006). Predictors
of mental health service enrollment among juvenile offenders. Youth Violence and
Juvenile Justice, 4(3), 266-280.
Luk, E. S., Staiger, P. K., Mathai, J., Wong, L., Birleson, P., & Adler, R. (2001). Children
with persistent conduct problems who drop-out of treatment. European Child &
Adolescent Psychiatry, 10(1), 28-36.

153

Lyons, J. S., Baerger, D. R., Quicgley, P., Erlich, J., & Griffith (2001). Mental health
service needs of juvenile offenders: A comparison of detention, incarceration, and
treatment settings. Children's Services: Social Policy, Research, in Practice,4(2),
69-85.
Lyons, J. S., Griffin, G., Quintenz, S., Jenuwine, M., & Shasha, M. (2003). Clinical and
forensic outcomes from the Illinois mental health juvenile justice initiative.
Psychiatric Services, 54(12), 1629-1634.
Maguire, J. F. (2012, February 6). Hon. Marilyn A. Moores. Indiana Court Times.
Retrieved from http://indianacourts.us/times/2012/02/hon-marilyn-a-moores/.
Last accessed on June 17, 2015.
Maschi, T., Hatcher, S. S., Schwalbe, C. S., & Rosato, N. S. (2008). Mapping the social
service pathways of youth to and through the juvenile justice system: A
comprehensive review. Children and Youth Services Review, 30(12), 1376-1385.
McDonald, C. J., Overhage, J. M., Tierney, W. M., Dexter, P. R., Martin, D. K., Suico, J.
G., ... & Wodniak, C. (1999). The Regenstrief medical record system: a quarter
century experience. International journal of medical informatics, 54(3), 225-253.
Menard, S. (2000). Coefficients of determination for multiple logistic regression
analysis. The American Statistician, 54(1), 17-24.
Moore, T., McArthur, M., & Saunders, V. (2013). Young people talk about transitioning
from youth detention to the community: Making good, Australian Social Work,
66(3), 328-343.
Mulvey, E. P., Schubert, C. A., & Chung, H. L. (2007). Service use after court
involvement in a sample of serious adolescent offenders. Children and Youth
Services Review, 29(4), 518-544.
Myner, J., Santman, J., Cappelletty, G. G., & Perlmutter, B. F. (1998). Variables related
to recidivism among juvenile offenders. International Journal of Offender
Therapy and Comparative Criminology, 42(1), 65-80.

154

Novins, D. K., Duclos, C. W., Martin, C., Jewett, C. S., & Manson, S. M. (1999).
Utilization of alcohol, drug, and mental health treatment services among
American Indian adolescent detainees. Journal of the American Academy of Child
and Adolescent Psychiatry, 38(9), 1102-1108.
O’Brien, R. M. (2007). A caution regarding rules of thumb for variance inflation factors.
Quality & Quantity, 41, 673-690.
Oruche, U. M., Downs, S., Holloway, E., Draucker, C., & Aalsma, M. C., (2014).
Barriers and facilitators to treatment participation by adolescents in the
community mental health clinic. Journal of Psychiatric and Mental Health
Nursing, 21, 241-248.
Office of Juvenile Justice and Delinquency Prevention (2013). Statistical briefing book
[Data file]. Retrieved from http://ojjdp.gov/ojstatbb/. Last accessed on February,
22, 2015.
Pajer, K. A., Kelleher, K., Gupta, R. A., Rolls, J., & Gardner, W. (2007). Psychiatric and
medical health care policies in juvenile detention facilities. Journal of The
American Academy Of Child And Adolescent Psychiatry, 46(12), 1660-1667.
Paskar, L. D. (2008). Utilization and Effectiveness of Mental Health Services for
Delinquent Youth with Substance Use Disorders: A Longitudinal Study. ProQuest.
Poe-Yamagata, E. (2009). And justice for some: Differential treatment of minority youth
in the justice system. DIANE Publishing.
Pumariega, A. J., Atkins, D. L., Rogers, K., Montgomery, L., Nybro, C., Caesar, R., &
Millus, D. (1999a). Mental health and incarcerated youth. II: Service utilization.
Journal of Child and Family Studies, 8(2), 205-215.
Rawal, P., Romansky, J., Jenuwine, M., & Lyons, J. S. (2004). Racial Differences in the
Mental Health Needs and Service Utilization of Youth in the Juvenile Justice
System. Journal of Behavioral Health Services & Research, 31(3), 242-254.
Ricks, E. P., & Louden, J. (2014). The relationship between officer orientation and
supervision strategies in community corrections. Law and Human Behavior,
39(2), 130-139.

155

Riley, F. (2014). Idaho's implementation of substance use disorder services for juvenile
justice populations. Corrections Today, May/June, 36-40.
Robertson, A. A., Dill, P. L., Husain, J., & Undesser, C. (2004). Prevalence of mental
illness and substance abuse disorders among incarcerated juvenile offenders in
Mississippi. Child Psychiatry and Human Development, 35(1), 55-74.
Rogers, K. M., Pumariega, A. J., Atkins, D. L., & Cuffe, S. P. (2006). Conditions
associated with identification of mentally ill youths in juvenile detention.
Community Mental Health Journal, 42(1), 25-40.
Rogers, K. M., Zima, B., Powell, E., & Pumariega, A. J. (2001). Who is referred to
mental health services in the juvenile justice system? Journal of Child and Family
Studies, 10(4), 485-494.
Roe, D. J., & Korn, E. L. (1993). Time‐period effects in longitudinal studies measuring
average rates of change. Statistics in Medicine, 12(9), 893-900.
Rozie-Battle, J. L. (2002). African American teens and the neo-juvenile justice system.
Journal of Health & Social Policy,15, 69-79.
Ryan, J. P., Williams, A.B., & Courtney, M.E. (2013). Adolescent neglect, juvenile
delinquency, and the risk of recidivism. Journal of Youth and Adolescence, 42,
454-465.
Samuel, I. A. (2015). Utilization of mental health services among African-American male
adolescents released from juvenile detention: examining reasons for within group
disparities and help seeking behaviors. Child and Adolescent Social Work
Journal, 32, 33-43.
Schaffner, L., (1998). Female juvenile delinquency: Sexual solutions, gender bias and
juvenile justice. Hastings Women's Law Journal, 9(1), 1-25.
Schwalbe, C.S., Gearing, R.E., MacKenzie M.J., Brewer K.B., & Ibrahim, R. (2012). A
meta-analysis of experimental studies of diversion programs for juvenile
offenders. Clinical Psychology Review, 32(1), 26-33.
Sedlak, A. J. (2009). Surveying Youths in Custody. Corrections Today, 71(6), 92-94.

156

Sedlak, A., & Bruce, C. (2010). Youth's characteristics and backgrounds: Findings from
the survey of youth in residential placement. US Department of Justice, Office of
Justice Programs, Office of Juvenile Justice and Delinquency Prevention.
Sedlak, A. J., & McPherson, K. S. (2010). Youth’s needs and services. OJJDP Juvenile
Justice Bulletin, 10-11.
Sexton, T., & C.W. Turner (2010). The effectiveness of functional family therapy for
youth with behavioral problems in a community practice setting. Journal of
Family Psychology, 24(3), 339-348.
Shufelt, J. L., & Cocozza, J. J. (2006). Youth with mental health disorders in the juvenile
justice system: Results from a multi-state prevalence study. Delmar, NY: National
Center for Mental Health and Juvenile Justice.
Shelton, D. (2001). Emotional disorders in young offenders. Journal of Nursing
Scholarship, 33(3), 259-263.
Shelton, D. (2005). Patterns of treatment services and costs for young offenders with
mental disorders. Journal of Child and Adolescent Psychiatric Nursing, 18(3),
103-112.
Soenksen, S., Stein, L.A., Brown, J. D., Stengel, J. R., Rossi, J. S., & Lebeau, R. (2015).
Canabis withdrawal amount detained adolescents: Exploring the impact of
nicotine and race, Journal of Child and Adolescent Substance Abuse, 24(2), 119124.
Steiner, H., Garcia, I. G., & Matthews, Z. (1997). Posttraumatic stress disorder in
incarcerated juvenile delinquents. Journal of the American Academy of Child and
Adolescent Psychiatry, 36(3), 357-365.
Steinhart, D. (2006). Juvenile detention risk assessment. A practice guide to juvenile
detention reform. Baltimore, MD: The Annie E. Casey Foundation.
Stiffman, A., Pescosolido, B. and Cabassa, L.(2004). Building a model to understand
youth service access: The gateway provider model. Mental Health Services
Research, 6(4), 189-198.

157

Survey of Youth in Residential Placement (2003). SYRP WesDax [Public data file].
Retrieved from https://syrp.org/online_analysis.html. Last accessed on September
12, 2014.
Teplin, L.A., Abram, K. M., McClelland, G. M., Dulcan, M. K.,Mericle, A. A. (2002).
Psychiatric disorders in youth in juvenile detention. Archives of General
Psychiatry, 59, 1122-1143.
Teplin, L. A., Abram, K. M., McClelland, G. M., Mericle, A. A., Dulcan, M. K.,
Washburn, J. J., (2006). Psychiatric Disorders of Youth in Detention. Juvenile
Justice Bulletin: Office of Juvenile Justice and Delinquency Prevention.
Teplin, L. A., Abram, K. M., McClelland, G. M., Washburn, J. J., & Pikus, A. K. (2005).
Detecting mental disorder in juvenile detainees: Who receives services? American
Journal of Public Health, 95(10), 1773-1780.
Teplin, L. A., Welty, L., Abram, K., Dulcan, M., &Washburn, J. (2012). Prevalence and
persistence of psychiatric disorders in youth after detention. Archives of General
Psychiatry, 69(10), 1031-1043.
Towberman, D. B. (1994). Racial bias in the criminal justice system: shifting the focus
from outcome to underlying causes. Juvenile and Family Court Journal, 45(1),
15-25.
Trupin, E. W., Turner, A. P., Stewart, D., & Wood, P. (2004). Transition planning and
recidivism among mentally ill juvenile offenders. Behavioral Sciences and the
Law, 22(4), 599-610.
Trupin, E. J., Kerns, S. E., Walker, S. C., DeRobertis, M. T., & Stewart, D. G. (2011).
Family integrated transitions: A promising program for juvenile offenders with
co-occurring disorders. Journal of Child and Adolescent Substance Abuse, 20(5),
421-436.
UCLA: Institute for Digital Research and Education. Annotated SPSS Output
Multinomial Logistic Regression. Retrieved from
http://www.ats.ucla.edu/stat/spss/output/mlogit.htm. Last accessed on June 22,
2015.

158

United States of America Census Bureau (2015). State and county quickfacts. [Public
Database]. Retrieved from http://quickfacts.
census.gov/qfd/states/18/1836003.html. Last accessed on June 6, 2015.
Vander Stoep, A., Evens, C. C., & Taub, J. (1997). I. Risk of juvenile justice system
referral among children in a public mental health system. The Journal of
Behavioral Health Services and Research, 24(4), 428-442.
Veysey, B. (2003). Adolescent girls with mental health disorders involved with the
juvenile justice system. National Center for Mental Health and Juvenile Justice.
Research and Program Brief, 2-6. Retrieved from
http://files.eric.ed.gov/fulltext/ED478687.pdf.
Vincent, G.M., Grisso T., Terry A., & Banks, S. (2008). Sex and race differences in
mental health symptoms in juvenile justice: the MAYSI-2 national meta-analysis.
Journal of the American Academy of Child and Adolescent Psychiatry, 47(3),
282-290.
Wasserman, G. A., Jensen, P. S., Ko, S. J., Cocozza, J., Trupin, E., Angold, A.,
Cauffman, E., & Grisso, T. (2003). Mental health assessments in juvenile justice:
report on the consensus conference. Journal of the American Academy of Child
and Adolescent Psychiatry, 10(7), 752-760.
Wasserman, G. A., McReynolds, L. S., Lucas, C. P., Fisher, P., & Santos, L. (2002). The
voice DISC-IV with incarcerated male youths: Prevalence of disorder. Journal of
the American Academy of Child and Adolescent Psychiatry, 41(3), 314-321.
Wasserman, G. A., McReynolds, L. S., Whited, A. L., Keating, J. M., Musabegovic, H.,
& Huo, Y (2008). Juvenile probation officers’ mental health decision making.
Administration and Policy in Mental Health and Mental Health Services
Research, 35(5): p. 410-422.
White, L. M., Aalsma, M. C., Salyers, M. P. & McGrew, J. H. (under review). Mental
health service utilization among detained adolescents: A meta-analysis of
prevalence and potential moderators. Journal of Correctional Healthcare.

159

Williams, R. J., & Chang, S. Y. (2000). A comprehensive and comparative review of
adolescent substance use treatment outcome. Clinical Psychology: Science and
Practice, 7(2), 138-166.
Williams, V. A., & Grisso T, (2011). Knowledge brief: Does mental health screening for
the Phillips promise? National Criminal Justice Reference Service. [Government
Website]. Retrieved from
https://www.ncjrs.gov/App/Publications/abstract.aspx?ID=261418. Last accessed
on March 3, 2015.
Williams, V. A., Grisso T, Valentine M, & Remsburg, N. (2008). Mental health
screening: Pennsylvania’s experience in juvenile detention. Corrections Today,
70, 24–27.
Willie, M. M. (2012). Analyzing of medical schemes complaints by means of parametric
proportional hazard frailty models. Public Health Research, 2(1), 1-7.
Winter, S. (1991). Youth authority treatment needs assessment. Sacramento, CA: Report
to Department of Youth Authority, Research Division.
Wolf, R. V. (2008). Race, bias, and problem-solving courts. National Black Law Journal,
21, 27-52.
Yazzie, R. A. (2011). Availability of treatment to youth offenders: Comparison of public
versus private programs from a national census. Children and Youth Services
Review, 33(6), 804-809.
Yoo, B. (2010). Impact of missing data on type 1 error rates in non-inferiority trials.
Pharmaceutical Statistics, 9(2), 87-99.
Young, D. W., Dembo, R., & Henderson, C. E. (2007). A national survey of substance
abuse treatment for juvenile offenders. Journal of Substance Abuse Treatment,
32(3), 255-262.

TABLES

160

Table 3.1

Sample Demographics

Gender
Total
Sample

Females

Variables

N (%)
N = 9664

N (%)
n = 3602

Male

6062
(62.7%)

--

Female

3602
(37.3%)

Black

Race/Ethnicity
Males

White

Age Cohorts

Older

Black

Other

Younger

Mid-Age

N (%)
N (%)
n = 6062 n = 3367

N (%)
n = 5667

N (%)
n = 630

N (%)
n = 2406

N (%)
n = 3924

--

2010
(59.7%)

3621
(63.9%)

431
(68.4%)

1626
(67.6%)

2422
(61.7%)

N (%)
n=
3334
2012
(60.3%)

--

--

1357
(40.3%)

2046
(36.1%)

199
(31.6%)

780
(32.4%)

1501
(37.3%)

1321
(39.7%)

5667
(58.6%)

2046
(56.8%)

3621
(59.7%)

--

--

--

1466
(6.1%)

2311
(58.9%)

1890
(56.7%)

White

3367
(34.8%)

1357
(37.7%)

2010
(33.2%)

--

--

--

782
(32.5%)

1332
(33.9%)

1253
(37.6%)

Other

630
(6.5%)

431
(12.0%)

199
(3.3%)

--

--

--

158
(6.6%)

281
(7.2%)

191
(5.7%)

Note. N = Number; % = Percentage. Other = Non-White Hispanic, American Indian/
Alaskan, and Asian/Pacific Islander.

13. RAI Score

12. Charge Severity

3. White Males
(y/n)
4. White Females
(y/n)
5. Black Males
(y/n)
6. Black Females
(y/n)
7. Age at
Detention Entry
8. Age at First
Contact
9. Pre-Detention
Arrest (y/n)
10. Number of Prior
Arrests
11. Number of
Charges

2. White (y/n)

1. Male (y/n)

Variables (N =
9664)
14
---------------------------

13
---------------------------

---------------------------

15
---------------------------

16

Correlation Matrix: Demographic and Criminal History Variables

r
1
2
3
4
5
6
7
8
9
10
11
12
p
r
------------p
------------r .042
-----------p .000
-----------r .395 .694
----------p .000 .000 -----------r -.524 .545 -.207 ---------p .000 .000 .000 ---------r .597 -.630 -.397 -.313 --------p .000 .000 .000 .000
--------r -.672 -.417 -.266 -.209 -.401 -------p .000 .000 .000 .000 .000 -------r -063 .045 -.009 .072 -.047 -------p .000 .000 .385 .000 .000 .324
-----r -.090 .068 -.014 .105 -.077 .012 .758 -----p .000 .000 .190 .000 .000 .266 .000 -----r .042 -.015 .012 -.024 .037 -.016 .001 -.052 ----p .000 .159 .240 .017 .000 121 .938 .000 ----r .161 -.096 .001 -.130 .148 -.076 .103 -.076 .229
---p .000 .000 .936 .000 .000 .000 .000 .000 .000
---r .014 .004 .015 -.013 -.002 -.002 -.005 .007 -.006 .026 --p .181 .720 .152 .218 .810 .874 .623 .526 .567 .013 --r .127 .004 .063 -.068 .070 -.086 -.033 .010 -.011 -.043 -.022 -p .000 .694 .000 .000 .000 .000 .002 .344 .311 .000 .034 -r .218 -.160 -.048 -.200 .187 -.093 -.105 -.089 .048 .070 .052 .104
p .000 .000 .016 .000 .000 .000 .000 .000 .014 .000 .009 .000

Table 3.2

---------------------------

17
---------------------------

18

---------------------------

19

---------------------------

20

---------------------------

21

---------------------------

22

161

161

r .038 -.033 .005 -.056 .017
p .000 .002 .628 .000 .090
r .083 -.002 .034 -.048 .042
p .000 .878 .001 .000 .000
r .110 -.090 -.031 -.078 .133
p .000 .000 .002 .000 .000
r .070 -.062 -.023 -.052 .091
p .000 .000 .025 .000 .000
r .138 -.094 -.023 -.091 .156
p .000 .000 .026 .000 .000
r .076 -.018 .034 -.065 .040
p .000 .214 .014 .000 .004
r .028 -.022 .005 -.029 .029
p .005 .034 .646 .004 .004
r -.019 -.002 -.10 .010 -.009
p .056 .867 .347 .322 .368
r .063 -.055 -.014 -.051 -.078
p .000 .000 .180 .000 .000
r .261 -.105 -.005 -.159 .213
p .000 .000 .628 .000 .000

.005
.638
-.057
.000
-.056
.000
-.035
.001
-.074
.000
-.027
.051
-.001
.894
.014
.154
-.025
.013
-.177
.000

-.152
.000
.008
.448
-.174
.000
-.189
.000
-.386
.000
-.311
.000
-.192
.000
-.008
.415
-.230
.000
.185
.000

-.066
.000
-.024
.021
-.155
.000
-.185
.000
-.336
.000
-.211
.000
-.215
.000
-.099
.000
-.240
.000
.177
.000

.054
.000
.003
.760
-.003
.778
-.045
.000
-.003
.744
.012
.398
-.073
.000
-.065
.000
-.068
.000
.026
.011

-.165
.000
.025
.020
-.032
.003
-.053
.000
-.033
.002
.037
.011
-.036
.001
-.057
.000
-.051
.000
-.165
.000

-.002
.830
.012
.252
-.023
.028
-.025
.015
-.017
.113
.029
.038
-.009
.371
-.013
.208
-.022
.032
.016
.021

-.161 .112
.000 .000
.019 .161
.073 .000
-.033 -.030
.001 .133
-.060 .033
.000 .091
-.031 .035
.002 .074
.039 .065
.006 .014
-.055 .052
.00 .008
-.055 .048
.000 .014
-.065 .051
.00 .010
.113 .043
.000 .029

--.103
.000
-.024
.020
-.053
.000
-.017
.104
.088
.000
-.044
.000
-.031
.003
-.059
.000
.033
.001

-----.012
.226
.048
.000
.001
.885
.002
.900
.108
.000
.033
.001
.089
.000
.055
.000

------------.557 -.000 -.591 .639
.000 .000
-.283 -.800
.000 .000
.237 .356
.000 .000
.035 .101
.001 .000
.510 .881
.000 .000
.058 -.108
.000 .000

Note. n = Number; r = Pearson correlation, two-tailed; p = Significance level; RAI = Risk Assessment Inventory;
a
Outcome (i.e., re-detention, incarceration) within two years of release from detention.

23. Year of Detention

17. Re-Detained
(y/n)a
18. Number of
Detentions
19. Days to 2nd
Detention
20. Incarcerated
(y/n)a
21. Number of
Incarcerations
22. Re-Detained/
Incarcerated (y/n)

16. Recidivist (y/n)

15. Length of Stay

14. Violent Offender
(y/n)

Table 3.2 Continued.

-----------.174
.000
.484
.000
.087
.000
.694
.000
-.112
.000

-------------.109
.000
-.057
.000
-.498
.000
-.135
.000

------------------------------------------.312 --.000 --.756 .229
-.000 .000
--.230 -.140 -.160
.000 .000 .000

162

162

163

Table 3.3

Prevalence of Detained Adolescents and Length of Stay for First
Detention Stay
Detained
Adolescents

Length of Stay (in Days)

Year of Detention

Number (%)
N = 9664

Mean (SD)

Range

1998 – 1999

1063 (11.0%)

9.55 (14.25)

0.0 – 102.0

2000 – 2001

1653 (17.1%)

9.73 (14.57)

0.0 – 90.0

2002 – 2003

1866 (19.3%)

9.57 (15.54)

0.0 – 156.0

2004 – 2005

2233 (23.1%)

7.32 (14.45)

0.0 – 180.0

2006 – 2007

915 (9.5%)

10.35 (16.00)

0.0 – 143.0

2008 – 2009

947 (9.8%)

12.63 (18.29)

0.0 – 141.0

2010 – 2011

987 (10.2%)

12.20 (15.83)

0.0 – 122.0

9664 (100.0%)

9.72 (15.45)

0.0 – 180.0

Total: 1998 – 2011

Note. N = Number; % = Percentage; SD = Standard Deviation.

164

Table 3.4

Variables/Categories

Criminal History Outcomes: Criminal Charges
Timing of Outcome
6 Months
Detention Entry
Post-Detention
Number (%)
Number (%)
N = 9664
N = 6401

Most Severe Charge
Conduct-Related
4445 (46.0%)
1024 (10.6%)
Property-Related
2567 (26.6%)
747 (7.7%)
Drug/Alcohol
439 (4.5%)
285 (2.9%)
Sexual
395 (4.1%)
42 (0.4%)
Trespass/Loitering
385 (4.0%)
136 (1.4%)
Warrant Arrest
272 (2.8%)
-Weapons
243 (2.5%)
28 (0.3%)
Modification or Probation Violation
170 (1.8%)
422 (4.4%)
Driving-Related
165 (1.7%)
35 (0.4%)
Resisting Arrest
114 (1.2%)
208 (2.2%)
Fraud
83 (0.9%)
-Severe
44 (0.5%)
23 (0.2%)
Other
24 (0.2%)
19 (0.2%)
Missing/Unknown
319 (3.3%)
-Reason for Detention Stay
-Awaiting Action (Pre- or Post-Adjudication) 3314 (34.3%)
-Court-Ordered Detention
487 (5.0%)
-Outright Arrest/New Arrest
4953 (51.3%)
-Warrant Arrest or Failure to Appear
406 (4.2%)
-Modification or Probation Violation
47 (0.5%)
-Other
34 (0.3%)
-Missing/Unknown
423 (4.4%)
-Release Decision
Released to Home Detention
4687 (48.5%)
-Community Adjustment
1221 (12.6%)
-Placed in County Jail or DOC Prison
405 (4.2%)
-Released to Parent/Guardian
2542 (26.3%)
-Released to Emergency Shelter or CPS
45 (0.5%)
-Released for Placement (Not Specified)
345 (3.6%)
-Released to Treatment or Hospital
20 (0.2%)
-Other
54 (0.5%)
-Missing/Unknown
344 (3.6%)
-Note. n = Number; % = Percentage; Dept. = Department; DOC = Department of
Corrections; CPS = Child Protective Services.

Females
N (%)
n = 3602
1235
(34.3%)
2093
(58.1%)
1729
(48.0%)
549
(15.2%)
2689
(74.7%)
2515
(69.8%)
1288
(35.8%)
1114
(30.9%)
617
(17.1%)
N (%)
n = 448
108
(24.1%)
264
(58.9%)
76
(17.0%)

N (%)
N
= 9664
Variables
3542
Violent Offender (y/n)
(36.7%)
6401
Recidivist (y/n)
(66.2%)
5227
Re-Detained (y/n)
(54.1%)
1538
Incarcerated (y/n)
(15.9%)
7015
Disorder (y/n)
(72.5%)
Mental Disorder
3571
(y/n)+
(37.0%)
Substance-Related
662
Disorder (y/n)+
(6.9%)
2782
Comorbid (y/n)+
(28.8%)
Pre-Detention
2438
Treatment (y/n)
(25.2%)
Risk Assessment
N (%)
Inventory Level
n = 2568
201
Low-Risk (y/n)++
(7.8%)
Medium1292
Risk(y/n)++
(50.3%)
1075
High-Risk (y/n)++
(41.9%)

N (%)
n = 6062
2307
(38.1%)
4308
(71.1%)
3498
(57.7%)
989
(16.3%)
4326
(71.4%)
3838
(63.3%)
2156
(35.6%)
1668
(27.5%)
1821
(30.0%)
N (%)
n = 2120
93
(4.4%)
1028
(48.5%)
999
(47.1%)

Males

Gender
White
N (%)
n = 3367
1153
(34.2%)
1963
(58.3%)
1673
(49.7%)
514
(15.3%)
2690
(79.9%)
2489
(73.9%)
1463
(43.5%)
1262
(37.5%)
841
(25.0%)
N (%)
n = 638
90
(14.1%)
329
(51.6%)
219
(34.3%)

χ,
p-value
13.83,
p < .001
169.66,
p < .001
85.65,
p < .001
1.95,
p = .168
12.29,
p < .001
42.51,
p < .001
0.85,
p =.861
12.82,
p < .001
199.64,
p < .001
U-test,
p-value

28842.0,
p < .001

2

-N (%)
n = 5667
2125
(37.5%)
4012
(70.8%)
3226
(56.9%)
943
(16.6%)
3847
(67.9%)
3428
(60.8%)
1744
(30.8%)
1325
(23.4%)
1396
(24.6%)
N (%)
n = 1678
89
(5.3%)
838
(49.9%)
751
(44.8%)

Black

Race

45638.50,
p < .001

χ,
p-value
9.67,
p = .002
147.23,
p < .001
44.57,
p < .001
2.94,
p = .086
152.29,
p < .001
168.64,
p < .001
148.23,
p < .001
205.49,
p < .001
0.72,
p = .366
U-test,
p-value
2

--

284
(53.9%)a
220
(41.7%)a

N (%)
n = 2406
1147
(47.7%)
1996
(83.0%)a
1665
(69.2%)a
593
(24.6%)a
1867
(77.6%)a
1739
(72.3%)
803
(33.4%)
675
(28.1%)a
492
(20.4%)a
N (%)
n = 527
23 (4.4%)a

Younger

495
(47.9%)b
476
(46.0%)b

N (%)
n = 3925
1421
(36.2%)
2828
(72.1%)b
2394
(61.0%)a
696
(17.7%)b
2869
(73.1%)b
2612
(66.6%)
1445
(36.8%)
1188
(20.3%)b
975
(24.8%)ab
N (%)
n = 1034
63 (6.1%)a

Mid-Age

N (%)
n = 3334
974
(29.2%)
1577
(47.3%)c
1168
(35.0%)b
249
(7.5%)b
2279
(68.4%)c
2002
(60.0%)
1196
(35.9%)
919
(27.6%)a
971
(29.1%)c
N (%)
n = 1007
115
(11.4%)b
513
(50.9%)bc
379
(37.6%)c

Older

Age Cohorts

Mental Health and Criminal History Outcomes by Gender, Race, and Age Cohorts

Total
Sample

Table 3.5

--

30266.50,
p < .001

χ 2,
p-value
205.62,
p < .001,
895.94,
p < .001
784.49,
p < .001
324.56,
p < .001
60.98,
p < .001
94.80,
p < .001
7.86,
p = .020
7.22,
p = .026
56.26,
p < .001
U-test,
p-value

165

165

Females

Males

Gender
--

White

Black

Race
--

Younger

Mid-Age

Older

Age Cohorts
--

++

Participants able to meet criteria for more than one type of disorder, so percentages do not add up to 100%.
Percentages calculated out of number of participants that obtained Risk Assessment Inventory scores.
+++
Percentages calculated out of total number of participants that completed the MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition.

+

N (%)
N (%)
N (%)
χ2,
N (%)
N (%)
χ2,
N (%)
N (%)
N (%)
χ2,
MAYSI-2
n = 2575
n = 467
n = 2065
p-value
n = 646
n = 1645
p-value
n = 524
n = 1023
n = 985
p-value
1758
379
1379
145.90,
489
1112
288.39,
376
727
655
35.59,
Positive Screen+++
(69.4%)
(81.2%)
(66.8%)
p < .001
(75.7%)
(67.6%)
p < .001
(71.8%)a
(71.1%)a
(66.5%)b
p = .001
Note. N = Number; % = Percentage; χ2 = 2 x 2 chi-square test; p-value = significance level; U-test = Mann-Whitney U-tests; MAYSI-2 = Massachusetts Youth
Screening Instrument-2nd Edition. For age cohorts, different superscripts indicate significant differences (p ≤ .01) and matching superscripts indicate nonsignificant differences (p > .01) based upon 2 x 2 chi-square tests between paired cohorts.

Total
Sample

Table 3.5 Continued.

166

166

167
Table 3.6

Post-Detention Criminal History Outcomes: Recidivism, Detention, and
Incarceration

Criminal History Outcomes
Recidivism (y/n)

Number (%)
N = 9664
6401 (66.2%)

Recidivism within 3 Months

1953 (20.2%)

Recidivism within 4-6 Months

1016 (10.5%)

Recidivism within 7-24 Months

801 (8.3%)

Recidivism after 2 Years

2631 (27.2%)

Re-Detained (y/n)a

5227 (54.1%)

No 2nd Detention

4437 (45.9%)

2nd Detention Stay within 1 Day

114 (1.2%)

2nd Detention Stay within 1 Month

797 (8.2%)

2nd Detention Stay within 2 Months

675 (7.0%)

2nd Detention Stay within 3 Months

498 (5.2%)

2nd Detention Stay within 4-6 Months

955 (9.9%)

2nd Detention Stay within 7-12 Months

969 (10.0%)

2nd Detention Stay within 1-2 Years

775 (8.0%)

2nd Detention Stay after 2 Years

444 (4.6%)

Incarcerated in Prison (y/n)
No Incarceration

1538 (15.9%)
8166 (84.5%)

Incarceration with 1 Day

122 (1.3%)

Incarceration within 1 Month

10 (0.1%)

Incarceration within 2 Months

34 (0.4%)

Incarceration within 3 Months

37 (0.4%)

Incarceration within 4-6 Months

127 (1.3%)

Incarceration within 7-12 Months

281 (2.9%)

Incarceration within 1-2 Years

424 (4.4%)

Incarceration after 2 Years
462 (4.8%)
a
Note. N = Number; % = Percentage. Re-Detained refers to second, separate
placement in detention, following first detention stay

168
Table 3.7

Massachusetts Youth Screening Instrument-2nd Edition (MAYSI-2) among
Cohort Two

MAYSI-2 Variables

Number (%)
N = 2575

Total Score (52 items)

Mean (SD)

Range

14.23 (9.29)

0.0 – 46.0

Alcohol/Drug Use (8 items)

--

1.62 (2.15)

0.0 – 8.0

Angry-Irritable (9 items)

--

4.19 (2.82)

0.0 – 9.0

Depressed-Anxious (9 items

--

2.41 (2.21)

0.0 - 9.0

Somatic Complaints (6 items)

--

2.55 (1.86)

0.0 – 6.0

Suicidal Ideation (5 items)

--

0.79 (1.42)

0.0 – 5.0

Thought Disturbances (5 items)

--

0.71 (0.98)

0.0 – 5.0

Traumatic Experiences (5 items)

--

1.92 (1.54)

0.0 – 5.0

--

--

Score within Caution or Warning Range
Alcohol/Drug Use (8 items)

522 (20.3%)

--

--

Angry-Irritable (9 items)

1212 (47.0%)

--

--

Depressed-Anxious (9 items

1036 (40.2%)

--

--

Somatic Complaints (6 items)

1256 (48.8%)

--

--

Suicidal Ideation (5 items)

523 (20.3%)

--

--

Thought Disturbances (5 items)

1156 (44.9%)

--

--

Traumatic Experiences (5 items)

1409 (55.6%)

--

--

1748 (67.9%)

--

--

≥1 Scale within Warning Range

867 (33.7%)

--

--

≥1 Scale within Caution Range

2065 (80.2%)

--

--

Number of Caution Subscales

--

2.79 (2.12)

0.0 – 7.0

Number of Warning Subscales

--

0.88 (1.46)

0.0 – 7.0

Positive Screen (y/n)

Note. Only participants from cohort two (i.e., detained during 2006 – 2011) completed the
Massachusetts Youth Screening Instrument-2nd Edition. n = Number; % = Percentage; SD
= Standard deviation; MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition.

Correlation Matrix: Demographic and Mental Health Variables

Variables

r
Male White W.M. W.F. B.M. B.F. Age 1
2
3
4
5
6
7
8
9
10 11 12
13
p
r
-.213 .139 .04 .176 -.180 .114 .010 ------------1. MAYSI-2 Scores
p
.000 .000 .042 .000 .000 .000 .324 ------------r
-.121 .079 .034 .090 -.099 .076 -.040 .680 ----------2. Positive Screen on
MAYSI-2 (y/n)
p
.000 .000 .090 .000 .000 .000 .039 .000 -----------r
-.196 .124 .033 .161 -.162 .104 -.049 .938 .753 ----------3. Number of Caution
Scales on MAYSI-2
p
.000 .000 .092 .000 .000 .000 .014 .000 .000 ----------r
-.036 .130 .058 .096 -.089 -.046 -.086 .139 .072 .107 ---------4. Any Disorder (y/n)
p
.000 .000 .000 .000 .000 .000 .000 .000 .000 .000 ---------r
-.066 .137 .049 .115 -.108 -.030 -.106 .145 .078 .113 .851 --------5. Mental Disorder
(y/n)
p
.000 .000 .000 .000 .000 .003 .000 .000 .000 .000 .000 --------r
.153 -.005 .072 -.098 .075 -.094 -.187 .097 .075 .100 .450 .529 -------6. Conduct Disorder
(y/n)
p
.000 .620 .000 .000 .000 .000 .000 .035 .000 .000 .000 .000 -------r
-.226 .146 -.024 .220 -.189 .066 .083 .021 -.015 -.002 .410 .482 -.489 ------7. Non-Conduct Mental
Disorder (y/n)
p
.000 .000 .020 .000 .000 .000 .000 .655 .480 .913 .000 .000 .000 ------r
-.002 .128 .062 .92 -.057 -.078 .020 .113 .076 .095 .457 .236 .166 .071 ------8. Substance-Related
Disorder (y/n)
p
.849 .000 .000 .000 .000 .000 .045 .000 .000 .002 .000 .000 .000 .000 ------r
-.036 .151 .050 .123 -.087 -.069 -.005 .133 .089 .112 .391 .459 .286 .176 .854 -----9. Comorbid (y/n)
p
.000 .000 .000 .000 .000 .000 .624 .000 .000 .000 .000 .000 .000 .000 .000 -----r
-.093 .178 .058 .164 -.134 -.040 -.091 .250 .168 .214 .465 .505 .462 .042 .440 .537 ----10. Number of Disorders
p
.000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 ----r
-.092 .150 .045 .141 -.122 -.023 -.116 .176 .168 .214 .437 .514 .495 .017 .279 .403 .962 ---11. Number of Mental
Disorders
p
.000 .000 .000 .000 .000 .023 .000 .000 .000 .000 .000 .000 .000 .099 .000 .000 .000 ---r
-.050 .172 .067 .149 -.102 -.068 .029 .173 .095 .115 .314 .220 .137 .094 .687 .658 .615 .376 --12. Number of
Substance Disorders
p
.000 .000 .000 .000 .000 .000 .005 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 --r
.196 -.053 .033 -.115 .147 -.131 .132 .060 .023 .041 -- .173 .345 -.247 -.045 .057 .198 .246 -.047 -13. Pre-Detention
Diagnosis (y/n)
p
.000 .000 .006 .000 .000 .000 .000 .005 .301 .061 .000 .000 .000 .000 .000 .000 .000 .000 .000 -r
-.053 .134 .055 .106 -.101 -.034 -.098 .005 .023 .025 .994 .837 .415 .432 .459 .383 .440 .412 .317 -.995
14. Post-Detention
Diagnosis (y/n)
p
.000 .000 .000 .000 .000 .001 .000 .790 .250 .207 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000
Note. W. M. = White Males; W. F. = White Females; B. M. = Black Males; B. F. = Black Females; r = Pearson correlation, two-tailed; p = Significance level.
MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition.

Table 3.8

169

169

170
Table 3.9

Mental Disorders and Substance-Related Disorders among Sample

Variables (N = 9664)
Number (%)
Mean (SD)
Range
Disorder (y/n)
7015 (72.6%)
--Mental Disorder (y/n)
6353 (65.7%)
----Primary Mental Health Diagnosis
Conduct-Related Disorder (CD, ODD)
1585 (16.4%)
--Mood Disorder (Depression, Bipolar)
1529 (15.8%)
--Attention Deficit/Hyperactivity Disorder
1219 (12.6%)
--Anxiety Disorder (Generalized, OCD, PTSD)
849 (8.8%)
--Adjustment or Somatization Disorder
576 (6.0%)
--Psychosis-Related Disorder
143 (1.5%)
--Other (Eating, Sleep, Feeding, Personality)
452 (4.7%)
--Meets Criteria for Following Diagnosesa
--Mood Disorder (Depression, Bipolar)
3872 (40.1%)
--Conduct-Related Disorder (CD, ODD)
3373 (34.9%)
--Anxiety (Generalized, OCD, PTSD)
2468 (25.5%)
--Attention Deficit/Hyperactivity Disorder
2138 (22.1%)
--Adjustment or Somatization Disorder
1412 (14.6%)
--Psychosis-Related Disorder
697 (7.2%)
--Other (Eating, Sleep, Feeding, Personality)
1448 (15.0%)
--Number of Mental Disorders
-2.90 (4.08)
0.0 – 35.0
1 Disorder
1634 (16.9%)
--2 – 3 Disorders
1970 (20.4%)
--4 – 5 Disorders
1076 (11.1%)
--≥6 Disorders
1673 (17.3%)
--Substance-Related Disorder (y/n)
3444 (35.6%)
--Primary Substance-Related Diagnosis
--Cannabis-Related Disorder
1647 (17.0%)
--Alcohol-Related Disorder
821 (8.5%)
--Drug-Related Disorder (Opioid, Tobacco)
760 (7.9%)
--Other (Use During Pregnancy, Fetal Alcohol)
210 (2.2%)
--Meets Criteria for Following Diagnosesa
--Cannabis-Related Disorder
2154 (22.3%)
--Drug-Related (Opioid, Tobacco)
1564 (16.2%)
--Alcohol-Related Disorder
1177 (12.2%)
--Number of Substance-Related Disorders
-0.72 (1.41)
0.0 – 17.0
1 Disorder
1848 (19.1%)
--2 – 3 Disorders
1171 (12.1%)
--≥4 Disorders
395 (4.1%)
--Total Number of Disorders
-3.62 (4.79)
0.0 – 43.0
1 Disorder
1629 (16.9%)
--2 – 3 Disorders
2061 (21.3%)
--4 – 5 Disorders
1140 (11.8%)
--6 – 7 Disorders
745 (7.7%)
--≥8 Disorders
1440 (14.9%)
--Note. N = Number; % = Percentage; y/n = yes/no; SD = Standard Deviation; CD = Conduct disorder; ODD
= Oppositional defiant disorder; OCD = Obsessive-compulsive disorder; PTSD = Post-traumatic stress
disorder. a Participants able to meet criteria for multiple disorders, so percentages do not total 100%.

Number (%)

Number (%)

Number (%)

1 Year Prior
to Detention
Number (%)

Number (%)

6 Months Prior 1 Month Prior
to Detention
to Detention

Number (%)

During
Detention

Any Treatment
2438 (25.2%) 2141 (22.2%) 1858 (19.2%) 1538 (15.9%)
918 (9.5%)
323 (3.3%)
Mental Health
2297 (23.8%) 1957 (20.3%) 1674 (17.3%) 1365 (14.2%)
797 (8.2%)
302 (7.1%)
Substance-Related
354 (3.7%)
320 (3.7%)
285 (2.9%)
214 (2.2%)
117 (1.2%)
14 (0.1%)
Either/Both
575 (5.9%)
200 (2.1%)
181 (1.9%)
124 (1.3%)
47 (0.5%)
15 (0.2%)
Outpatient Treatment
2435 (25.2%) 2027 (21.0%) 1766 (18.3%) 1477 (15.3%)
873 (9.0%)
315 (3.3%)
Mental Health
2270 (23.5%) 1905 (19.7%) 1647 (17.0%) 1344 (13.9%)
774 (8.0%)
294 (3.0%)
Substance-Related
300 (4.1%)
212 (2.2%)
175 (1.3%)
121 (1.3%)
91 (0.9%)
14 (0.2%)
Either/Both
537 (5.6%)
138 (1.4%)
130 (1.3%)
103 (1.1%)
44 (0.5%)
15 (0.2%)
Non-Outpatient Treatment
718 (7.4%)
611 (6.3%)
488 (5.0%)
365 (3.7%)
157 (1.6%)
16 (0.2%)
Mental Health
571 (6.1%)
368 (4.7%)
344 (3.6%)
251 (2.6%)
89 (0.9%)
14 (0.1%)
Substance-Related
153 (1.6%)
134 (1.4%)
132 (1.4%)
109 (1.1%)
66 (0.7%)
2 (0.02%)
Either/Both
191 (2.0%)
88 (0.9%)
62 (0.64%)
30 (0.3%)
6 (0.1%)
0 (0.0%)
Inpatient Stay
456 (4.7%)
334 (3.5%)
257 (2.7%)
174 (1.7%)
34 (0.3%)
10 (0.1%)
Mental Health
416 (4.3%)
294 (3.0%)
215 (2.2%)
148 (1.5%)
30 (0.3%)
10 (0.1%)
Substance-Related
11 (0.1%)
5 (0.04%)
5 (0.04%)
3 (0.03%)
0 (0.0%)
0 (0.0%)
Either/Both
63 (0.7%)
54 (0.6%)
54 (0.6%)
29 (0.3%)
5 (0.1%)
0 (0.0%)
Emergency Department Visit
426 (4.4%)
385 (3.9%)
313 (3.2%)
239 (2.5%)
128 (1.3%)
6 (0.1%)
Mental Health
285 (2.9%)
246 (2.5%)
199 (2.1%)
145 (1.5%)
63 (0.7%)
4 (0.03%)
Substance-Related
146 (1.5%)
131 (1.4%)
127 (1.3%)
105 (1.1%)
66 (0.7%)
2 (0.02%)
Either/Both
155 (1.6%)
27 (0.7%)
15 (0.2%)
2 (0.02%)
2 (0.02%)
0 (0.0%)
Note. N = Number; % = Percentage. Participants may have obtained more than one treatment, so percentages do not add to 100%.

2 Years Prior
to Detention

Any Time Prior
to Detention

Timing of Treatment Utilization

Prevalence of Pre-Detention and During Detention Treatment Utilization: Subdivided by Treatment Type and
Treatment Setting

Treatment Variables
(N = 9664)

Table 3.10

171

171

1. Pre-Detention
Treatment (y/n)
2. Pre-Detention MH
Treatment (y/n)
3. Pre-Detention SR
Treatment (y/n)
4. Pre-Detention E/B
Treatment (y/n)
5. Pre-Detention
Treatment Number
6. Pre-Detention MH
Treatment Number
7. Pre-Detention SR
Treatment Number
8. Pre-Detention E/B
Treatment Number
9. Pre-Detention NonOutpatient (y/n)
10. Pre-Detention
Outpatient (y/n)
11. Pre-Detention NonOutpatient Number
12. Pre-Detention
Outpatient Number

Variables (N = 9664)

r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
.145
.000
.139
.000
.055
.000
.019
.058
.085
.000
.084
.000
.042
.00
.019
.056
.023
.022
.144
.000
.008
.451
.086
.000

.004
.715
-.005
.657
.065
.000
.085
.000
-.015
.156
-.019
.075
.028
.009
.043
.000
.082
.000
.003
.762
.059
.000
-.017
.105

.071
.000
.050
.000
.131
.000
.065
.000
.015
.142
.009
.368
.088
.000
.054
.000
.051
.000
.074
.000
.024
.018
.014
.157

MaleWhite Age
--.943
.000
.327
.000
.426
.000
.449
.000
.439
.000
.182
.000
.212
.000
.479
.000
.982
.000
.286
.000
.447
.000

1
----.176
.000
.400
.000
.472
.000
.465
.000
.106
.000
.208
.000
.448
.000
.960
.000
.286
.000
.470
.000

2
------.296
.000
.123
.000
.094
.000
.552
.000
.223
.000
.331
.000
.331
.000
.201
.000
.119
.000

3
--------.390
.000
.364
.000
.168
.000
.497
.000
.746
.000
.433
.000
.560
.000
.377
.000

4
----------.998
.000
.110
.000
.296
.000
.442
.000
.458
.000
.500
.000
1.00
.000

5

7

-- --- --- --- --- --- --- --- --- --- --- --- -.066 -.000 -.240 .304
.000 .000
.427 .313
.000 .000
.448 .215
.000 .000
.479 .487
.000 .000
.998 .284
.000 .000

6
----------------.163
.000
.184
.000
.112
.000
.108
.000

8
------------------.486
.000
.598
.000
.429
.000

9

11

----------------------------------------.291 -.000 -.455 .474
.000 .000

10
-------------------------

12

-------------------------

13

-------------------------

14

-------------------------

15

Table 3.11 Correlation Matrix: Demographics, Pre-Detention Treatment Utilization, and During Detention Treatment Utilization

-------------------------

16

172

-------------------------

17

172

.303
.000
.216
.000
.026
.011

.065
.000
.024
.019
.126
.000

.184
.000
.118
.000
.056
.000

.055 -.778 .233 .005 .049 .000 .000
.020 .003 .012 .067 .064 .090 .112
.029 .009 -.033 .231 .220 .043 .130
.005 .404 .001 .000 .000 .000 .000
.261 -.105 .185 .571 .560 .207 .276
.000 .000 .000 .000 .000 .000 .000

-.030 .336
.004 .000
-.036 .223
.000 .000
.018 .067
.074 .000

r
p
r
p
r
p

.014
.194
.008
.469
.015
.158

.049
.000
.025
.013
.023
.026

r
p
r
p
r
p
.046
.000
.297
.000
.320
.000

.328
.000
.297
.000
.018
.071
.041
.000
.297
.000
.345
.000

.086
.000
.080
.000
.168
.000

.054
.000
.028
.007
.134
.000

.326 .105 .042
.000 .000 .000
.297 .073 .018
.000 .000 .076
.016 .016 .056
.119 .127 .000
.049
.000
.168
.000
.305
.000

.231
.000
.161
.000
.056
.000
.068
.000
.215
.000
.576
.000

.298
.000
.208
.000
.051
.000

.047
.000
.296
.000
.351
.000

.326
.000
.295
.000
.018
.071

.020
.054
.179
.000
.214
.000

.200
.000
.180
.000
.011
.262

.212
.000
.704
.000
.176
.000

--.680
.000
.186
.000

.030
.003
.991
.000
.117
.000

----.044
.000

.096
.000
.151
.000
.044
.000

-------

--------.114
.000
.059
.000

Note. n = Number; r = Pearson correlation, two-tailed; p = Significance level; MH = Mental Health; SR = Substance-Related; E/B =
Either/Both.

18. Year of Detention

Detention Number

17. Treatment During

Detention (y/n)

13. Treatment During
Detention (y/n)
14. MH Treatment During
Detention (y/n)
15. S-R Treatment During
Detention (y/n)
16. E/B Treatment During

Table 3.11 Continued.

173

----.124
.000

-------

173

Timing of Post-Detention Treatment Utilization
1 Month
6 Months
1 Year
18 Months
2 Years
Any time
Number (%)
Number (%)
Number (%)
Number (%)
Number (%)
Number (%)
Treatment Variables
N = 9502
N = 9319
N = 9156
N = 9015
N = 8915
N = 9664
Any Treatment
971 (10.2%)
2055 (22.1%)
2659 (29.0%)
3002 (33.3%)
3269 (36.7%) 6437 (66.6%)
Mental Health
849 (8.9%)
1733 (18.6%)
2263 (24.7%)
2560 (28.4%)
2794 (31.4%) 5496 (56.9%)
Substance-Related
60 (0.5%)
399 (4.3%)
617 (6.7%)
788 (8.7%)
930 (10.4%)
2573 (26.6%)
Either/Both
61 (0.5%)
199 (2.1%)
325 (3.5%)
401 (4.4%)
483 (5.4%)
1652 (17.1%)
Outpatient Treatment
910 (9.6%)
2018 (21.7%)
2639 (28.8%)
2941 (32.6%)
3207 (36.0%) 6234 (64.5%)
Mental Health
825 (8.7%)
1717 (18.4%)
2242 (24.5%)
2541 (28.2%)
2751 (31.0%) 5364 (55.5%)
Substance-Related
52 (0.5%)
367 (3.9%)
563 (7.1%)
707 (7.8%)
833 (9.4%)
2268 (23.5%)
Either/Both
56 (0.5%)
186 (2.0%)
305 (3.3%)
360 (4.0%)
445 (5.0%)
1454 (15.1%)
Non-Outpatient
73 (0.8%)
259 (2.7%)
424 (5.6%)
562 (6.2%)
1119 (8.0%)
2515 (26.0%)
Mental Health
61 (0.6%)
193 (2.1%)
304 (3.3%)
404 (4.5%)
486 (5.5%)
1795 (18.6%)
Substance-Related
8 (0.1%)
57 (0.6%)
121 (1.3%)
183 (2.0%)
233 (2.6%)
977 (10.1%)
Either/Both
5 (0.1%)
30 (0.3%)
49 (0.5%)
66 (0.7%)
97 (1.0%)
547 (5.7%)
Inpatient Treatment
44 (0.4%)
153 (1.6%)
234 (2.6%)
296 (3.3%)
373 (4.2%)
1398 (14.5%)
Mental Health
42 (0.4%)
114 (1.2%)
185 (2.0%)
247 (2.7%)
289 (3.2%)
1002 (10.4%)
Substance-Related
0 (0.0%)
3 (0.03%)
12 (0.1%)
16 (0.2%)
25 (0.3%)
251 (2.6%)
Either/Both
3 (0.03%)
29 (0.3%)
43 (0.5%)
55 (0.6%)
76 (0.9%)
453 (4.7%)
Emergency Dept.
33 (0.3%)
161 (1.7%)
277 (3.0%)
382 (4.2%)
496 (5.6%)
1891 (19.6%)
Mental Health
23 (0.2%)
107 (1.1%)
163 (1.8%)
222 (2.5%)
288 (3.3%)
1276 (13.2%)
Substance-Related
8 (.1%)
54 (0.5%)
109 (1.2%)
167 (1.9%)
209 (2.4%)
808 (8.4%)
Either/Both
2 (.02%)
2 (0.02%)
8 (0.1%)
14 (0.3%)
27 (0.3%)
164 (1.7%)
Note. Dept. = Department; N = Number. Participants may have obtained more than one treatment, so percentages do not add up to
100%. Due to attrition, total number of participants (N) for analyses decreased as length of follow-up increased.

Table 3.12 Prevalence of Post-Detention Treatment Utilization: Subdivided by Treatment Type and Treatment Setting

174

174

Variables (N = 8915)

r
Male White W.M. W.F. B. M. B. F. Age
1
2
3
4
5
6
7
8
9
10
11
12
p
1. Post-Detention
r .150 .041 .090 -.064 .054 -.126 -.103 ------------Treatment (y/n)
p .000 .000 .000 .000 .000 .000 .000
------------2. Post-Detention MH
r .103 .042 .080 -.047 .026 -.088 -.156 .889
-----------Treatment (y/n)
p .000 .000 .000 .000 .016 .000 .000 .000
-----------3. Post-Detention SR
r .115 .038 .061 -.031 .043 -.107 .070 .449 .179 ----------Treatment (y/n)
p .000 .000 .003 .000 .000 .000 .000 .000 .000 ----------4. Post-Detention E/B
r .076 .048 .063 -.013 .017 -.075 .016 .319 .254 .303
---------Treatment (y/n)
p .000 .000 .000 .233 .116 .000 .137 .000 .000 .000
---------5. Post-Detention
r .062 .022 .054 -.031 .017 -.042 .156 .463 .487 .150 .246 --------Treatment Number
p .000 .420 .000 .003 .111 .000 .000 .000 .000 .000 .000 --------6. Post-Detention MH
r .440 .019 .045 -.024 .009 -.028 -.167 .420 .473 .047 .148 .970
-------Treatment Number
p .000 .075 .000 .024 .387 .008 .000 .000 .000 .000 .000 .000
-------7. Post-Detention SR
r .078 .003 .031 -.036 .042 -.059 .036 .235 .088 .523 .194 .171 -.001 ------Treatment Number
p .000 .761 .003 .001 .000 .000 .001 .000 .000 .000 .000 .000 .954 ------8. Post-Detention E/B
r .045 .019 .300 -.016 .010 -.037 -.005 .139 .116 .148 .436 .252 .065 .117
-----Treatment Number
p .000 .086 .004 .141 .338 .000 .671 .000 .000 .000 .000 .000 .000 .000
-----9. Post-Detention
r .147 .031 .083 -.068 .059 -.119 -.109 .981 .893 .416 .320 .469 .426 .234 .141
----Outpatient (y/n)
p .000 .005 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000
----10. Post-Detention Nonr -.003 .104 .070 .049 -.067 -.040 -.030 .389 .343 .326 .277 .311 .278 .145 .125 .336 ---outpatient (y/n)
p .794 .000 .000 .000 .000 .000 .005 .000 .000 .000 .000 .000 .000 .000 .000 .000 ---11. Post-Detention
r .006 .077 .053 .038 -.035 -.043 -.035 .254 .250 .174 .259 .300 .274 .087 .137 .234 .649 --Outpatient Number
p .570 .000 .000 .000 .001 .000 .001 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 --12. Post-Detention Nonr .063 .020 .052 -.033 .018 -.041 -.157 .458 .483 .143 .238 .999 .970 .168 .250 .465 .288 .263 -Outpatient Number
p .000 .070 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 -r .261 -.105 -.005 -.159 .213 -.177 .185 .426 .364 .225 .239 .206 .168 .137 .141 .420 .161 .109 .204
13. Year of Detention
p .000 .000 .628 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000
Note. n = Number; r = Pearson correlation, two-tailed; p = Significance level; W. M. = White Male; W. F. = White Female; B. M. = Black Male; B. F. = Black
Female; MH = Mental Health; SR = Substance-Related; E/B = Either/Both.

Table 3.13 Correlation Matrix: Demographics and Post-Detention Treatment Utilization

175

175

944
(27.6%)
860
(25.3%)
202
(5.9%)
109
(3.2%)
918
(27.0%)

3269
(36.7%)

2794
(31.4%)

930
(10.4%)

483
(5.4%)

3207
(36.0%)

722
(8.1%)

Any
Treatment

Mental Health

SubstanceRelated

Either/Both

Outpatient

NonOutpatient

--

52.63,
p < .001

443
(8.0%)

0.08,
p = .782

2289
190.47,
(41.5%) p < .001

374
(6.8%)

728
118.97,
(13.2%) p < .001

1934
94.03,
(35.1%) p < .001

2335
200.22,
(42.3%) p < .001

N (%)
χ2,
n = 5515 p-value

Males

Gender
Black

--

Younger Mid-Age

Older

Age Cohorts

367
(11.6%)

1179
(37.3%)

209
(6.6%)

367
(11.6%)

1057
(33.4%)

1221
(38.6%)

13.96,
947
p < .001 (41.4%)a

15.41,
p < .001

91
(4.0%)a

12.23,
147
p < .001 (6.4%)a

302
(5.8%)

88.87,
200
p < .001 (8.8%)a

1776
7.84,
938
(34.3%) p = .002 (41.0%)a

239
(4.6%)

478
(9.2%)

1525
14.66,
905
(29.4%) p < .001 (39.6%)a

1791
(34.5%)

χ2,
p-value

--

146
(4.8%)a

26.88,
p < .001

315
(8.8%)a

207
(6.8%)b

11.20,
p = .014

1397
872
114.52,
a
b
(39.2%) (28.5%) p < .001

246
(6.9%)b

423
360
3.84,
b
b
(11.9%) (11.8%) p = .147

1219
670
212.76,
(34.2%)b (21.9%)c p < .001

1419
903
1054.42,
(39.8%)a (29.5%)b p < .001

N (%)
N (%)
χ2,
N (%)
N (%)
N (%)
n = 3164 n = 5185 p-value n = 2287 n = 3568 n = 3060

White

Race

Note. N = Number; % = Percentage; χ2 = Chi-square test; p-value = Significance level. For age cohorts, different superscripts
indicate significant differences (p ≤ .01) and matching superscripts indicate non-significant differences (p > .01) based upon 2 x 2
chi-square tests between paired cohorts.

279
(8.2%)

N (%)
N (%)
N = 8915 n = 3400

Females

Treatment
Variables

Total
Sample

Table 3.14 Prevalence of Post-Detention Treatment Utilization within Two Years Post-Detention: Demographic Groups

176

176

r
p
1. MAYSI-2 Scores
-r
2. MAYSI-2 – Positive
Screen (y/n)
p
r
3. Number of Caution
Scales
p
r
4. Disorder (y/n)
p
r
5. Mental Disorder (y/n)
p
r
6. Substance-Related
Disorder (y/n)
p
r
7. Comorbid Disorder (y/n)
p
r
8. Conduct Disorder (y/n)
p
r
9. Non-Conduct Disorder
(y/n)
p
r
10. Number of Disorders
p
r
11. Number of Mental
Disorders
p
r
12. Number of Substance
Disorders
p
r
13. Pre-Detention
Diagnosis (y/n)
p
r
14. Post-Detention
Diagnosis (y/n)
p

Variables (N = 8915)

2
---.753
.000
.072
.000
.078
.000
.076
.000
.089
.000
.075
.001
-.01
.661
.168
.000
.168
.000
.088
.000
.023
.301
.069
.001

1
-.680
.000
.938
.000
.139
.000
.145
.000
.113
.000
.133
.000
.095
.038
.020
.672
.250
.000
.174
.000
.175
.000
.060
.005
.131
.000

-----.107
.000
.113
.000
.095
.000
.112
.000
.098
.000
.002
.923
.214
.000
.202
.000
.108
.000
.041
.061
.101
.000

3
-------.851
.000
.457
.000
.391
.000
.444
.000
.419
.000
.465
.000
.438
.000
.318
.000
.329
.000
.994
.000

4

Table 3.15

---------.236
.000
.460
.000
.522
.000
.493
.000
.505
.000
.515
.000
.232
.000
.173
.000
.837
.000

5
-----------.850
.000
.162
.000
.080
.000
.440
.000
.287
.000
.689
.000
-.045
.024
.459
.000

6
-------------.281
.000
.188
.000
.537
.000
.412
.000
.658
.000
.057
.000
.383
.000

7
----------------.057
.000
.032
.003
.101
.000
-.236
.000
.439
.000

---.485
.000
.452
.000
.487
.000
.125
.000
.332
.000
.411
.000

9

-------------

8
-------------------.961
.000
.623
.000
.198
.000
.446
.000

10
---------------------.383
.000
.236
.000
.414
.000

11

13

-----------------------------------------------.033 -.008 -.319 -.463
.000 .000

12
----------------------------

14
----------------------------

15
----------------------------

16
----------------------------

17
----------------------------

18

Variables and Post-Detention Treatment Utilization Variables

----------------------------

19
----------------------------

20

----------------------------

21

----------------------------

22

----------------------------

23

----------------------------

24

177

----------------------------

25

177

r .146 .098 .120 .471 .443 .237 .250 .563 -.120 .350 .366 .137 .396 .090 -----------15. Post-Detention
Treatment (y/n)
p .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 -----------r .159 .102 .134 .419 .493 .094 .203 .598 -.105 .254 .420 .068 .396 .138 .89 ----------16. Post-Detention MH
Treatment (y/n)
p .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 ----------r .051 .031 .031 .211 .073 .458 .352 .157 -.085 .019 .316 .133 .160 -.065 .449 .179 ---------17. Post-Detention SR
Treatment (y/n)
p .010 .115 .120 .000 .000 .000 .000 .000 .000 .152 .000 .000 .000 .000 .000 .000 ---------18. Post-Detention E/B
r .091 .054 .077 .150 .177 .321 .376 .246 -.069 .102 .276 .157 .184 .014 .319 .254 .303 --------Treatment (y/n)
p .000 .007 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .191 .000 .000 .000 --------19. Number of Postr .103 .079 .086 .218 .242 .113 .158 .375 -.134 .556 .412 .093 .306 -.029 .463 .487 .500 .246 -------Detention Treatment
p .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .006 .000 .000 .000 .000 -------20. Number of Postr .094 .072 .078 .198 .233 .054 .106 .356 -.124 .555 .409 .039 .285 -.02 .420 .473 .047 .148 .970 ------Detention MH Tx
p .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .060 .000 .000 .000 .000 .000 ------21. Number of Postr .038 .03 .033 .111 .033 .240 .179 .083 -.050 .015 .031 .109 .096 -.048 .235 .088 .523 .194 .171 -.001 -----Detention SR Treatment p .056 .125 .10 .000 .002 .000 .000 .000 .000 .000 .003 .000 .000 .000 .000 .000 .000 .000 .000 .954 -----22. Number of Postr .030 .023 .027 .066 .077 .142 .166 .123 -.046 .106 .087 .149 .106 -.015 .139 .116 .148 .436 .252 .065 .117 ----Detention E/B Tx
p .124 .244 .175 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .154 .000 .000 .000 .000 .000 .000 .000 ----23. Post-Detention
r .139 .092 .112 .464 .441 .222 .239 .563 -.123 .238 .368 .128 .392 .092 .981 .893 .416 .320 .469 .426 .234 .141 --Outpatient (y/n)
p .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 --24. Post-Detention Nonr .173 .142 .152 .184 .184 .217 .240 .244 -.064 .218 .375 .220 .201 .006 .389 .343 .326 .277 .311 .278 .145 .125 .336 --Outpatient Tx (y/n)
p .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .579 .000 .000 .000 .000 .000 .000 .000 .000 .000 --25. Number of Postr .099 .075 .082 .216 .240 .108 .13 .372 -.134 .554 .403 .086 .304 -.029 .458 .483 .143 .238 .999 .970 .168 .250 .465 .288 -Detention Outpatient
p .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .007 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 -26. Number of Post-Detention r .148 .115 .110 .120 .130 .150 .177 .196 -.067 .258 .388 .199 .147 -.005 .254 .250 .174 .259 .300 .274 .087 .137 .234 .649 .263
Non-Outpatient
p .000 .001 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .670 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000
Note. n = Number; r = Pearson correlation, two-tailed; p = Significance level; MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition; MH = Mental Health; SR
= Substance-Related; E/B = Either/Both; Tx = Treatment.

Table 3.15 Continued.

178

178

445
(7.5%)a
177
(2.8%)a
1934
(26.8%)a

2800
(31.4%)

930
(10.4%)

496
(5.6%)

3207
(36.0%)

723
(8.1%)

Mental
Health

SubstanceRelated

Either/Both

Outpatient

NonOutpatient

39
(5.9%)a

329
(54.6%)b

67
(10.2%)b

114
(17.4%)b

291
(44.4%)b

354
(54.0%)b

N (%)
n = 603

255
(17.4%)b

944
(64.04%)c

225
(15.4%)c

306
(20.9%)c

807
(55.1%)c

963
(65.3%)bc

N (%)
n = 1474

218.31,
p < .001

770.27,
p < .001

378.83,
p < .001

273.25,
p < .001

555.40,
p < .001

228
(37.0%)b

N (%)
n = 617

228
(37.0%)b

212
(6.5%)a

1606
(49.3%)a

38
(6.2%)a

214
(34.7%)b

5 (0.2%)a 0 (0.0%)a

0 (0.0%)a

473
(18.4%)b

1387
(54.0%)c

491
(19.1%)b

702
(27.3%)c

1186
(46.2%)c

1427
(55.6%)c

N (%)
n = 2567

Substance
Comorbid
Disorder

1614
0 (0.0%)b
(49.5%)a

1614
(49.5%)a

N (%)
n = 3259

χ2,
p-value
788.48,
p < .001

Mental
Disorder

--

Disorder Type

599.17,
p < .001

2003.20,
p < .001

547.87,
p < .001

1918.03,
p < .001

2171.72,
p < .001

2058.99,
p < .001

χ2,
p-value

--

328
(4.8%)

1728
(25.3%)

206
(3.0%)

488
(7.2%)

1452
(21.3%)

1761
(25.8%)

N (%)
n = 6823

395
(18.9%)

1479
(70.7%)

290
(13.7%)

442
(21.2%)

1348
(64.4%)

1503
(72.1%)

N (%)
n = 2077

No Prior
Prior
Treatment Treatment

429.81,
p < .001

1432.69,
p < .001

362.99,
p < .001

335.68,
p < .001

1385.79,
p < .001

1477.71,
p < .001

χ2,
p-value

--

Pre-Detention Treatment

Note: N = Number; % = Percentage; χ2 = Chi-square test; p-value = Significance level; MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition. For
MAYSI-2 groups and disorder type groups, different superscripts indicate significant differences (p ≤ .01) and matching superscripts indicate non-significant
differences (p > .01) based upon 2 x 2 chi-square tests between paired groups.

358
(5.7%)a

1505
(23.9%)a

1974
(28.9%)a

3269
(36.7%)

Any
Treatment

N (%)
n = 6838

N (%)
N = 8915

No
Non-Positive Positive
MAYSI-2
Screen
Screen

Treatment
Variables

Total
Sample

MAYSI-2

Table 3.16 Prevalence of Treatment Utilization within Two Years Post-Detention: Mental Health Groups

179

179

1. Pre-Detention
Treatment (y/n)
2. Pre-Detention MH
Treatment (y/n)
3. Pre-Detention SR
Treatment (y/n)
4. Pre-Detention E/B
Treatment (y/n)
5. Pre-Detention
Treatment Number
6. Pre-Detention MH
Treatment Number
7. Pre-Detention SR
Treatment Number
8. Pre-Detn E/B
Treatment Number
9. Pre-Detention
Outpatient (y/n)
10. Pre-Detn NonOutpatient (y/n)
11. Pre-Detention
Outpatient Number
12. Pre-Detn NonOutpatient Number
13. During Detention
Treatment (y/n)
14. During Detn MH
Treatment (y/n)

Variables (N = 9664)

r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p

--.940
.000
.331
.000
.422
.000
.448
.000
.438
.000
.179
.000
.217
.000
.981
.000
.473
.000
.445
.000
.278
.000
.336
.000
.220
.000

1

----.171
.000
.389
.000
.471
.000
.465
.000
.101
.000
.209
.000
.958
.000
.433
.000
.469
.000
.275
.000
.301
.000
.213
.000

2
------.311
.000
.133
.000
.101
.000
.537
.000
.229
.000
.335
.000
.350
.000
.128
.000
.216
.000
.047
.000
.018
.082

3
--------.366
.000
.339
.000
.176
.000
.514
.000
.43
.000
.743
.000
.352
.000
.545
.000
.167
.000
.104
.000

4
----------.998
.000
.130
.000
.325
.000
.456
.000
.413
.000
.999
.000
.488
.000
.319
.000
.298
.000

5
------------.080
.000
.270
.000
.446
.000
.397
.000
.998
.000
.465
.000
.318
.000
.299
.000

6

.202
.000
.182
.000
.165
.000
.128
.000
.111
.000
.037
.000
.014
.183

--------------

7
----------------.221
.000
.323
.000
.311
.000
.544
.000
.101
.000
.065
.000

8
------------------.480
.000
.454
.000
.283
.000
.294
.000
.205
.000

9
--------------------.399
.000
.589
.000
.208
.000
.145
.000

10
----------------------.460
.000
.318
.000
.297
.000

11

13

------------------------------------------------.170 -.000 -.148 .674
.000 .000

12
-----------------------------

14
-----------------------------

15
-----------------------------

16
-----------------------------

17
-----------------------------

18
-----------------------------

19
-----------------------------

20
-----------------------------

21
-----------------------------

22

-----------------------------

23

Table 3.17 Correlation Matrix: Pre-Detention, During Detention, and Post-Detention Treatment Utilization

-----------------------------

24

-----------------------------

25

180

-----------------------------

26

-----------------------------

-----------------------------

27 28

180

15. During Detn SR
r .059 .013 .107 .054 .010 .007 .051 .019 .039 .057 .010 .013 .154 .036 ------- -- ----- -- -- -Treatment (y/n)
p .000 .219 .000 .000 .349 .507 .000 .070 .000 .000 .360 .225 .000 .001 ------- -- ----- -- -- -16. During Detn E/B
r .068 .065 .106 .122 .057 .049 .063 .112 .069 .061 .057 .026 .217 .035 .058 ------ -- ----- -- -- -Treatment (y/n)
p .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .001 .000 ------ -- ----- -- -- -17. During Detention
r .228 .216 .036 .117 .299 .30 .023 .075 .211 .153 .298 .149 .695 .992 .134 .113 ----- -- ----- -- -- -Treatment Number p .000 .000 .001 .000 .000 .000 .028 .000 .000 .000 .000 .000 .000 .000 .000 .000 ----- -- ----- -- -- -18. Post-Detention
r .406 .393 .122 .218 .249 .244 .067 .128 .404 .238 .247 .165 .216 .148 .036 .050 .153 ---- -- ----- -- -- -Treatment (y/n)
p .00 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 ---- -- ----- -- -- -19. Post-Detention MH r .394 .398 .073 .204 .259 .257 .047 .113 .390 .228 .258 .167 .230 .162 .025 .037 .165 .889 --- -- ----- -- -- -Treatment (y/n)
p .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .016 .000 .000 .000 --- -- ----- -- -- -20. Post-Detention SR r .192 .154 .187 .104 .057 .050 .099 .068 .193 .109 .056 .050 .052 .024 .057 .054 .033 .449 .179 -- -- ----- -- -- -Treatment (y/n)
p .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .023 .000 .000 .002 .000 .000 -- -- ----- -- -- -21. Post-Detention E/B r .202 .177 .119 .178 .115 .10 .073 .166 .198 .157 .113 .103 .101 .045 .053 .119 .058 .319 .254 .303 -- ----- -- -- -Treatment (y/n)
p .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 -- ----- -- -- -22. Number of Postr .310 .305 .056 .188 .291 .288 .046 .138 .303 .230 .289 .180 .315 .238 .010 .025 .237 .463 .487 .150 .256 ----- -- -- -Detn Treatment
p .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .343 .017 .000 .000 .000 .000 .000 ----- -- -- -23. Post- Detn MH
r .283 .287 .014 .166 .283 .283 .019 .092 .276 .216 .282 .174 .314 .244 -.002 .003 .241 .420 .473 .047 .148 .970 ---- -- -- -Treatment Number p .000 .000 .180 .000 .000 .000 .071 .000 .000 .000 .000 .000 .000 .000 .815 .744 .000 .000 .000 .000 .000 .000 ---- -- -- -24. Number of Postr .124 .092 .163 .059 .038 .032 .087 .060 .125 .057 .038 .033 .021 -.003 .050 .025 .004 ,235 .088 .523 .194 .171 -.001 --- -- -- -Detn SR Treatment p .000 .000 .000 .000 .000 .002 .000 .000 .000 .000 .000 .002 .050 .806 .000 .017 .703 .000 .000 .000 .000 .000 .954 --- -- -- -25. Post- Detn E/B
r .115 .095 .103 .113 .080 .065 .087 .237 .114 .083 .080 .048 .059 .023 .028 .106 .033 .139 .116 .148 .436 .252 .065 .117 -- -- -- -Treatment Number p .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .032 .008 .000 .002 .000 .000 .000 .000 .000 .000 .000 -- -- -- -26. Post-Detention
r .400 .391 .114 .208 .250 .245 .066 .126 .397 .229 .248 .162 .219 .150 .036 .051 .155 .981 .893 .416 .320 .469 .426 .234 .141 -- -- -Outpatient (y/n)
p .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .001 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 -- -- -27. Post-Detn Nonr .220 .195 .112 .150 .144 .138 .043 .132 .215 .231 .139 .175 .141 .096 .033 .021 .10 .389 .343 .326 .277 .311 .278 .145 .125 .336 -- -Outpatient (y/n)
p .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .002 .048 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 -- -28. Number of Postr .307 .303 .053 .182 .290 .287 .045 .132 .301 .226 .289 .174 .317 .241 .011 .025 .241 .458 .483 .143 .238 ,999 ,970 .168 .250 .465 .288 -Detn Outpatient
p .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .312 .017 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 -29. Post-Detn Nonr .164 .145 .071 .198 .126 .118 .023 .184 .151 .197 .119 .241 .111 .075 .003 .010 .075 .254 .25 .174 .259 .30 .274 .087 .137 .234 .649 .263
Outpatient Number p .000 .000 .000 .000 .000 .000 .032 .000 .000 .000 .000 .000 .000 .000 .754 .346 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000
Note. n = Number; r = Pearson correlation, two-tailed; p = Significance level; MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition; MH = Mental Health; SR
= Substance-Related; E/B = Either/Both; Detn = Detention.

Table 3.17 Continued.

181

181

14. Number of
Incarcerations
15. Re-Detained/
Incarcerated (y/n)*

13. Incarceration (y/n)a

12. Days to 2nd Detention

11. Number of
Detentionsb

10. Re-Detention (y/n)a

9. Recidivism (y/n)

8. Length of Stay

6. Risk Assessment
Inventory Scores
7. Violent Offender
(y/n)

5. Charge Severity

4. Number of Charges

1. Age at First Contact
2. Pre-Detention Arrests
(y/n)
3. Number of PreDetention Arrests

Variables (N = 8915)

r
p
-r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
--.046
.000
-.076
.000
.007
.260
.017
.097
-.089
.00
-.071
.000
-.025
.024
-.155
.000
-.182
.000
-.092
.000
-.218
.000
-.216
.000
-.264
.000
-.084
.000

1

3

------.229 -.000 --.006 .026
.567 .013
-.011 -.043
.000 .000
.048 .070
.014 .000
.056 -.026
.000 .013
.000 .116
.978 .000
-.006 .170
.549 .000
-.046 .137
.000 .000
-.059 .087
.000 .000
.014 -.107
.327 .000
-.070 .035
.000 .001
-.042 .018
.000 .085
-.063 .116
.000 .000

2
--------.022
.034
.052
.009
-.005
.830
.014
.182
-.022
.044
-.023
.030
-.015
.162
.029
.042
-.005
.653
-.006
.558
-.007
.520

4
---------.104
.000
-.165
.000
.025
.020
-.032
.003
-.053
.000
-.033
.002
.037
.011
-.036
.001
-.057
.000
-.051
.000

5
-----------.118
.000
.163
.000
-.030
.174
.047
.033
.039
.078
.060
.032
.049
.025
.046
.037
.067
.002

6
-------------.101
.000
-.022
.039
-.049
.000
-.036
.001
.092
.000
-.045
.000
-.030
.005
-.026
.015

7
----------------.010
.344
.049
.000
.012
.245
.002
.877
.077
.000
.054
.000
.029
.006

8
-----------------.563
.000
.145
.000
.000
.000
.245
.000
.553
.000
.045
.000

9

11

12

13

---- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- -.350 --- -.000 --- -.000 -.105 -- -.000 .000 -- -.337 .149 -.097 -.000 .000 .000 -.896 .322 .276 .364
.000 .000 .000 .000
.92 .440 -.051 .300
.000 .000 .000 .000

10

--------------.109
.000

-------------

14
------------------------------

15
------------------------------

16
------------------------------

17
------------------------------

18
------------------------------

19
------------------------------

20

------------------------------

21

------------------------------

22

------------------------------

23

Table 3.18 Correlation Matrix: Criminal History Variables and Treatment Utilization within Two Years of Detention Release

------------------------------

24

182

------------------------------

25

182

------------------------------

26

-.068 .035 .202 -.008 .008 -.036 .052 .073 .207 .152 -.021 -.096 -.040 -.064 .134 ---- -- -- -- ---.000 .001 .000 .441 .435 .068 .000 .000 .000 .000 .052 .000 .000 .000 .000 ---- -- -- -- ----.118 .033 .159 -.006 .005 -.063 .068 .070 .191 .140 -.012 -.079 -.018 -.054 .129 .889 --- -- -- -- ---.000 .002 .000 .590 .610 .002 .000 .000 .000 .000 .248 .00 .000 .000 .000 .000 --- -- -- -- ---.072 .004 .141 -.007 -.006 -.016 -.054 .006 .145 .118 .006 -.074 -.023 -.028 .100 .449 .179 -- -- -- -- ---.000 .711 .000 .492 .563 .431 .000 .590 .000 .000 .550 .000 .029 .008 .000 .008 .000 -- -- -- -- ---.014 .011 .136 -.002 -.010 -.015 -.024 .023 .110 .091 .007 -.049 -.020 -.015 .076 .319 .254 .303 -- -- -- ---.216 .308 .000 .867 .367 .463 .021 .032 .000 .000 .499 .001 .057 .167 .000 .000 .000 .000 -- -- -- ----.112 .007 .095 -.007 -.006 -.093 .040 .048 .146 .098 .00 -.043 .002 -.022 .10 .463 .487 .150 .246 -- -- ---.00 .491 .000 .517 .560 .000 .000 .000 .000 .000 .972 .003 .821 .040 .000 .000 .000 .000 .000 -- -- ----.121 .005 .071 -.007 -.007 -.096 .050 .047 .129 .084 .001 -.033 .006 -.020 .090 .420 .473 .047 .148 .970 -- ---.000 .616 .000 .517 .495 .000 .000 .000 .000 .000 .895 .023 .584 .055 .000 .000 .000 .000 .000 .000 -- ---.041 .005 .079 -.002 -.004 -.017 -.036 .006 .083 .069 -.002 -.038 -.003 -.016 .056 .235 .088 .523 .194 .171 -.001 ---.00 .626 .000 .865 .693 .388 .001 .549 .000 .000 .870 .008 .786 .122 .000 .000 .000 .000 .000 .000 .954 ----.011 .008 .082 -.001 .008 .001 -.014 .015 .055 .038 -.008 -.035 -.016 .001 .029 .139 .116 .148 .436 .252 .065 .117 --.327 .436 .082 .893 .443 .964 .193 .153 .000 .000 .427 .016 .133 .925 .006 .000 .000 .000 .000 .000 .000 .000 ---.074 .031 .200 -.008 .007 -.032 .057 .076 .203 .148 -.021 -.092 -.040 -.065 .131 .981 .893 .416 .320 .469 .436 .234 .141 -.000 .003 .000 .439 .512 .105 .000 .000 .000 .000 .045 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 --.012 .021 .097 -.003 -.014 -.107 .013 .027 .121 .085 .000 -.054 -.027 -.021 .073 .389 .343 .326 .277 .311 .278 .145 .125 .202
.294 .051 .000 .773 .202 .000 .224 .010 .000 .000 .975 .000 .010 .046 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000
-.114 .007 .093 -.007 -.006 -.089 .041 .054 .143 .096 .000 -.042 .002 -.022 .099 .458 .483 .143 .238 .999 .907 .168 .250 .465
.000 .498 .000 .489 .591 .000 .000 .000 .000 .000 .972 .004 .817 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000
-.015 .012 .068 .001 -.011 -.101 .018 .015 .076 -.057 -.005 -.024 -.011 -.019 .051 .254 .250 .174 .259 .30 .274 .087 .137 .234
27. Number of Post-Detn
Non-Outpatient
p .178 .248 .00 .978 .306 .000 .092 .145 .000 .000 .656 .099 .318 .079 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000
Note. n = Number; r = Pearson correlation, two-tailed; p = Significance level; MH = Mental Health; SR = Substance-Related; E/B = Either/Both; Detn = Detention.
a
Outcome (i.e., re-detention, incarceration) within two years of release from detention. b Number of detentions includes first detention stay.

16. Post-Detention
Treatment (y/n)
17. Post-Detention MH
Treatment (y/n)
18. Post-Detention SR
Treatment (y/n)
19. Post-Detention E/B
Treatment (y/n)
20. Number of PostDetention Treatment
21. Number of Post-Detn
MH Treatment
22. Number of Post-Detn
SR Treatment
23. Number of Post-Detn
E/B Treatment
24. Post-Detention
Outpatient (y/n)
25. Post-Detention NonOutpatient (y/n)
26. Number of Post-Detn
Outpatient

r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r
p
r

Table 3.18 Continued.

183

---------------------------------------.288 -.000 -.649 .263
.000 .000

183

661
(11.7%)
339
(6.0%)
1918
1289
29.47,
1537
(33.9%) (39.6%) p < .001 (28.2%)

930
(10.4%)

496
(5.6%)

3207
(36.0%)

723
(8.1%)

SubstanceRelated

Either/Both

Outpatient

NonOutpatient

279
(8.6%)

157
(4.8%)

269
(8.3%)

1.48,
p = .227

5.32,
p = .021

25.71,
p < .001

298
(5.5%)

193
(3.5%)

376
(6.9%)

--

108.79,
p < .001

146
(3.6%)

425
130.86,
(12.3%) p < .001

237
(5.9%)

1670
365.68,
1163
(48.2%) p < .001 (29.0%)

303
(8.7%)

283
(7.1%)

1475
326.02,
994
(42.5%) p < .001 (24.8%)
554
186.63,
(16.0%) p < .001

Yes

486
(9.9%)

2044
(41.7%)

350
(7.1%)

647
(13.2%)

1806
(36.8%)

2085
(42.5%)

N (%)
N (%)
n = 4010 n = 4905

No

47.32,
p < .001

153.75,
p < .001

51.28,
p < .001

88.82,
p < .001

148.25,
p < .001

160.10,
p < .001

χ2,
p-value

--

Re-Detained/Incarcerated++

1706
383.37,
1184
(49.2%) p < .001 (29.5%)

N (%)
χ2,
n = 3468 p-value

Yes

Recidivist+

Note. N = Number; % = Percentage; χ2 = 2 x 2 chi-square test; p-value = Significance level. + Recidivism within two years postdetention. ++ Re-detained in detention center or incarcerated in prison within two years post-detention.

444
(7.8%)

1648
(29.0%)

2800
(31.4%)

Mental Health

1157
40.97,
1325
(35.6%) p < .001 (24.3%)

1968
1301
24.22,
1563
(34.8%) (40.0%) p < .001 (28.7%)

No

3269
(36.7%)

--

Any
Treatment

Yes

N (%)
N (%)
N (%)
χ2,
N (%)
n = 8915 n = 5661 n = 3254 p-value n = 5447

No

Violent Offender

Treatment
Variables

Total
Sample

Table 3.19 Prevalence of Treatment Utilization within Two Years of Detention Release: Criminal History Groups

184

184

185
Table 3.20 Prevalence of Treatment Utilization within Two Years Post-Detention:
Cohort One versus Cohort Two
Cohort Groups
Total Sample
(1998-2011)

Cohort One
(1998-2005)

Cohort Two
(2006-201)

--

Treatment Variables

Number (%)
N = 8915

Number (%)
n = 6574

Number (%)
n = 2341

χ2,
p-value

Any Treatment

3269 (36.7%)

1935 (29.4%)

1334 (56.9%)

564.21, p < .001

Mental Health

2800 (31.4%)

1711 (26.0%)

1089 (46.5%)

336.48, p < .001

Substance-Related

930 (10.4%)

488 (7.4%)

442 (18.9%)

242.54, p < .001

Either/Both

496 (5.6%)

202 (3.1%)

294 (12.6%)

295.65, p < .001

1900 (28.9%)

1307 (55.8%)

543.51, p < .001

2767 (31.0%)

1688 (25.7%)

1079 (46.1%)

336.11, p < .001

Substance-Related

833 (9.4%)

435 (6.6%)

398 (17.0%)

219.75, p < .001

Either/Both

457 (5.1%)

175 (2.7%)

271 (11.6%)

288.64, p < .001

Treatment

1119 (8.0%)

579 (8.8%)

599 (25.6%)

423.85, p < .001

Mental Health

486 (5.5%)

287 (4.4%)

199 (8.5%)

68.43, p < .001

Substance-Related

233 (2.6%)

126 (1.9%)

107 (4.5%)

50.03, p < .001

Either/Both

97 (1.0%)

44 (0.6%)

53 (2.1%)

47.73, p < .001

Outpatient Treatment 3207 (36.0%)
Mental Health

Non-Outpatient

Note. n = Number; % = Percentage; χ2 = 2 x 2 chi-square test; p-value = Significance level.

19.70
(30.05)
20.04
(30.96)
5.31
(8.20)
6.71
(13.46)
19.87
(30.25)
2.06
(2.26)

M (SD)

M (SD)
20.37
(31.33)
20.49
(31.41)
4.12
(9.756)
5.06
(6.578)
20.38
(31.62)
1.96
(2.14)

Range
1.0 –
369.0
1.0 –
369.0
1.0 –
132.0
1.0 –
147.0
1.0 –
369.0
1.0 –
20.0

Females
n = 934

19.43
(29.52)
19.85
(30.75)
5.64
(7.69)
7.20
(14.88)
19.67
(20.69)
2.11
(2.33)

M (SD)

Gender
Males
n = 2335

Race
Black
n = 1791

t-test,
M (SD)
M (SD)
p-value
Total Number of Services
0.66,
20.34
19.97
p = .417
(32.29)
(29.51)
0.25,
20.477
20.52
p = .615
(22.10)
(30.53)
8.06,
4.79
5.80
p = .002
(6.62)
(9.52)
6.63,
6.37
6.65
p = .01
(13.08)
(12.80)
0.38,
20.69
20.02
p = .545
(32.71)
(29.60)
0.08,
2.13
2.02
p = .387
(2.34)
(2.15)

--

White
n = 1222

0.11,
p = .741
0.01,
p = .971
8.86,
p = .003
1.77,
p = .470
0.33,
p = .567
0.37,
p = .545

t-test,
p-value

--

28.22
(36.13)a
27.64
(35.61)a
5.61
(12.71)a
8.58
(21.54)a
28.36
(36.23)a
2.34
(2.73)a

M (SD)

Younger
n = 947

18.25
(29.23)b
27.96
(30.07)a
5.30
(6.89)a
6.93
(11.74)b
18.35
(29.50)b
2.00
(2.19)b

M (SD)

13.00
(20.84)c
13.56
(22.64)c
5.21
(7.23)a
5.22
(8.91)bc
13.19
(20.87)c
1.86
(1.78)b

M (SD)

Age Cohorts
Mid-Age
Older
n = 1419 n = 903

64.17,
p < .001
46.25,
p < .001
34.26,
p = .08
2.96,
p = .053
62.33,
p < .001
2.40,
p = .092

F-test,
p-value

--

Number of Outpatient Session Per Month
Any
1.31
0.04 –
1.29
1.32
0.15,
1.40
1.29
2.14,
1.54
1.23
1.19
11.65,
Outpatient
(1.81)
30.42
(1.75)
(1.83) p = .695
(1.94)
(1.79)
p = .143
(1.91)a
(1.79)b
(1.70)b p < .001
Mental
1.40
0.04 –
1.34
1.43
1.16,
1.45
1.41
0.26,
1.58
1.32
1.28
5.81,
Health
(1.92)
30.42
(1.79)
(1.98) p = .283
(2.02)
(1.93)
p = .608
(1.90)a
(1.93)b
(1.93)b p = .003
Substance0.84
0.04 –
0.89
0.92
4.26,
0.99
0.78
0.60,
0.76
0.79
0.90
4.26,
Related
(.97)
9.18
(1.50)
(0.83)
p = .04
(1.24)
(0.82)
p = .61
(1.00)a
(0.99)ab (0.95)bc
p = .04
1.19
0.04 –
0.99
1.24
0.84,
1.46
1.04
2.94,
1.95
1.00
1.23
3.39,
Either/Both
(1.52)
10.14
(1.11)
(1.61)
p = .36
(1.90)
(1.23)
p = .088
(2.96)a
(1.16)b
(1.28)b
p = .036
Note. n = Number; M = Mean; SD = Standard deviation; p-value = Significance level. For age cohorts, different superscripts indicate significant differences
(p ≤ .01) and matching superscripts indicate non-significant differences (p > .01) based upon t-tests between paired cohorts. * Participants who utilized
treatment within two years post-detention. Analyses exclude participants with 0 treatment services.

NonOutpatient

Outpatient

Either/Both

Any
Treatment
Mental
Health
SubstanceRelated

Treatment
Variables

Total Sample
N = 3269*

Table 3.21 Quantity of Treatment Services within Two Years Post-Detention: Demographic Groups

186

186

M (SD) Range

M (SD)

M (SD)

M (SD)

--

Mental
Disorder
n = 1614

Disorder Type
Substance
Comorbid
Disorder
n = 1427
n = 228
--

No
Yes
n = 6823 n = 2092

--

Pre-Detention Treatment

F-test,
F-test,
t-test,
M (SD)
M (SD) M (SD)
M (SD)
M (SD)
p-value
p-value
p-value
Total Number of Treatment Services
19.70 1.0 –
17.76
19.73
23.66
12.55,
19.52
5.58
22.16
30.49,
14.62
25.72
114.48,
Treatment
(30.05) 369.0
(27.27)a
(27.67)b
(35.59)c p < .001
(29.77)a
(6.23)b
(31.98)c p < .001 (23.79)
(35.16) p < .001
Mental
20.04 1.0 –
18.51
20.06
23.31
6.64,
19.83
0.0
20.76
44.93,
15.43
25.11
69.91,
Health
(30.96) 369.0
(28.46)a
(28.72)a
(26.12)b p < .001
(30.03)a
(0.00)b
(32.50)a p < .001 (25.07)
(35.66) p < .001
Substance5.31
1.0 –
4.38
5.79
6.70
2.62,
0.0
5.58
5.22
64.85,
4.73
5.96
6.46,
Related
(8.20) 132.0
(5.32)a
(8.56)ab
(11.27)b p = .073
(0.00)a
(6.23)b
(8.75)b p < .001 (6.195)
(9.94)
p = .011
6.71
1.0 –
4.72
8.77
7.95
0.48,
1.0
6.76
6.70
68.15,
4.94
8.00
34.20,
Either/Both
(13.46) 147.0
(8.05)a
(12.39)b
(17.04)b p = .622
(0.00)a
(13.51)b (13.45)b p < .001
(10.99) (14.90) p < .001
19.88 1.0 –
17.91
20.01
23.85
12.30,
19.62
5.83
22.34
28.18,
14.75
25.97
113.34,
Outpatient
(30.25) 369.0
(27.38)a
(28.02)b
(35.75)c p < .001
(29.89)a
(6.37)b
(32.25)c p < .001 (23.92)
(35.43) p < .001
Non2.06
1.0 –
1.94
1.36
2.09
2.29,
1.91
1.02
2.20
5.41,
1.96
2.13
1.026,
Outpatient
(2.26) 20.0
(1.99)a
(1.06)b
(2.20)a p = .102
(2.22)a
(0.16)b
(2.36)ac p = .002
(2.26)
(2.27)
p = .312
Number of Outpatient Sessions per Month
1.31 0.04 –
1.16
1.43
1.57
17.23,
1.33
0.74
1.37
11.52
0.94
1.75
168.58,
Outpatient
(1.81) 30.42
(1.51)a
(2.23)b
(2.14)bc
p < .001
(1.92)a
(0.85)b
(1.77)a p < .001
(1.29)
(2.20)
p < .001
Mental
1.41 0.04 –
1.24
1.60
1.70
18.12,
1.23
0.00
1.52
28.12,
1.11
1.84
0.25,
Health
(1.92) 30.42
(1.59)a
(2.51)b
(2.31)c
p < .001
(1.92)a
(0.00)b
(1.92)a p < .001
(1.39)
(2.31)
p = .617
Substance0.84 0.04 –
0.75
0.79
1.01
3.22,
0.00
0.74
0.84
3.84,
0.63
1.13
22.98,
Related
(0.97)
9.18
(0.94)a
(0.80)a
(1.09)b p = .041
(0.00)a
(0.85)b
(0.97)b p = .051
(1.09)
(1.34)
p < .001
1.19 0.04 –
0.85
1.48
1.44
15.60,
0.66
0.00
1.20
125.44,
0.94
1.63
5.58,
Either/Both
(1.52) 10.14
(1.07)a
(1.68)b
(1.77)b
p = .001
(0.84)a
(0.00)b
(1.53)c p < .001
(1.29)
(1.73)
p = .018
Note. n = Number; M = Mean; SD = Standard deviation; p-value = Significance level. MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition. For
MAYSI-2 groups and disorder type groups, different superscripts indicate significant differences (p ≤ .01) and matching superscripts indicate non-significant
differences (p > .01) based upon t-tests between paired groups. * Participants who utilized treatment within two years post-detention. Analyses exclude participants
with 0 treatment services.

Treatment
Variables

Total Sample
N = 3269*

MAYSI-2
No
Non-Positive Positive
MAYSI-2
Screen
Screen
n = 1974
n = 332
n = 963

Table 3.22 Quantity of Treatment Services within Two Years Post-Detention: Mental Health Groups

187

187

--

Recidivist+
No
Yes
n = 1563 n = 1706
--

Re-Detained/Incarcerated++
No
Yes
-n = 1100 n = 2085
t-test,
M (SD)
M (SD)
p-value

M (SD)

Range

M (SD)

M (SD)

t-test,
t-test,
M (SD)
M (SD)
p-value
p-value
Total Number of Services
19.70
1.0 –
19.00
20.76
2.69,
16.86
22.30
26.90,
16.77
21.37
17.79,
Any Treatment
(30.05)
369.0
(30.10)
(29.95)
p = .101
(26.55)
(32.72)
p < .001
(25.29)
(32.33)
p < .001
20.04
1.0 –
19.52
20.76
1.10,
18.43
25.12
70.00,
18.51
21.03
6.92,
Mental Health
(30.96)
369.0
(29.77)
(32.50)
p = .294
(25.10)
(32.66)
p < .001
(25.45)
(27.72)
p < .001
Substance5.31
1.0 –
5.49
4.90
0.12,
5.05
5.49
4.80,
5.15
5.39
4.93,
Related
(8.19)
132.0
(8.66)
(6.91)
p = .734
(6.56)
(9.15)
p = .029
(6.62)
(9.80)
p = .027
6.71
1.0 –
6.84
6.41
0.23,
6.20
7.03
0.08,
7.14
6.52
0.03,
Either/Both
(13.46)
147.0
(13.30)
(13.85)
p = .633
(11.97)
(14.35)
p = .774
(15.36)
(12.58)
p = .860
19.88
1.0 –
19.23
20.84
2.17,
17.03
22.50
26.36,
16.91
21.59
17.70,
Outpatient
(30.25)
369.0
(30.37)
(30.06)
p = .141
(26.88)
(32.85)
p < .001
(25.46)
(32.56)
p < .001
Non2.07
1.0 –
1.96
2.20
1.84,
2.11
2.02
0.28,
2.02
2.07
0.08,
Outpatient
(2.26)
20.0
(1.83)
(2.82)
p = .176
(2.46)
(2.12)
p = .597
(2.33)
(2.23)
p = .778
Number of Outpatient Session Per Month
1.31
0.04 –
1.32
1.29
0.25,
1.23
1.39
5.80,
1.28
1.33
0.68,
Any Outpatient
(1.81)
30.42
(1.95)
(1.58)
p = .650
(1.64)
(1.94)
p = .016
(1.57)
(1.93)
p = .409
1.41
0.04 –
1.42
1.38
8.52,
1.29
1.51
1.83,
1.33
1.44
0.30,
Mental Health
(1.92)
30.42
(2.11)
(1.65)
p = .004
(1.72)
(2.09)
p = .177
(1.63)
(2.09)
p = .583
Substance0.84
0.04 –
0.89
0.71
3.25,
1.12
0.79
1.26,
0.94
0.78
3.07,
Related
(0.97)
9.18
(1.04)
(0.77)
p = .072
(1.26)
(0.91)
p = .262
(1.11)
(0.87)
p = .08
1.19
0.04 –
1.20
0.98
0.80,
1.32
1.11
0.74,
1.28
1.13
0.90,
Either/Both
(1.52)
10.14
(1.60)
(1.32)
p = .372
(1.67)
(1.42)
p = .392
(1.57)
(1.51)
p = .372
Note. n = Number; M = Mean; SD = Standard deviation; p-value = Significance level. + Recidivism within two years post-detention. ++ Re-detained in
detention center or incarcerated in prison within two years post-detention * Participants who utilized treatment within two years post-detention. Analyses
exclude participants with 0 treatment services.

Total Sample
N = 3269*

Violent Offender+
No
Yes
n = 1968 n = 1301

Table 3.23 Quantity of Treatment Services within Two Years Post-Detention: Criminal History Groups

188

188

0.84 (0.97)
1.20 (1.52)

SubstanceRelated

Either/Both

Range

4.29 (5.76)
19.05 (28.06)
1.98 (2.06)

1.0 – 147.0
1.0 – 369.0
1.0 – 20.0

1.0 – 17.0

1.0 – 314.0

1.0 – 40.0

1.0 – 39.0

1.0 – 312.0

1.0 – 314.0

2.14 (2.49)

21.10 (33.15)

8.37 (16.65)

6.36 (10.39)

20.74 (22.68)

20.97 (31.14)

0.70 (0.73)
0.74 (0.93)

0.04 – 10.14

1.25 (1.59)

0.04 – 30.42
0.04 – 9.18

1.17 (1.51)

0.04 – 30.42

0.04 – 3.80

0.04 – 4.92

0.04 – 19.45

0.04 – 19.45

1.48 (1.73)

0.97 (1.14)

1.63 (2.33)

1.50 (2.13)

0.81, p = .368

3.53, p = .079

11.23, p < .001

13.99, p < .001

0.90, p = .344

3.56, p = .059

t-test, p-value

--

1.06, p = .305
0.06 – 10.14 11.36, p = .001

0.04 – 9.18

0.04 – 30.42 24.85, p < .001

0.04 – 30.42 12.85, p < .001

1.0 – 20.0

1.0 – 369.0

1.0 – 147.0

1.0 – 132.0

1.0 – 369.0

1.0 – 369.0

Range

Cohort Two
n = 1334
Mean (SD)

Number of Outpatient Sessions per Month

4.36 (5.33)

19.60 (29.08)

1.0 – 369.0
1.0 – 132.0

19.25 (32.03)

Total Number of Treatment Services

Mean (SD)

1.0 – 369.0

Range

Cohort One
n = 1935

Note. n = Number; SD = Standard deviation; p-value = Significance level. * Participants who utilized treatment within 2 years postdetention. Analyses exclude participants with 0 treatment services.

1.40 (1.92)

1.38 (1.79)

Outpatient

Mental Health

2.06 (2.26)

Non-Outpatient

6.71 (13.46)

Either/Both
19.87 (30.25)

5.31 (8.20)

SubstanceRelated

Outpatient

20.04 (30.96)

19.70 (30.05)

Treatment

Mental Health

Mean (SD)

Treatment
Variables

Total Sample
N = 3269*

Cohort Groups

Table 3.24 Quantity of Treatment Services within Two Years Post-Detention: Cohort One versus Cohort Two

189

189

190

Table 3.25 Frequency of Treatment Services within Two Years Post-Detention
Treatment Users

Time between Services (in Days)+

Treatment
Variables

Number (%)
N = 3269*

Mean (SD)

Range

Outpatient

3207 (98.1%)

33.92 (54.93)

1.0 – 680.0

Mental Health

2751 (84.2%)

33.56 (53.27)

1.0 – 680.0

Substance-Related

833 (25.5%)

34.43 (56.56)

1.0 – 593.0

Either/Both

445 (13.6%)

38.03 (72.61)

2.0 – 471.0

Non-Outpatient

1119 (34.2%)

117.69 (128.13)

0.0 – 723.0

Mental Health

486 (14.9%)

117.50 (140.64)

1.0 – 615.0

Substance-Related

233 (9.8%)

415.19 (212.24)

52.0 – 723.0

Either/Both

97 (3.0%)

67.72 (49.37)

0.0 – 137.0

Treatment Users

Length of Stay (in Days)++

373 (11.4%)

11.37 (24.36)

0.0 – 496.0

Mental Health

289 (8.8%)

12.56 (35.16)

0.0 – 496.0

Substance-Related

25 (0.8%)

10.37 (17.34)

0.0 – 61.0

Either/Both

76 (2.3%)

10.73 (24.73)

0.0 – 212.0

Inpatient Stay

Note. N = Number; SD = Standard deviation. + Time between services based upon time
actively involved in treatment, so excludes time to first service from detention release.
++
Length of stay refers to only inpatient treatment services. All outpatient sessions and
emergency department visits were ≤1 day. * Participants who utilized treatment within 2
years post-detention. Analyses exclude participants with 0 treatment services.

Treatment
Variables

Mean
(SD)

Females
n = 844

Gender
Males
n = 2212
--

White
n = 1102

Race
Black
n = 1714
--

Younger
n = 895

Age Cohorts
Mid-Age
Older
n = 1344
n = 817

--

Mean
t-test,
Mean
Mean
t-test,
Mean
Mean
Mean
F-test,
(SD)
p-value
(SD)
(SD)
p-value
(SD)
(SD)
(SD)
p-value
Time Between Treatment Services (in Days)+
33.92
1.0 –
35.54
32.56
1.72,
35.12
32.32
1.61,
29.35
35.51
34.62
3.51,
Outpatient
(54.93) 680.0
(58.25)
(52.08) p = .191
(60.72)
(50.32) p = .205 (44.81)a
(57.67)b
(56.66)b p = .030
33.56
1.0 –
35.64
32.72
1.58,
36.12
32.07
3.07,
30.67
33.53
38.02
3.25,
Mental Health
(53.27) 680.0
(59.97)
(50.00) p = .209
(60.88)
(49.24)
p = .08 (48.84)a
(51.55)a
(61.59)b p = .039
Substance34.43
1.0 –
36.11
33.37
0.12,
28.80
37.19
6.51,
24.57
38.76
31.00
1.45,
Related
(56.56) 593.0
(40.13)
(57.55) p = .726
(35.59)
(65.21) p = .012 (16.34)a
(61.14)b
(54.37)c
p = .237
38.03
2.0 –
60.68
30.72
5.20,
52.55
24.06
1.79,
24.61
37.29
37.79
0.31,
Either/Both
(72.61) 471.0
(114.16)
(53.48) p = .024
(105.30)
(17.53) p = .183 (21.07)a
(72.43)b
(73.96)b p = .736
117.69
0.0 –
150.44
96.14
5.62,
115.27
121.25
0.07,
87.79
116.52
167.97
3.52,
Non-Outpatient
(128.13) 586.0
(136.20)
(118.57) p = .019 (136.72)
(121.96) p = .799 (105.98)a (126.23)b (151.33)c p = .033
Length of Stay (in Days)++
11.37
0.0 –
11.12
11.53
0.03,
12.32
10.25
0.58,
9.58
10.26
15.58
1.85,
Inpatient Stays
(24.36) 290.0
(24.92)
(24.96) p = .88
(27.85)
(22.61) p = .447 (16.61)a
(15.31)a
(41.08)b p = .158
12.56
0.0 –
13.46
12.27
0.08,
13.74
12.16
0.09,
10.24
10.05
22.13
2.83,
Mental Health
(35.16) 496.0
(45.03)
(27.81) p = .78
(31.00)
(42.93) p = .768 (17.54)a
(15.73)a
(70.26)b p = .06
Substance10.37
0.0 –
2.75
12.00
0.93,
8.40
12.18
0.21,
1.50
15.70
9.33
0.98,
Related
(17.34)
61.0
(3.09)
(18.86) p = .345
(18.64)
(19.13) p = .652
(1.29)a
(20.15)b
(17.89)c p = .391
10.73
0.0–
9.47
10.49
0.03,
7.96
14.21
1.16,
5.55
12.43
8.67
0.64,
Either/Both
(24.73) 212.0
(11.71)
(28.82) p = .865
(9.37)
(37.33) p = .284
(3.48)a
(35.78)b
(8.93)ab p = .529
Note. n = Number; SD = Standard deviation; p-value = Significance level. * Participants who utilized treatment within 2 years post-detention. Analyses exclude
participants with 0 treatment services. For age cohorts, different superscripts indicate significant differences (p ≤ .01) and matching superscripts indicate nonsignificant differences (p > .01) based upon t-tests between paired cohorts. + Time between services based upon time actively involved in treatment, so excludes
time to first service from detention release. ++ Length of stay refers to only inpatient treatment services. All outpatient sessions and emergency department visits
were ≤1 day.

Total Sample
N = 3269*
Mean
Range
(SD)

Table 3.26 Frequency of Treatment Services within Two Years Post-Detention: Demographic Groups

191

191

MAYSI-2
Disorder Type
Pre-Detention Treatment
No
Non-Positive Positive
Mental Substance
Comorbid
No
Yes
Total Sample
MAYSI-2
Screen
Screen
-Disorder Disorder
--n = 1388
n = 1498 n = 1558
N = 3269*
n = 1677
n = 366 n = 1013
n = 1504 n = 164
Treatment
Mean
Mean
Mean
Mean
F-test,
Mean
Mean
Mean
F-test,
Mean
Mean
t-test,
Range
Variables
(SD)
(SD)
(SD)
(SD)
p-value
(SD)
(SD)
(SD)
p-value
(SD)
(SD)
p-value
Time Between Treatment Services (in days) +
33.92
1.0 –
33.53
33.19
34.38
0.10,
32.77
27.79
34.80
1.36,
32.75
34.09
1.25,
Outpatient
(54.93) 680.0
(52.65)a
(50.43)a
(58.93)a p = .906 (47.67)a
(28.16)b (62.09)a p = .256 (50.96) (56.75) p = .264
Mental
33.56
1.0 –
32.75
31.79
35.74
0.98,
33.39
0.0
33.93
112.93,
32.41
34.68
1.14,
Health
(53.27) 680.0
(50.88)ab
(42.37)a (60.15)bc p = .375 (48.92)a
(0.00)b
(59.22)a p < .001 (48.13) (57.41) p = .286
Substance34.43
1.0 –
33.20
41.59
31.00
0.77,
0.0
28.93
37.37
229.49,
30.37
39.60
2.52,
Related
(56.56) 593.0
(444.57)a (78.61)b
(57.11)a p = .463 (0.00)a
(36.96)b (65.37)c p < .001 (36.44) (77.12) p = .113
38.03
2.06 –
54.05
29.63
27.11
2.93,
24.75
0.0
36.15
320.09
43.55
30.61
1.59,
Either/Both
(72.61) 471.0
(98.58)a
(46.40)b
(49.68)b p = .056 (1.00)a
(0.00)b
(69.37)c p < .001 (87.53) (51.49) p = .211
Non117.69
0.0 –
106.66
65.62
127.35
0.66,
113.45
0.0
117.69
583.32,
86.41
129.74
4.43,
Outpatient
(128.13) 586.0 (123.20)a (51.22)b
(131.71)a p = .516 (129.47)a
(0.00)b (128.13)a p < .001 (106.83) (135.31) p = .037
Length of Stay (in Days)++
11.37
0.0 –
11.27
13.44
11.28
0.06,
10.51
22.285
11.52
0.76,
14.54
8.34
6.39,
Inpatient
(24.36) 290.0
(29.74)a
(14.29)a
(16.46)a p = .94 (14.46)a
(21.06)b (29.35)a p = .471 (34.11) (13.08) p = .012
Mental
12.56
0.0 –
13.67
14.77
11.13
0.18,
10.51
0.0
14.64
20.04,
18.10
8.36
5.89,
Health
(35.16) 496.0
(44.72)a
(16.35)a
(17.14)b p = .834 (14.46)a
(0.00)b (47.09)c p < .001 (52.69) (13.11) p = .016
Substance10.37
0.0 –
7.09
1.50
14.92
0.86,
0.0
22.29
6.26
304.38,
12.51
7.88
0.75,
Related
(17.34)
61.0
(17.67)a
(0.71)b
(18.52)c p = .435 (0.00)a
(21.06)b (14.56)c p < .001 (33.78)
(9.82) p = .391
10.73
0.0–
6.95
15.75
13.36
0.75,
0.0
0.0
10.14
12.36
7.86
0.48,
Either/Both
-(24.73) 212.0
(6.01)a
(9.53)b
(36.52)b p = .477 (0.00)a
(0.00)a (24.46)b
(17.91) (16.58) p = .496
Note. n = Number; SD = Standard deviation; p-value = Significance level; MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition. * Participants
who utilized treatment within 2 years post-detention. Analyses exclude participants with 0 treatment services. For MAYSI-2 groups and disorder groups,
different superscripts indicate significant differences (p ≤ .01) and matching superscripts indicate non-significant differences (p > .01) based upon t-tests between
paired groups. + Time between services based upon time actively involved in treatment, so excludes time to first service from detention release. ++ Length of stay
refers to only inpatient treatment services. All outpatient sessions and emergency department visits were ≤1 day.

Table 3.27 Frequency of Treatment Services within Two Years Post-Detention: Mental Health Groups

192

192

Violent Offender
Recidivista
Re-Detained/Incarceratedb
Total Sample
No
Yes
No
Yes
No
Yes
---N = 3269*
n = 1836 n = 1220
n = 1449 n = 1607
n = 1159 n = 1897
Treatment
Mean
Mean
Mean
t-test,
Mean
Mean
t-test,
Mean
Mean
t-test,
Range
Variables
(SD)
(SD)
(SD)
p-value
(SD)
(SD)
p-value
(SD)
(SD)
p-value
Time Between Treatment Services (in days) +
33.92
1.0 –
32.63
34.06
0.28,
34.74
32.21
1.53,
35.61
32.17
2.59,
Outpatient
(54.93)
680.0
(55.85)
(53.35)
p = .599
(58.83)
(49.00)
p = .217
(62.304)
(48.50)
p = .108
33.56
1.0 –
33.56
33.66
0.01,
35.56
31.83
3.05,
36.38
31.93
4.08,
Mental Health
(53.27)
680.0
(55.49)
(50.02)
p = .961
(58.35)
(48.05)
p = .081
(60.38)
(48.33)
p = .043
Substance34.43
1.0 –
31.51
36.71
1.71,
30.01
37.52
1.79,
30.01
36.35
1.22,
Related
(56.56)
593.0
(50.70)
(65.43)
p = .192
(51.92)
(58.74)
p = .182
(53.23)
(57.06)
p = .270
38.03
2.06 –
36.23
35.77
0.002,
48.05
28.05
3.75,
46.16
30.20
2.29,
Either/Both
(72.61)
471.0
(71.01)
(65.23)
p = .968
(94.66)
(43.27)
p = .054
(94.06)
(48.65)
p = .131
117.69
0.0 –
119.02
115.98
0.02,
140.72
98.67
13.43,
96.67
159.73
7.11,
Non-Outpatient
(128.13) 586.0
(133.00) (122.76) p = .896
(147.70) (106.78) p < .001
(115.89)
(141.95)
p < .001
Length of Stay (in Days)++
11.37
0.0 –
11.04
11.86
0.10,
11.89
10.95
0.23,
13.36
10.23
1.37,
Inpatient
(24.36)
290.0
(23.45)
(25.67)
p = .753
(31.96)
(17.42)
p = .716
(35.28)
(16.36)
p = .243
12.56
0.0 –
12.03
13.78
0.17,
14.76
11.26
0.69,
14.73
10.57
2.00,
Mental Health
(35.16)
496.0
(27.84)
(45.09)
p = .68
(50.79)
(18.39)
p = .407
(55.82)
(17.27)
p = .158
Substance10.37
0.0 –
6.37
22.0
4.58,
10.0
11.25
0.03,
10.27
10.80
0.01,
Related
(17.34)
61.0
(9.71)
(28.41)
p = .043
(17.28)
(18.82)
p = .861
(18.42)
(17.71)
p = .942
10.73
0.0–
11.66
6.23
0.78,
7.38
12.08
0.59,
8.43
11.086
0.21,
Either/Both
(24.73)
212.0
(28.32)
(7.72)
p = .380
(7.83)
(32.42)
p = .446
(8.32)
(29.88)
p = .648
Note. n = Number; SD = Standard deviation; p-value = Significance level. * Participants who utilized treatment within two years post-detention. Analyses
exclude participants with 0 treatment services. a Recidivism within two years post-detention. b Re-detained in detention center or incarcerated in prison
within two years post-detention. + Time between services based upon time actively involved in treatment, so excludes time to first service from detention
release. ++ Length of stay refers to only inpatient treatment services. All outpatient sessions and emergency department visits were ≤1 day.

Table 3.28 Frequency of Treatment Services within Two Years Post-Detention: Criminal History Groups

193

193

194
Table 3.29 Frequency of Treatment Services within Two Years Post-Detention:
Cohort One versus Cohort Two

Total Sample
N = 3269*
Treatment
Variables

Cohort One
n = 1935

Cohort Groups
Cohort Two
n = 1334

Mean (SD) Range Mean (SD) Range Mean (SD)

--

Range

t-test,
p-value

Time Between Treatment Services (in Days)+
Outpatient

33.92
(54.93)

1.0 –
680.0

33.14
(55.02)

1.0 –
680.0

33.74
(52.17)

1.0 –
607.0

1.21,
p = .298

Mental
Health

33.56
(53.27)

1.0 –
680.0

32.40
(52.07)

1.0 –
680.0

33.77
(49.33)

1.0 –
607.0

0.38,
p = .534

SubstanceRelated

34.43
(56.56)

1.0 –
593.0

30.71
(38.59)

1.0 –
329.0

36.19
(65.23)

2.0 –
593.0

0.92,
p = .338

Either/Both

38.03
(72.61)

2.0 –
471.0

64.14
(131.84)

2.0 –
388.0

29.10
(48.36)

2.0 –
471.0

5.73,
p = .018

117.69
(128.13)

0.0 –
586.0

104.59
(128.05)

0.0 –
586.0

129.81
(123.97)

2.0 –
585.0

1.30,
p = .256

NonOutpatient

Length of Stay (in Days)++
Inpatient
Stays

11.37
(24.36)

0.0 –
290.0

11.38
(28.29)

0.0 –
290.0

11.25
(16.92)

0.0 –
91.0

0.00,
p = .988

Mental
Health

12.56
(35.16)

0.0 –
496.0

13.71
(42.29)

0.0 –
496.0

10.54
(16.84)

0.0 –
91.00

0.60,
p = .438

10.37
0.0 –
2.30
0.0 –
15.12
0.0 –
3.81,
(17.34)
61.0
(2.36)
7.0
(20.55)
61.00
p = .062
10.73
0.0 –
6.87
0.0 –
14.57
0.0 –
2.01,
Either/Both
(24.73) 212.0
(5.35)
30.0
(34.39)
212.0
p = .160
Note. n = Number; SD = Standard deviation; p-value = Significance level. * Participants
who utilized treatment within 2 years post-detention. Analyses exclude participants with 0
treatment services. + Time between services based upon time actively involved in
treatment, so excludes time to first service from detention release. ++ Length of stay refers
to only inpatient treatment services. All outpatient sessions and emergency department
visits were ≤1 day.
SubstanceRelated

195
Table 3.30 Intensity of Treatment Services within Two Years Post-Detention:
Treatment Type, Treatment Setting, and User Intensity Level
Number (%)
Treatment Categories
N = 8915
Treatment Type
-Only Mental Health Treatment
2084 (23.4%)
Only Substance-Related Treatment
370 (4.2%)
Only Either/Both Treatment
58 (0.7%)
Mental Health and Substance-Related Treatment
319 (3.6%)
Mental Health and Either/Both Treatment
197 (2.2%)
Substance-Related and Either/Both Treatment
41 (0.5%)
All Three Treatment Types
200 (2.2%)
Treatment Setting
-Only Outpatient Treatment
2554 (28.6%)
Only Non-Outpatient Treatment
70 (0.8%)
Outpatient and Inpatient Treatment
223 (2.5%)
Outpatient and ED Visit
291 (3.3%)
Outpatient, Inpatient, and ED Visit
139 (1.6%)
Number of Treatment Settings
-One Setting
2614 (29.3%)
Two Settings
516 (5.8%)
Three Settings
139 (1.6%)
Outpatient User Intensity Level
3207 (36.0%)
Non-User
5708 (64.0%)
Low User (1-2 Sessions)
700 (7.9%)
Low-to-Moderate User (3-7 Sessions )
771 (8.6%)
Moderate User (8-12 Sessions)
446 (5.0%)
Moderate-to-High User (13-17 Sessions)
296 (3.3%)
High User (18-22 Sessions)
226 (2.5%)
Extreme User (≥23 Sessions)
830 (9.3%)
Non-Outpatient User Intensity Level
723 (8.13%)
Non-User
8204 (92.0%)
Low User (1 Stay/Visit)
449 (5.0%)
Low-to-Moderate User (2 Stays/Visits)
121 (1.4%)
Moderate User (3-4 Stays/Visits)
75 (0.9%)
High User (5-6 Stay/Visit)
33 (0.4%)
Extreme User (≥7 Stays/Visits)
33 (0.3%)
Note. N = Number; % = Percentage; ED = Emergency Department.

1-2 Sessions

Three Settings

Two Settings

One Setting
(Outpatient)
One Setting
(Non-Outpatient)

All Three Types

Mental Health and
Substance-Related
Mental Health and
Either/Both
Substance-Related
and Either/Both

Only Either/Both

Only SubstanceRelated

Only Mental Health

192
(5.6%)

700
(7.9%)

508
(9.2%)

656
1891
(25.8%) (34.3%)
21
46
(0.6%) (0.8%)
208
308
(6.1%) (5.6%)
49
90
(1.4%) (1.6%)

666
1418
(19.6%) (25.7%)
59
311
(1.7%) (5.6%)
11
47
(0.3%) (0.9%)
95
224
(2.8%) (4.1%)
55
142
(1.6%) (2.6%)
4
37
(0.1%) (0.7%)
44
156
(1.3%) (2.8%)

2547
(28.6%)
67
(0.8%)
516
(5.8%)
139
(1.6%)

2084
(23.4%)
370
(4.2%)
58
(0.7%)
319
(3.6%)
197
(2.2%)
41
(0.5%)
200
(2.2%)

N (%)
N (%)
n = 3164 n = 5185

Treatment Type
44.03,
745
1183
p < .001
(23.5%) (22.8%)
80.58,
120
213
p < .001
(3.8%)
(4.1%)
9.10,
25
30
p = .003
(0.8%)
(0.6%)
9.79,
137
153
p = .002
(4.3%)
(3.0%)
8.92,
84
100
p = .003
(2.7%)
(1.9%)
14.06,
17
19
p < .001
(0.2%)
(0.4%)
22.58,
93
93
p < .001
(2.9%)
(1.8%)
Treatment Setting
230.75,
852
1492
p < .001 (26.9%) (28.8%)
1.32,
43
19
p = .312
(1.4%)
(0.4%)
1.10,
247
230
p = .304
(7.8%)
(4.4%)
0.50,
78
52
p = .538
(2.5%)
(1.0%)
Outpatient User Intensity Level
36.92,
250
383
p < .001
(7.9%)
(7.4%)

χ2,
p-value

Black

White

Males

--

Race

Gender

N (%)
N (%)
N (%)
N
=
8915
n
=
3400
n
= 5515
Treatment Variables

Total
Sample Females

0.74,
p = .394

3.32,
p = .036
26.26,
p < .001
41.44,
p < .001
27.41,
p < .001

0.59,
p = .229
0.51,
p = .257
1.34,
p = .154
11.15,
p < .001
4.81,
p = .031
1.34,
p = .162
11.84,
p < .001

χ2,
p-value

--

Older

141
(6.2%)a

748
(32.7%)a
12
(0.5%)a
145
(6.3%)a
43
(1.9%)a

460
(15.0%)c
188
(6.1%)b
27
(0.9%)bc
100
(3.3%)a
56
(1.8%)a
17
(0.6%)b
55
(1.8%)a

309
(8.7%)b

250
(8.2%)b

1105
694
(31.0%)a (22.7%)b
28
27
(0.8%)a (0.9%)a
226
145
(6.3%)a (4.7%)b
61
35
(1.7%)a (1.1%)a

752
872
(32.9%)a (24.4%)b
31
151
(1.4%)a (4.2%)b
6
25
(0.3%)a (0.7%)ab
71
148
(3.1%)a
(4.1%)a
42
99
(1.8%)a (2.8%)b
2
22
(0.1%)a (0.6%)b
43
102
(1.9%)a (2.9%)b

N (%)
N (%)
N (%)
n = 2285 n = 3566 n = 3059

Younger Mid-Age

Age Cohorts

Table 3.31 Intensity of Treatment Services within Two Years Post-Detention: Demographic Groups

12.64
p = .002

81.49,
p < .001
2.33,
p = .312
9.41,
p = .01
5.50,
p = .064

236.54,
p < .001
75.53,
p < .001
8.02,
p = .018
5.70,
p = .058
8.79,
p = .012
9.46,
p = .106
10.31,
p = .006

χ 2,
p-value

--

196

196

226
(6.6%)
125
(3.7%)
80
(2.4%)
63
(1.9%)
248
(7.3%)

545
27.86,
279
429
0.75,
171
346
254
9.41,
(9.9%) p < .001 (8.8%)
(8.3%) p = .395
(7.5)ac
(9.7%)b (8.3%)ab
p = .01
321
20.34,
181
239
5.08,
115
200
131
6.08,
8-12 Sessions
(5.8%) p < .001
(5.7%)
(4.6%) p = .024 (5.0%)a
(5.6%)a (4.3%)b p = .048
216
16.02,
140
154
2.30,
81
141
74
12.54
13-17 Sessions
(3.9%) p < .001 (4.42%
(3.0%) p = .349 (3.5%)ab
(4.0%)b (2.4%)ac p = .002
163
10.35,
75
124
0.004,
75
105
46
20.71,
18-22 Sessions
(3.0%) p < .001
(2.4%)
(2.4%) p = .99
(3.3%)a
(2.9%)a (1.5%)b p < .001
582
26.45,
323
462
3.88,
364
318
148
191.39,
≥23 Sessions
(10.6%) p < .001 (10.2%) (8.9%) p = .053 (15.9%)a (8.9%)b (4.8%)b p < .001
Non-Outpatient User Intensity Level
449
176
273
0.23,
216
194
40.05,
117
202
130
6.92,
1 Stay/Visit
(5.0%)
(5.2%) (5.0%) p = .654
(6.8%)
(3.7%) p < .001 (5.1%)a
(5.7%)a
(4.2%)b p = .031
121
47
74
0.03,
78
38
43.04,
35
49
37
1.02,
2 Stays/Visits
(1.4%)
(1.4%) (1.3%) p = .925 (2.5%)
(0.7%) p < .001 (1.5%)a
(1.4%)a
(1.2%)a p = .600
75
29
46
0.01,
43
27
16.61,
19
36
20
2.50,
3-4 Stays/Visits
(0.9%)
(0.9%) (0.8%) p = .906 (1.4%)
(0.5%) p < .001 (0.8%)a
(1.0%)b
(0.7%)a p = .287
33
11
22
0.32,
12
20
0.002,
12
13
2.46,
5-6 Stay/Visit
8 (0.3%)b
(0.4%)
(0.3%) (0.4%) p = .720
(0.4%)
(0.4%) p = .998 (0.5%)a
(0.4%)a
p = .292
33
10
23
0.30,
20
12
5.43,
14
12
7
2.89,
≥7 Stays/Visits
(0.1%)
(0.3%) (0.5%) p = .777 (0.6%)
(0.3%) p = .040 (0.6%)a
(0.03%)ab (0.2%)bc p = .236
Note. n = Number; % = Percentage; χ2 = 2 x 2 chi-square test; p-value = Significance level. For age cohorts, different superscripts indicate
significant differences (p ≤ .01) and matching superscripts indicate non-significant differences (p > .01) based upon 2 x 2 chi-square tests between
paired cohorts.

3-7 Sessions

771
(8.6%)
446
(5.0%)
296
(3.3%)
226
(2.5%)
830
(9.3%)

Table 3.31 Continued.

197

197

67
(0.8%)
516
(5.8%)
139
(1.6%)

One Setting
(Non-Outpatient)

Three Settings

Two Settings

2547
(28.6%)

2084
(23.4%)
370
(4.2%)
58
(0.7%)
319
(3.6%)
197
(2.2%)
41
(0.5%)
200
(2.2%)

One Setting
(Outpatient)

All Three Types

Mental Health and
Substance-Related
Mental Health and
Either/Both
Substance-Related
and Either/Both

Only Either/Both

Only Mental
Health
Only SubstanceRelated

Treatment
Variables

48
(0.7%)a
302
(4.4%)a
75
(1.1%)a

1551
(22.7%)a

1362
(19.9%)a
212
(3.1%)a
26
(0.4%)a
193
(2.8%)a
73
(1.1%)a
16
(0.2%)a
92
(1.3%)a

3
(0.5%)ab
30
(5.0%)a
3
(0.5%)a

296
(29.1%)b

191
(31.7%)b
50
(8.3%)b
4
(0.7%)a
28
(4.6%)b
28
(4.6%)b
10
(1.7%)ab
21
(3.5%)b

16
(1.1%)bc
184
(12.5%)b
61
(4.1%)b

700
(47.5%)c

531
(36.0%)b
108
(7.3%)b
28
(1.9%)b
98
(6.6%)b
96
(6.5%)b
15
(1.0%)bc
87
(5.9%)b

MAYSI-2
Total
No
Non-Positive Positive
Sample MAYSI-2
Screen
Screen
N (%)
N (%)
N (%)
N (%)
N = 8915 n = 6838 n = 603 n = 3164

0.95,
p = .229
145.48,
p < .001
77.84,
p < .001

12
(0.4%)ac
171
(5.2%)a
28
(0.9%)a

16
(2.6%)b
22
(3.6%)b
0
(0.0%)a

39
(1.5%)ab
323
(12.6%)c
111
(4.3%)b

73.48,
p < .001
376.55,
p < .001
187.19,
p < .001

42
(0.6%)
239
(2.5%)
51
(0.7%)

25
(1.2%)
277
(13.3%)
88
(4.2%)

7.41,
p = .009
282.95,
p < .001
126.50,
p < .001

Disorder Type
Pre-Detention Treatment
Mental Substance
-Comorbid
-No
Yes
-Disorder Disorder
χ2,
N (%)
N (%)
N (%)
χ 2,
N (%)
N (%)
χ 2,
p-value n = 3257 n = 617 n = 2567 p-value
n = 6838 n = 2077
p-value
Treatment Type
200.49,
1609
0
475
2205.86,
1166
918
655.44,
p < .001 (49.4%)a (0.0%)b (18.5%)c p < .001 (17.1%) (44.2%)
p < .001
82.34,
0
228
142
1929.37,
225
115
13.09
p < .001
(0.0%)a (37.0%)b (5.5%)c
p < .001 (3.7%)
(5.5%)
p < .001
43.31,
0
0
35
23
8.74,
58 (0.7%)b
-p < .001
(0.0%)a
(0.0%)a
(0.5%)
(1.1%)
p = .005
53.58,
0
0
319
143
176
188.10,
-p < .001
(0.0%)a
(0.0%)a (12.4%)b
(2.1%)
(8.5%)
p < .001
184.11,
5
0
192
463.47,
83
114
134.73,
p < .001
(0.2%)a
(0.0%)a (7.5%)b
p < .001 (42.1%) (57.9%)
p < .001
36.55,
0
0
20
21
17.97,
41 (1.6%)b
-p < .001
(0.0%)a
(0.0%)a
(0.3%)
(1.0%)
p < .001
119.34,
0
0
200
72
128
189.67,
-p < .001 (0.0%)a
(0.0%)a (7.8%)b
(1.1%)
(6.2%)
p < .001
Treatment Setting
498.41,
1403
191
953
1418.44,
1445
1102
797.87,
p < .001 (43.1%)a (31.0%)b (37.1%)c
p < .001 (21.1%) (53.1%)
p < .001

Table 3.32 Intensity of Treatment Services within Two Years Post-Detention: Mental Health Groups

198

198

3-7 Sessions

771
(8.6%)
446
(5.0%)
296
(3.3%)
226
(2.5%)
830
(9.3%)

460 (6.7%)a

229
(8.9%)a

327.86,
p < .001

483
(7.1%)

217
(10.4%)

25.22,
p < .001

81
230
139.68,
388
76
307
323.92,
482
289
95.04,
(13.4%)b (15.6%)b p < .001 (11.9%)ac (12.3%)b (12.0%)ab p = .101 (7.0%) (37.5%) p < .001
274
44
128
62.97,
223
31
102
186.52,
259
187
91.19,
8-12 Sessions
(4.05%)a (7.3%)a
(8.7%)b p < .001
(50.0%)a (5.0%)b (7.5%)c
p < .001 (3.8%)
(9.0%)
p < .001
30
92
57.14,
134
7
155
159.51,
151
145
113.06,
13-17 Sessions
174 (2.5%)a
(5.0%)ab (6.2%)bc p < .001
(4.1%)a
(1.1%)b (6.0%)a
p < .001 (2.2%)
(7.0%)
p < .001
24
66
35.86,
96
9
121
118.72,
93
133
164.01,
18-22 Sessions
136 (2.0%)a
(4.0%)ab (4.5%)bc p < .001 (2.9%)ab
(1.5%)a (4.7%)bc p < .001 (41.2%) (6.4%)
p < .001
84
295
274.92,
400
7
423
492.73,
306
524
812.68,
≥23 Sessions
451 (6.6%)a
(13.9%)a (20.0%)c p < .001
(12.3%)a
(1.1%)b (16.5%)c p < .001 (4.5%) (25.2%) p < .001
Non-Outpatient User Intensity Level
449
262
30
157
117.92,
240
37
272
301.92,
215
234
219.73,
1 Stay/Visit
(5.0%)
(3.8%)a
(5.0%)a (10.7%)b p < .001
(4.3%)a
(6.0%)a (10.6%)b p < .001 (3.1%) (11.3%) p < .001
121
69
3
49
52.12,
33
1
122.65,
56
63.53,
2 Stays/Visits
87 (3.4%)b
65 (3.1%)
(1.4%)
(1.0%)a
(0.5%)a (3.3%)b
p < .001
(1.0%)a
(0.2%)a
p < .001 (0.8%)
p < .001
75
46
1
28
25.42,
18
9
57
88.01,
26
74.79,
3-4 Stays/Visits
49 (2.4%)
(0.9%)
(0.7%)ac (0.2%)ab (1.9%)c
p < .001
(0.6%)a
(1.5%)ab (2.2%)bc p < .001 (0.4%)
p < .001
33
16
2
15
20.22,
7
0
42.36,
12
30.16,
5-6 Stay/Visit
26 (1.0%)b
21 (1.0%)
(0.4%)
(0.2%)a
(0.3%)a
(1.0%)b
p < .001 (0.2%)a
(0.0%)a
p < .001 (0.2%)
p < .001
33
22
0
11
3.67,
8
9
25
20.20,
16
6.80,
≥7 Stays/Visits
17 (0.8%)
(0.3%) (0.3%)ab (0.0%)ac (0.7%)b
p = .159
(0.3%)a
(1.5%)bc (0.9%)ba p < .001 (0.2%)
p = .016
Note. n = Number; % = Percentage; χ2 = 2 x 2 chi-square test; p-value = Significance level; MAYSI-2 = Massachusetts Youth Screening Instrument-2nd
Edition. For MAYSI-2 groups and disorder groups, different superscripts indicate significant differences (p ≤ .01) and matching superscripts indicate nonsignificant differences (p > .01) based upon 2 x 2 chi-square tests between paired groups.

1-2 Sessions

Outpatient User Intensity Level
700
69
152
29.85,
373
98
479 (7.0%)a
(7.9%)
(11.4%)b (10.3%)b p < .001
(11.4%)a (15.9%)b

Table 3.32 Continued.

199

199

Two Settings

One Setting
(Outpatient)
One Setting
(Non-Outpatient)

All Three Types

Mental Health and
Substance-Related
Mental Health and
Either/Both
Substance-Related
and Either/Both

Only Either/Both

Only Mental
Health
Only SubstanceRelated

Treatment
Variables

Violent Offender
Recidivist+
Total
Sample
No
Yes
-No
Yes
2
N (%)
N (%)
N (%)
χ,
N (%)
N (%)
N = 8915 n = 5661 n = 3254 p-value n = 5447 n = 3468
Treatment Type
2084
1147
937
84.01,
1080
1004
(23.4%) (20.3) (28.8%) p < .001 (19.8%) (29.0%)
370
259
111
7.04,
187
183
(4.2%)
(4.6%)
(3.4%) p = .004 (3.4%) (5.3%)
58
35
23
0.25,
32
26
(0.7%)
(0.6%)
(0.7%) p = .682 (0.6%) (0.7%)
319
223
96
5.85,
103
216
(3.6%)
(3.9%)
(3.0%) p = .023 (1.9%) (6.2%)
197
125
72
0.001,
75
122
(2.2%)
(2.2%)
(2.2%)
p = .99
(1.4%) (3.5%)
41
31
10
2.61,
19
22
(0.5%)
(0.5%)
(0.3%) p = .142 (0.3%) (0.6%)
200
148
52
9.73,
67
133
(2.2%)
(2.6%)
(1.6%) p = .002 (1.2%) (3.8%)
Treatment Setting
2547
1523
1024
21.29,
1266
1281
(28.6%) (26.9%) (31.5%) p < .001 (23.2%) (37.0%)
67
48
19
1.93,
29
38
(0.8%)
(0.8%)
(0.6%) p = .202 (0.5%) (1.1%)
516
311
205
2.46,
213
303
(5.8%)
(5.5%)
(6.3%) p = .120 (3.9%)
(8.7%)
195.53,
p < .001
9.05,
p = .004
90.54,
p < .001

98.46,
p < .001
18.11,
p < .001
0.86,
p = 348
115.54,
p < .001
44.95,
p < .001
3.77,
p = .055
65.57,
p < .001

-χ,
p-value
2

948
1599
(23.7%) (32.6%)
22
45
(0.5%) (0.9%)
177
339
(4.4%) (6.9%)

818
1266
(20.4%) (25.8%)
149
221
(3.7%) (4.5%)
27
31
(0.7%) (0.6%)
71
248
(1.8%) (5.1%)
56
141
(1.4%) (2.9%)
14
27
(0.3%) (0.6%)
49
151
(1.2%) (3.1%)

86.26,
p < .001
4.02,
p = .049
25.23,
p < .001

36.07,
p < .001
3.46,
p = .035
0.06,
p = .895
69.03,
p < .001
22.30,
p = .011
1.33,
p = .023
34.68,
p < .001

Re-Detained/Incarcerated++
No
Yes
-N (%)
N (%)
χ2,
n = 4010 n = 4905 p-value

Table 3.33 Intensity of Treatment Services within Two Years Post-Detention: Criminal History Groups

200

200

85
(1.5%)

54
0.34,
56
83
25.73,
38
101
17.76,
(1.7%) p = .594 (1.0%) (2.4%) p < .001 (0.9%) (2.1%) p < .001
Outpatient User Intensity Level
700
418
282
4.69,
392
308
8.311,
291
409
3.57,
1-2 Sessions
(7.9%)
(7.4%)
(8.7%) p = .033 (7.2%) (8.9%) p = .004 (7.3%) (8.3%) p = .063
771
475
296
1.20,
385
386
44.26,
294
S477
15.99,
3-7 Sessions
(8.6%)
(8.4%)
(9.1%) p = .257 (7.1%) (11.1%) p < .001 (7.3%) (9.7%) p < .001
446
279
167
0.10,
215
231
32.84,
162
284
14.22,
8-12 Sessions
(5.0%)
(4.9%)
(5.1%) p = .687 (3.9%) (6.7%) p < .001 (4.0%) (5.8%) p < .001
296
193
103
0.38,
134
162
32.28,
108
188
8.92,
13-17 Sessions
(3.3%)
(3.4%)
(3.2%) p = .581 (2.5%)
(4.7%) p < .001 (2.7%) (3.8%) p = .003
226
142
84
0.05,
93
133
38.82,
72
154
16.11,
18-22 Sessions
(2.5%)
(2.5%)
(2.6%) p = .834 (1.7%)
(3.8%) p < .001 (1.8%) (3.1%) p < .001
830
461
369
25.00,
344
486
148.73,
257
573
72.65,
≥23 Sessions
(9.3%)
(8.1%) (11.3%) p < .001 (6.3%) (14.0%) p < .001 (6.4%) (11.7%) p < .001
Non-Outpatient User Intensity Level
449
276
173
0.84,
187
262
75.264,
156
293
20.02,
1 Stay/Visit
(5.0%) (4.9%)
(5.3%)
p = 365
(3.4%) (7.6%) p < .001 (3.9%) (5.9%) p < .001
121
78
43
0.05,
46
75
27.50,
27
94
25.47,
2 Stays/Visits
(1.4%) (1.4%)
(1.3%)
p = .85
(0.8%) (2.2%) p < .001 (0.7%) (1.9%) p < .001
75
45
30
0.40,
32
43
10.81,
27
48
2.47,
3-4 Stays/Visits
(0.9%) (0.8%)
(0.9%) p = .548 (0.6%) (1.2%) p < .001 (0.7%) (1.0%) p = .130
33
24
9
1.21,
10
23
13.22,
8
25
5.76,
5-6 Stay/Visit
(0.4%) (0.4%)
(0.3%) p = .365 (0.2%) (0.7%) p < .001 (0.2%) (0.5%) p = .021
33
17
16
0.18,
17
16
0.001,
12
21
0.22,
≥7 Stays/Visits
(0.3%) (0.3%)
(0.5%) p = .779 (0.3%) (0.4%) p = .999 (0.3%) (0.5%) p = .783
Note. n = Number; % = Percentage; χ2 = 2 x 2 chi-square test; p-value = Significance level. + Recidivism within 2 years postdetention. ++ Re-detained in detention center or incarcerated in prison within 2 years post-detention.

Three Settings

139
(1.6%)

Table 3.33 Continued

201

201

202

Table 3.34 Intensity of Treatment Services within Two Years Post-Detention: Cohort
One versus Cohort Two
Cohort Groups
Cohort Two
2006-2011
Number (%)
n = 2341

Total Sample Cohort One
-1998-2011
1998-2005
Number (%) Number (%)
χ2,
Treatment Variables
N = 8915
n = 6574
p-value
Treatment Type
Only Mental Health
2084 (23.4%) 1354 (20.6%)
730 (31.2%)
106.84, p < .001
Only Substance-Related
370 (4.2%)
187 (2.8%)
183 (7.8%)
197.30, p < .001
Only Either/Both
58 (0.7%)
23 (0.3%)
35 (1.5%)
35.08, p < .001
MH and S-R
319 (3.6%)
191 (2.9%)
128 (5.5%)
32.58, p < .001
MH and E/B
197 (2.2%)
69 (1.0%)
128 (5.5)
155.93, p < .001
S-R and E/B
41 (0.5%)
14 (0.2%)
27 (1.2%)
33.34, p < .001
All Three Types
200 (2.2%)
96 (1.5%)
104 (4.4%)
70.01, p < .001
Treatment Setting
One Setting – Outpatient 2547 (28.6%) 1518 (23.1%) 1029 (44.0%) 379.30, p < .001
One Setting – Non67 (0.8%)
41 (0.7%)
26 (1.0%)
4.32, p = .032
Outpatient
Two Settings
516 (5.8%)
305 (4.6%)
211 (9.0%)
60.90, p < .001
Three Settings
139 (1.6%)
71 (1.08%)
68 (2.9%)
37.59, p < .001
Outpatient User Intensity Level
1-2 Sessions
700 (7.9%)
454 (6.9%)
246 (10.5%) 30.96, p < .001
3-7 Sessions
771 (8.6%)
437 (6.6%)
334 (14.3%) 126.87, p < .001
8-12 Sessions
446 (5.0%)
265 (4.0%)
181 (7.7%)
49.74, p < .001
13-17 Sessions
296 (3.3%)
168 (2.6%)
128 (5.5%)
45.61, p < .001
18-22 Sessions
226 (2.5%)
133 (2.0%)
93 (4.0%)
26.55, p < .001
≥23 Sessions
830 (9.3%)
478 (7.3%)
352 (15.0%) 123.28, p < .001
Non-Outpatient User Intensity Level
1 Stay/Visit
449 (5.0%)
267 (4.1%)
182 (7.8%)
49.76, p < .001
2 Stays/Visits
121 (1.4%)
65 (1.0%)
56 (2.4%)
23.40, p < .001
3-4 Stays/Visits
75 (0.9%)
42 (0.6%)
33 (1.4%)
12.29, p = .001
5-6 Stay/Visit
33 (0.4%)
18 (0.3%)
15 (0.6%)
6.30, p = .017
≥7 Stays/Visits
33 (0.3%)
19 (0.4%)
14 (0.6%)
5.85, p = .022
2
Note. n = Number; % = Percentage. χ = 2 x 2 chi-square test; p-value = Significance
level; MH = Mental Health; S-R = Substance-Related; E/B = Either/Both.

203
Table 3.35 Outpatient Treatment Dropouts and Treatment Gaps within Two Years
Post-Detention
Number (%)
Treatment Variables
N = 3207*
Treatment Dropout within 1-3 Sessions
721 (22.5%)
Mental Health (n = 2751)
512 (16.0%)
Substance-Related (n = 833)
160 (5.0%)
Either/Both (n = 445)
48 (1.5%)
Dropout after 1 Session
342 (10.7%)
Mental Health
243 (7.6%)
Substance-Related
74 (2.3%)
Either/Both
24 (0.7%)
Dropout after 2 Sessions
230 (7.2%)
Mental Health
163 (5.1%)
Substance-Related
51 (1.6%)
Either/Both
16 (0.5%)
Dropout after 3 Sessions
149 (4.6%)
Mental Health
106 (3.3%)
Substance-Related
35 (1.1%)
Either/Both
8 (0.2%)
Reasons for Dropouta
n = 721
Detention/Prison
110 (16.4%)
Inpatient Stay
4 (0.6%)
No Clear Reason
605 (84.2%)
Gaps between Outpatient Sessions
N = 2789b
No gaps
1656 (59.4%)
Gap(s) between Sessions
1133 (40.6%)
Treatment Type for Gap(s)
Mental Health
968 (34.7%)
Substance-Related
112 (3.9%)
Either/Both
52 (1.9%)
Number of Gaps
1 Gap
700 (25.1%)
2-3 Gaps
262 (9.4%)
≥4 Gaps
71 (2.6%)
c
Reasons for Gaps
n = 1133
Detention/Prison
130 (11.5%)
Inpatient Stay
38 (3.6%)
No Clear Reason
965 (85.2%)
Note. N = Number; % = Percentage. * Participants who utilized outpatient treatment
within two years post-detention. a Percentages calculated out of total number of dropouts
(n = 721). b Participants with ≥2 outpatient sessions within 2 years post-detention. A total
of 419 participants obtained only one outpatient session, so not included in analyses for
gaps between outpatient sessions. c Percentages calculated out of total number of
participants with ≥1 gaps (n = 1133).

Age Cohorts
Mid-Age
Older
N (%)
N (%)
n = 1395 n = 871
286
(32.8%)c
29
(10.2%)b
1
(0.3%)a
256
(89.5%)b
N (%)
n = 749
255
(34.1%)b
20
(7.8%)c
9
(45.0%)c
226
(88.6%)b

323
(23.1%)b
44
(13.6%)b
2
(1.8%)b
277
(85.7%)b

Younger
N (%)
n = 935
112
(12.0%)a
37
(33.03)a
1
(0.1%)a
74
(66.1%)a

N (%)
N (%)
n = 855 n = 1244
402
476
(47.1%)a (38.7%)b
43
67
(10.7%)ac (14.1%)ab
13
16
(30.2%)a (23.9%)b
346
393
a
(57.1%)
(82.6%)b

χ2,
p-value
25.17,
p < .001
9.59,
p = .008
2.39,
p = .303
19.93,
p < .001

113.15,
p < .001
4.82,
p = .09
0.20,
p = .904
115.44,
p = .007

-χ2,
p-value

Note. N = Number; % = Percentage; χ2 = 2 x 2 chi-square test; p-value = significance level. + Participants who utilized outpatient treatment within Two Years of
detention release. ++ Participants with ≥2 outpatient sessions within 2 years post-detention.* Percentages calculated out of total number of dropouts (n = 721).
** Percentages calculated out of total number of participants with ≥1 gaps (n = 1133). For age cohorts, different superscripts indicate significant differences (p
≤ .01) and matching superscripts indicate non-significant differences (p > .01) based upon 2 x 2 chi-square tests between paired cohorts.

Gender
Race
Total
Sample Females Males
-White
Black
-χ2,
χ2,
N (%)
N (%)
N (%)
N (%)
N (%)
Treatment
N = 3207+ n = 918 n = 2283 p-value n = 1177 n = 1772 p-value
Variables
Dropouts from Outpatient Treatment
Outpatient
721
197
524
0.74
227
425
8.89,
Dropouts
(22.5%) (21.5%) (23.0%) p = .209 (19.3%) (23.9%) p = .003
Detention/
110
22
88
15.53,
41
62
0.16,
Prison*
(16.4%) (11.2%) (1.6%) p < .001
(1.8%)
(14.6%) p = .684
Inpatient
4
4
0
6.49,
1
3
0.28,
Stay*
(0.6%)
(2.0%)
(0.0%) p = .021
(0.4%)
(0.7%)
p = .512
No Clear
605
171
436
1.02,
185
360
8.91,
Reason*
(84.0%) (86.8%) (83.2%) p = .322 (81.5%) (84.7%) p = .003
Gaps between Outpatient Sessions
χ2,
χ2,
N (%)
N (%)
N (%)
N (%)
N (%)
++
N = 2789 n = 808 n = 2040 p-value n = 1078 n = 1556 p-value
Outpatient
1113
330
803
0.66,
425
615
0.13,
Gaps+
(40.6%) (40.8%) (39.4%) p = .417 (39.4%) (39.5%) p = .377
Detention/
130
34
96
0.40,
42
74
0.68,
Prison**
(11.5%) (10.3%) (2.0%) p = .30
(9.9%)
(12.0%) p = .234
Inpatient
38
19
19
2.28,
16
17
1.68,
Stay**
(33.6%)
(5.7%)
(2.4%) p = .091 (38.1%) (22.9%) p = .141
No Clear
965
277
688
0.13,
367
524
0.31,
Reason** (85.2%) (83.9%) (85.7%) p = .378 (52.0%) (65.0%) p = .303

Table 3.36 Outpatient Treatment Dropouts and Treatment Gaps within Two Years Post-Detention: Demographic Groups

204

204

Outpatient Treatment Dropouts and Treatment Gaps within Two Years Post-Detention: Mental Health Groups

Total
Sample
N (%)
N = 3207+

MAYSI-2
Disorder Type
Pre-Detention Treatment
No
Non-Positive Positive
Mental Substance
-Comorbid
-No
Yes
-MAYSI-2
Screen
Screen
Disorder Disorder
2
2
Treatment
N (%)
N (%)
N (%)
χ,
N (%)
N (%)
N (%)
χ,
N (%)
N (%)
χ 2,
Variables
n = 1934
n = 329
n = 944
p-value
n = 1606 n = 214
n = 1387 p-value
n = 1728 n = 1479 p-value
Dropouts from Outpatient Treatment
Outpatient
721
460
73
188
5.46,
402
102
217
120.83,
474
247
49.18,
Dropouts
(22.5%)
(23.8%)a
(22.2%)ab (19.9%)bc p = .065
(25.0%)a (47.7%)b (15.7%)c p < .001
(27.2%) (16.8%) p < .001
Detention/
110
67
10
33
16.73,
47
5
44
15.96,
75
35
4.53,
Prison*
(16.4%)
(14.6%)a
(13.7%)a (17.5%)b
p < .001
(11.7%)a (4.9%)b
(20.3%)c p < .001
(15.8%) (14.2%)
p = .04
Inpatient
4
4
0
0
1.22,
3
0
1
3.02,
3
1
0.01,
Stay*
(0.6%)
(0.8%)b
(0.0%)a
(0.0%)a
p = .545
(0.18%)a (0.0%)a
(0.5%)b
p = .388
(0.2%)
(0.1%)
p = .654
No Clear
605
389
63
155
3.21,
352
97
172
123.46,
396
211
40.77,
Reason*
(84.0%)
(84.6%)ab
(86.3%)a (82.5%)bc p = .201
(87.6%)ab (95.1%)b (79.3%)ac p < .001
(83.5%) (85.4%) p < .001
Gaps between Outpatient Sessions
N (%)
N (%)
N (%)
N (%)
χ2,
N (%)
N (%)
N (%)
χ2,
N (%)
N (%)
χ 2,
N = 2789++ n = 1685
n = 290
n = 873
p-value
n = 1387 n = 170 n = 1291
p-value
n = 1485 n = 1363 p-value
Outpatient
1113
662
119
352
0.11,
546
41
546
19.48,
497
636
46.51,
Gaps+
(40.6%)
(39.3%)a
(41.0%)a (40.3%)a p = .946
(39.4%)a (24.1%)b (42.3%)a p < .001
(33.5%) (46.7%)
p < .001
Detention/
130
59
21
50
13.36,
48
4
78
16.12,
53
77
9.95,
Prison**
(11.5%)
(8.9%)a
(17.7%)b (14.2%)c p < .001
(8.8%)a
(9.8%)a (14.3%)b p < .001
(10.5%) (12.1%) p = .002
Inpatient
38
24
2
12
6.26,
11
0
27
37.50,
20
18
12.37,
Stay**
(33.6%)
(3.7%)a
(1.7%)b
(3.4%)a
p = .044
(2.0%)a
(0.0%)b
(5.0%)c
p < .001
(4.2%)
(2.8%)
p < .001
No Clear
965
579
96
290
0.94,
487
37
441
9.26,
424
541
35.30,
Reason**
(85.2%)
(87.5%)a
(80.7%)b (82.4%)a p = .626
(89.2%)a (90.2%)a (80.7%)b
p = .010 (85.3%) (85.1%) p < .001
Note. N = Number; % = Percentage; χ2 = 2 x 2 chi-square test; p-value = Significance level. MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition.
+
Participants who utilized outpatient treatment within 2 years of detention release. ++ Participants with ≥2 outpatient sessions within 2 years post-detention.
* Percentages calculated out of total number of dropouts (n = 721). ** Percentages calculated out of total number of participants with ≥1 gaps (n = 1133). For MAYSI2 and disorder type, different superscripts indicate significant differences (p ≤ .01) and matching superscripts indicate non-significant differences (p > .01) based upon
2 x 2 chi-square tests between paired groups.

Table 3.37

205

205

721
(22.5%)
110
(16.4%)
4
(0.6%)
605
(84.0%)

N=
2789++
1113
(40.6%)
130
(11.5%)
38
(33.6%)
965
(85.2%)

Treatment
Variables

Outpatient
Dropouts
Detention/
Prison*
Inpatient
Stay*
No Clear
Reason*

Outpatient
Gaps+
Detention/
Prison**
Inpatient
Stay**
No Clear
Reason**

Violent Offender+
Recidivista
No
Yes
-No
Yes
χ2,
N (%)
N (%)
N (%)
N (%)
n = 1918 n = 1289 p-value n = 1537 n = 1670
Dropouts from Outpatient Treatment
458
263
5.26,
401
320
(23.9%)
(20.4%) p = .022 (26.1%) (19.2%)
76
34
1.50,
36
74
(16.6%)
(12.9%) p = .233
(8.9%) (23.1%)
0
3
2.56,
1
3
(0.0%)
(1.2%)
p = .141
(0.2%)
(0.9%)
382
226
5.26,
364
243
(83.4%)
(85.9%) p = .024 (90.8%) (75.9%)
Gaps between Outpatient Sessions
χ2,
N (%)
N (%)
N (%)
N (%)
n = 1474 n = 1125
p-value n = 1330 n = 1518
656
447
4.49,
520
613
(4.5%)
(39.7%) p = .019 (39.1%) (40.4%)
89
41
4.21,
27
103
(13.6%)
(9.2%)
p = .024
(5.2%) (16.8%)
22
16
0.52,
18
20
(3.4%)
(3.1%)
p = .288
(3.5%)
(3.3%)
545
420
9.88,
475
490
(83.1%)
(93.9%) p = .003 (91.4%) (79.9%)
308
(26.5%)
18
(5.8%)
2
(0.6%)
288
(93.5%)
N (%)
n = 1016
380
(32.7%)
0
(0.0%)
14
(3.7%)
366
(96.3%)

22.04,
p < .001
37.72,
p < .001
0.21,
p = .645
26.93,
p < .001

χ2,
p-value
0.61,
p = .228
40.00,
p < .001
3.03,
p = .059
3.51,
p = .033

-χ2,
p-value

N (%)
n = 1832
753
(36.8%)
130
(17.2%)
24
(3.2%)
599
(79.5%)

413
(20.2%)
92
(22.3%)
2
(0.5%)
319
(77.2%)

χ2,
p-value
3.43,
p = .035
77.16,
p < .001
1.02,
p = .199
3.58,
p = .032

16.61,
p < .001
36.85,
p < .001
0.04,
p = .609
20.25,
p < .001

Re-Detained/Incarceratedb
No
Yes
-χ2,
N (%)
N (%)
n = 1163 n = 2044 p-value

Note. N = Number; % = Percentage; χ2 = 2 x 2 chi-square test; p-value = Significance level. + Participants who utilized outpatient treatment within 2 years of
detention release. ++ Participants with ≥2 outpatient sessions within 2 years post-detention a Recidivism within 2 years post-detention. b Re-detained in detention
center or incarcerated in prison within 2 years post-detention. * Percentages calculated out of total number of dropouts (n = 721). ** Percentages calculated out
of total number of participants with ≥1 gaps (n = 1133).

Total
Sample
N (%)
N = 3207+

Table 3.38 Outpatient Treatment Dropouts and Treatment Gaps within Two Years Post-Detention: Criminal History Groups

206

206

207

Table 3.39 Outpatient Treatment Dropouts and Treatment Gaps within Two Years
Post-Detention: Cohort One versus Cohort Two

Treatment Variables
Dropout from Outpatient
Mental Health
Substance-Related
Either/Both
Timing of Dropout
Dropout after 1 Session
Dropout after 2 Sessions
Dropout after 3 Sessions

Total Sample
Number (%)
N = 3207+
721 (22.5%)
512 (16.0%)
160 (5.0%)
48 (1.5%)

Cohort One
Number (%)
n = 1900
429 (22.6%)
322 (16.9%)
80 (4.2%)
27 (1.4%)

Cohort Groups
Cohort Two
Number (%)
n = 1307
291 (22.3%)
190 (14.5%)
80 (6.1%)
21 (1.6%)

-χ2 ,
p-value
0.06, p = .830
0.01, p = .959
0.21, p = .694
9.55, p = .003

342 (10.7%)
219 (11.5%)
123 (9.4%)
3.63, p = .062
230 (7.2%)
132 (6.9%)
98 (7.5%)
0.35, p = .578
149 (4.6%)
79 (4.2%)
70 (5.4%)
2.51, p = .124
Number (%)
Number (%)
Number (%)
χ2 ,
Reasons for Dropouts
n = 721
n = 429
n = 291
p-value
Detention/Prison*
60 (8.3%)
34 (7.54%)
26 (8.9%)
5.18, p = .075
Inpatient Stay*
55 (7.7%)
43 (10.0%)
12 (4.1%)
0.56, p = .540
No Clear Reason*
605 (84.0%)
352 (82.1%)
253 (86.9%)
2.29, p = .134
Number (%)
Number (%)
Number (%)
χ2 ,
++
n = 2789
n = 1617
n = 1172
p-value
Gaps between Sessions (y/n) 1133 (40.6%)
655 (40.5%)
478 (42.2%)
1.73, p = .629
Mental Health
968 (34.7%)
589 (36.4%)
379 (32.3%)
8.48, p < .001
Substance-Related
112 (4.0%)
51 (3.1%)
61 (5.2%)
3.49, p = .022
Either/Both
52 (1.9%)
15 (0.01%)
38 (3.24%)
12.25, p = .003
Number of Gaps
1 Gap
700 (21.8%)
389 (20.5%)
311 (23.8%)
129.51, p < .001
2-3 Gaps
262 (11.2%)
218 (11.5%)
144 (12.3%)
35.62, p = .393
≥4 Gaps
71 (2.2%)
48 (2.5%)
23 (1.6%)
1.39, p = .278
χ2 ,
Number (%)
Number (%)
Number (%)
n = 1133
n = 655
n = 478
p-value
Reasons for Gaps
Detention/Prison**
130 (11.5%)
77 (11.7%)
53 (11.1%)
0.01, p = .538
Inpatient Stay**
38 (3.4%)
23 (3.5%)
15 (3.1%)
2.44, p = .094
No Clear Reason**
965 (85.2%)
555 (84.7%)
410 (85.8%)
0.13, p = .374
Note. N = Number; % = Percentage; χ2 = 2 x 2 chi-square test; p-value = Significance level.
+
Participants who utilized outpatient treatment within 2 years of detention release.
++
Participants with ≥2 outpatient sessions within two years post-detention * Percentages
calculated out of total number of dropouts (n = 721). ** Percentages calculated out of total
number of participants with ≥1 gaps (n = 1133).

Number of
Outpatient
Sessions per
Month*

Total Number of
Treatment
Services*

Main Outcomes
Prevalence Rates
for Post-Detention
Treatment
Utilization*

Demographic Groups
Mental Health Groups
Age
MAYSI-2
Disorder
Prior
Variables
Gender Race Cohorts Positive Screen
Type
Treatment
Y > O;
PS > No;
C > MD >
User >
Any Treatment
M>F W>B
MA > O Non-PS > No
SD
Non-User
Y > MA PS > Non-PS MD > C >
User >
Mental Health
M>F W>B
>O
> No
SD
Non-User
PS > Non-PS SD > C >
User >
Substance-Related
M>F W>B
> No
MD
Non-User
MA > Y; PS > Non-PS C > MD;
User >
Either/Both
M>F W>B
MA > O
> No
C > SD
Non-User
Y > O; PS > Non-PS C > MD >
User >
Outpatient
M>F W>B
MA > O
> No
SD
Non-User
PS > Non-PS; C > MD;
User >
Non-Outpatient
W>B
PS > No
C > SD
Non-User
Y > MA PS > Non-PS C > MD >
User >
Any Treatment
>O
> No
SD
Non-User
Y > MA
PS > No;
MD > SD;
User >
Mental Health
>O
PS > Non-PS
C > SD
Non-User
C > MD;
Substance-Related
M>F W>B
SD > MD
C > MD;
User >
Either/Both
M>F
SD > MD Non-User
Y > MA PS > Non-PS C > MD >
User >
Outpatient
>O
> No
SD
Non-User
C > SD;
Non-Outpatient
MD > SD
Y > MA; Non-PS > No; MD > SD;
User >
Any Outpatient
Y>O
PS > No
C > SD
Non-User
PS > Non-PS MD > SD;
Mental Health Outpatient
> No
C > SD
Viol >
Non-Viol

Rec >
Non-Rec

Re-Det/Inc
> Not

2>1

2>1

2>1

Criminal History Groups
Violent
Re-Detained/ Cohort
Offender Recidivist Incarcerated Groups
Viol >
Rec >
Re-Det/Inc
2>1
Non-Viol Non-Rec
> Not
Viol >
Rec >
Re-Det/Inc
2>1
Non-Viol Non-Rec
> Not
Viol >
Rec >
Re-Det/Inc
2>1
Non-Viol Non-Rec
> Not
Rec >
Re-Det/Inc
2>1
Non-Rec
> Not
Viol >
Rec >
Re-Det/Inc
2>1
Non-Viol Non-Rec
> Not
Rec >
Re-Det/Inc
2>1
Non-Rec
> Not
Rec >
Re-Det/Inc
Non-Rec
> Not
Rec >
Re-Det/Inc
Non-Rec
> Not

Table 3.40 Summary of Results: Group Differences in Prevalence and Quantity of Treatment Services

208

208

Either/Both Inpatient

Substance-Related Inpatient

Mental Health Inpatient

Any Inpatient

Non-Outpatient

Either/Both Outpatient

Substance-Related Outpatient

Mental Health Outpatient

Any Outpatient

PS > Non-PS; C > MD >
Non-PS > No
SD

User >
NonUser

Rec >
Non-Rec

Re-Det/Inc
> Not

2>1

Note. MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition; RAI = Risk Assessment Inventory. M = Male; F = Female; W = White; B = Black; Y =
Younger; MA = Mid-Age; O = Older; PS = Positive screen; Non-PS = Non-positive screen; No = No MAYSI-2 Score; MD = Mental disorder; SD = Substancerelated disorder; C = Comorbid; User = Pre-detention treatment user; Non-User = Not a pre-detention treatment user; Viol = Violent offender; Non-Viol = Nonviolent offender; Rec = Recidivist; Non-Rec = not a recidivist; Re-Det/Inc = Re-detained or incarcerated within two years post-detention; Not = not re-detained or
incarcerated within two years post-detention; 2 = Cohort Two; 1 = Cohort One. * Treatment utilization within two years of detention release.

Length of
Inpatient Stay*

Time between
Treatment
Services*

Either/Both Outpatient

Substance-Related
Outpatient

Table 3.40 Continued.

209

209

Outpatient
User Intensity
Levels*

Treatment
Setting User*

Main
Outcomes
Treatment
Type User*

13-17 Sessions

8-12 Sessions

3-7 Sessions

1-2 Sessions

Three Settings

Only NonOutpatient
Outpatient and
Non-Outpatient

Only Outpatient

All Three Types

Mental Health and
Substance
Mental Health and
Either/Both
Substance and
Either/Both

Only Either/Both

Variables
Only Mental
Health
Only SubstanceRelated

Demographic Groups
Mental Health Groups
Age
MAYSI-2
Disorder
Prior
Gender Race
Cohorts Positive Screen
Type
Treatment
Y > MA Non-PS > No; MD > C >
User >
M>F
>O
PS > No
SD
Non-User
O > MA Non-PS > No; DS > C >
User >
M>F
>Y
PS > No
MD
Non-User
PS > No;
User >
M>F
PS > Non-PS
Non-User
Non-PS > No;
User >
M>F W>B
PS > No
Non-User
Non-PS > No; C > MD;
User >
M>F
PS > No
C > SD
Non-User
User >
M>F
PS > No
Non-User
MA > Y; Non-PS > No;
User >
M>F W>B
MA > O
PS > No
Non-User
Y > O;
PS >
MD > C >
User >
M>F
MA > O Non-PS > No
SD
Non-User
User >
W>B
SD > MD
Non-User
Y > O;
PS > No;
C > MD >
User >
W>B
MA > O PS > Non-PS
SD
Non-User
PS > No;
C > MD; C
User >
W>B
PS > Non-PS
> SD
Non-User
MA > Y; Non-PS > No; SD > MD;
User >
M>F W>B
O>Y
PS > No
SD > C
Non-User
Non-PS > No;
User >
M>F
MA > Y
PS > No
Non-User
PS > No;
MD > C >
User >
M>F
PS > Non-PS
SD
Non-User
MD > SD;
User >
M>F
MA > O
PS > No
C > SD
Non-User

Viol >
Rec >
Re-Det/Inc
Non-Viol Non-Rec
> Not
Viol >
Rec >
Re-Det/Inc
Non-Viol Non-Rec
> Not
Rec >
Non-Rec
Rec >
Re-Det/Inc
Non-Rec
> Not
Rec >
Re-Det/Inc
Non-Rec
> Not
Rec >
Non-Rec
Rec >
Re-Det/Inc
Non-Rec
> Not
Rec >
Re-Det/Inc
Non-Rec
> Not
Rec >
Re-Det/Inc
Non-Rec
> Not

Rec >
Re-Det/Inc
Non-Rec
> Not
Rec >
Non-Rec

Criminal History Groups
Violent
Re-Detained/
Offender Recidivist Incarcerated
Viol >
Rec >
Re-Det/Inc
Non-Viol Non-Rec
> Not
Viol >
Rec >
Non-Viol Non-Rec

Table 3.41 Summary of Results: Group Differences in Treatment Intensity, Dropouts, and Gaps

2>1

2>1

2>1

2>1

2>1

2>1

2>1

2>1

2>1

2>1

2>1

2>1

2>1

2>1

Cohort
Groups

210

210

M>F

23+ Sessions

Inpatient

Detention/Prison

Dropout

7+ Stays/Visits

W>B

W>B

W>B

5-6 Stay/Visit

MA > Y;
O>Y
Y > MA;
Y>O

O > MA
>Y

PS > No;
PS > Non-PS

Y>O
PS > No
MA > O
Y > MA
PS >
>O
Non-PS > No
Y > O;
PS > No;
MA > O PS > Non-PS
PS > No;
PS > Non-PS
PS > No;
PS > Non-PS
PS > No;
PS > Non-PS
C > MD;
C > SD
SD >MD
SD > MD
>C
C > MD >
SD

C > MD

C > MD >
SD
C > MD;
C > SD
C > MD;
C > SD

C > SD

User >
Non-User

Rec >
Re-Det/Inc
Non-Rec
> Not
Rec >
Re-Det/Inc
Non-Rec
> Not

User >
Rec >
Re-Det/Inc
Non-User
Non-Rec
> Not
User >
Viol >
Rec >
Re-Det/Inc
Non-User Non-Viol Non-Rec
> Not
User >
Rec >
Re-Det/Inc
Non-User
Non-Rec
> Not
User >
Rec >
Re-Det/Inc
Non-User
Non-Rec
> Not
User >
Rec >
Non-User
Non-Rec
User >
Rec >
Non-User
Non-Rec

2>1

2>1

2>1

2>1

2>1

No Clear Reason

SD > MD
User >
Rec >
Re-Det/Inc
>C
Non-User
Non-Rec
> Not
Outpatient
Non-PS >
MD > SD;
User >
Outpatient Gap
Gaps*
PS > No
C > SD
Non-User
C > MD;
User >
Rec >
Re-Det/Inc
Detention/Prison
MA > O
C > SD
Non-User
Non-Rec
> Not
C > MD
User >
Inpatient
> SD
Non-User
MA > Y;
MD > C;
User >
Viol >
No Clear Reason
O>Y
SD > C
Non-User Non-Viol
Note. MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition; RAI = Risk Assessment Inventory. M = Male; F = Female; W = White; B = Black; Y =
Younger; MA = Mid-Age; O = Older; PS = Positive screen; Non-PS = Non-positive screen; No = No MAYSI-2 Score; MD = Mental disorder; SD = Substancerelated disorder; C = Comorbid; User = Pre-detention treatment user; Non-User = Not a pre-detention treatment user; Viol = Violent offender; Non-Viol = Nonviolent offender; Rec = Recidivist; Non-Rec = not a recidivist; Re-Det/Inc = Re-detained or incarcerated within two years post-detention; Not = not re-detained or
incarcerated within two years post-detention; 2 = Cohort Two; 1 = Cohort One. * Treatment utilization within two years of detention release.

Outpatient
Dropouts*

W>B

W>B

W>B

3-4 Stays/Visits

M>F

M>F

18-22 Sessions

Non-Outpatient
1 Stay/Visit
User Intensity
Levels*
2 Stays/Visits

Outpatient
User Intensity
Levels*

Table 3.41 Continued.

211

211

212
Table 3.42 Multicollinearity Statistics of Predictor Variables
Multicollinearity Statistics
Predictor Variables

Tolerance

Variance Inflation Factor

Male (y/n)

0.93

1.08

Black race (y/n)

0.88

1.14

Age at Detention Entry

0.84

1.20

Number of Prior Arrests

0.88

1.14

Number of Charges

0.98

1.02

Charge Severity (1-5)

0.45

2.90

Length of Stay in Detention

0.92

1.09

Violent Offender (y/n)

0.94

1.06

Risk Assessment Inventory

0.88

1.13

MAYSI-2 Positive Screen (y/n)

0.96

1.04

Conduct-Related Disorder (y/n)

0.35

2.88

Non-Conduct Mental Disorder (y/n)

0.48

2.11

Substance-Related Disorder (y/n)

0.88

1.14

Number of Disorders

0.30

3.36

Pre-Detention Outpatient Treatment (y/n)

0.28

3.55

Pre-Detention Non-Outpatient Treatment (y/n)

0.72

1.39

MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition

213
Table 3.43 Model Summary Statistics for Logistic Regression Models Predicting
Treatment Utilization within Two Years Post-Detention
Model Summary Statistics (N = 8058)
-2 Log
Nagelkerke
Classification
Treatment
χ2
df
P
2
R
Percentage
Outcomes Likelihood
Post-Detention Treatment (Y/N)
Stage 1
10134.05
.044
258.20
3
.000
63.7%
Stage 2
9861.06
.089
530.63
8
.000
65.6%
Stage 3
7007.15
.476
3395.05
14
.000
80.0%
Post-Detention Mental Health Treatment (Y/N)
Stage 1
9552.82
.048
273.67
3
.000
69.0%
Stage 2
9356.00
.081
470.58
8
.000
69.4%
Stage 3
5793.67
.561
4032.81
14
.000
82.2%
Post-Detention Substance-Related Treatment (Y/N)
Stage 1
5089.61
.040
156.34
3
.000
89.8%
Stage 2
4980.90
.068
265.05
8
.000
89.8%
Stage 3
3041.58
.501
2204.37
14
.000
89.9%
Post-Detention Either/Both Treatment (Y/N)
Stage 1
3301.02
.024
66.08
3
.000
94.5%
Stage 2
3199.19
.061
167.92
8
.000
94.5%
Stage 3
2015.26
.453
1351.86
14
.000
94.6%
Post-Detention Outpatient Treatment (Y/N)
Stage 1
10070.35
.043
255.27
3
.000
64.4%
Stage 2
9804.65
.087
520.88
8
.000
66.0%
Stage 3
7013.70
.469
3311.89
14
.000
80.0%
Post-Detention Non-Outpatient Treatment (Y/N)
Stage 1
4187.08
.030
101.15
3
.000
92.4%
Stage 2
4117.57
.051
170.63
8
.000
92.4%
Stage 3
3281.36
.285
1006.83
14
.000
92.5%
Note. N = Number; df = Degrees of freedom; χ2 = Chi-square test (comparing model with
predictors to null model with no predictors); p = Significance level. Stage 1 includes the
following demographic predictors: gender, race, and age. Stage 2 includes the following
criminal history variables: number of prior arrests, number of charges at detention entry,
charge severity, length of stay, and violent offender (y/n). Stage 3 includes the following
mental health predictor variables: conduct disorder (y/n), non-conduct mental disorder
(y/n), substance-related disorder (y/n), number of disorders, pre-detention outpatient
treatment (y/n), and pre-detention non-outpatient treatment (y/n).

Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
.532
.066
1.70
1.50 – 1.94 .000
.365
Black (y/n)
-.041
.063
0.96
0.85 – 1.09 .520 -.065
Age
-.006
.019
0.98
0.96 – 0.99 .010 -.068
Stage 2: Criminal History Variables
Number of Prior Arrests
.086
.013
1.09
1.06 – 1.12 .000
.047
Charge Number
-.012
.059
0.99
0.88 – 1.18 .835
.038
Charge Severity
.089
.032
1.09
1.03 – 1.16 .005
.079
Length of Stay
.001
.002
1.00
0.99 – 1.01 .886 -.002
Violent Offender (y/n)
.165
.064
1.18
1.04 – 1.34 .010
.123
Stage 3: Mental Health Variables
Conduct Disorder (y/n)
3.303
.090
7.18 2.41 – 22.97 .000
.072
Non-Conduct Mental Disorder (y/n) 1.585
.093
4.88
4.07 – 5.82 .000
.012
Substance-Related Disorder (y/n)
.792
.069
2.21
1.93 – 2.53 .000 -.468
Number of Disorders
-.011
.009
0.99
0.97 – 1.00 .232
.045
Pre-Detention Outpatient (y/n)
.015
.002
1.02
1.00 – 1.21 .000
.011
Pre-Detention Non-Outpatient
.359
.090
1.43
1.20 – 1.71 .000
.182
Treatment (y/n)
Note. N = Number; SE = Standard error; CI = Confidence interval; p = Significance level.

1.44
0.91
0.94
1.05
1.04
1.08
1.00
1.13
2.08
1.01
0.63
1.05
1.11
1.20

.070
.067
.020
.014
.064
.034
.002
.068
.004
.001
.074
.009
.002
.073

.000
.003
.001
.001
.555
.020
.474
.069
.000
.000
.000
.000
.000
.013

1.02 – 1.08
0.92 – 1.28
1.01 – 1.16
0.99 – 1.01
0.99 – 1.29
1.67 – 12.08
1.00 – 2.02
0.54 – 0.72
1.03 – 1.06
1.00 – 1.02
1.04 – 1.39

p
1.26 – 1.65
0.82 – 1.00
0.90 – 0.97

Treatment Outcomes (N = 8058)
Any Treatment Utilization (y/n)
Mental Health Treatment (y/n)
LogOdds
LogOdds
SE
95% CI
p
SE
95% CI
Odds
Ratios
Odds
Ratios

Table 3.44 Odds Ratios and Associated Statistics for Logistic Regression Models Predicting Treatment and Mental Health
Treatment within Two Years Post-Detention

214

214

Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
.727
.104
2.07 1.69 – 2.54 .000
.654
Black (y/n)
-.035
.093
0.97 0.81 – 1.16 .708 -.050
Age
.197
.029
1.33 1.15 – 1.59 .000
.141
Stage 2: Criminal History Variables
Number of Prior Arrests
.080
.017
1.08 1.05 – 1.12 .000
.093
Charge Number
.002
.090
1.00 0.84 – 1.20 .979
.030
Charge Severity
.014
.046
1.01 0.93 - 1.11 .766
.001
Length of Stay
-.004
.003
0.99 0.98 – 1.00 .208 -.007
Violent Offender (y/n)
-.207
.099
0.81 0.67 – 0.99 .037 -.143
Stage 3: Mental Health Variables
Conduct Disorder (y/n)
.343
.133
1.41 1.09 – 1.83 .010 1.285
Non-Conduct Mental Disorder (y/n) -.738
.135
1.99 1.48 – 2.62 .000 1.961
Substance-Related Disorder (y/n)
2.503
.888
2.02 1.74 – 2.92 .000
.824
Number of Disorders
-.044
.012
0.96 0.94 – 0.98 .000
.025
Pre-Detention Outpatient
.003
.002
1.00 0.99 – 1.01 .136
.001
Treatment (y/n)
Pre-Detention Non-Outpatient
-.023
.072
0.98 0.85 – 1.13 .750
.025
Treatment (y/n)
Note. N = Number; SE = Standard error; CI = Confidence interval; p = Significance level.

1.92
0.95
1.15
1.10
1.03
1.00
0.99
0.87
1.02
1.00
4.46
1.03
1.00
1.03

.135
.117
.037
.019
.114
.060
.004
.124
.447
.091
.454
.012
.002
.072

.000
.669
.000
.000
.794
.994
.136
.250
.000
.728
.000
.028
.623
.727

1.48 – 2.51
0.76 – 1.20
1.08 – 1.24
1.06 – 1.14
0.82 – 1.29
0.89 – 1.12
0.98 – 1.00
0.68 – 1.11
1.01 – 1.92
0.99 – 1.01
1.14 – 12.88
1.00 – 1.05
0.99 – 1.01
0.89 – 1.18

Substance-Related Treatment Utilization (y/n) Either/Both Treatment Utilization (y/n)
LogOdds
LogOdds
SE
95% CI
p
SE
95% CI
p
Odds
Ratios
Odds
Ratios

Treatment Outcomes (N = 8058)

Table 3.45 Odds Ratios and Associated Statistics for Logistic Regression Models Predicting Substance-Related Treatment and
Either/Both Treatment within Two Years Post-Detention

215

215

Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
.521 .066 1.68
1.48 – 1.92 .000
Black (y/n)
.027 .063 1.03
0.91 – 1.16 .675
Age
-.014 .019 0.99
0.95 – 1.03 .461
Stage 2: Criminal History Variables
Number of Prior Arrests
.082 .013 1.09
1.05 – 1.12 .000
Charge Number
-.012 .059 0.98
0.88 – 1.11 .842
Charge Severity
.085 .032 1.09
1.02 – 1.16 .007
Length of Stay
.001 .002 1.00
0.99 – 1.02 .655
Violent Offender (y/n)
.186 .064 1.21
1.06 – 1.37 .003
Stage 3: Mental Health Variables
Conduct Disorder (y/n)
3.298 .099 7.05 2.26 – 13.85 .000
Non-Conduct Mental Disorder (y/n)
1.605 .094 4.98
4.14 – 6.00 .000
Substance-Related Disorder (y/n)
.684 .069 1.98
1.73 – 2.28 .000
Number of Disorders
-.002 .009 0.99
0.98 – 1.02 .801
Pre-Detention Outpatient Treatment (y/n)
.015 .002 1.05
1.01 – 1.22 .000
Pre-Detention Non-Outpatient
.230 .081 1.26
1.08 – 1.48 .004
Treatment (y/n)
Note. N = Number; SE = Standard error; CI = Confidence interval; p = Significance level.

Outpatient Treatment Utilization (y/n)
LogOdds
SE
95% CI
p
Odds
Ratios
.105
.097
.030
.017
.093
.049
.003
.101
.202
.210
.107
.010
.002
.062

.006
-.470
.005
.062
.009
.009
-.005
-.035
1.630
.671
.720
.128
-.002
.212

1.24

5.105
1.96
2.054
1.14
1.00

1.06
1.09
1.01
1.00
0.96

1.01
0.63
1.01

.954
.000
.866
.000
.922
.854
.177
.727
.000
.001
.000
.000
.337
.001

0.92 – 1.23
0.52 – 0.76
0.95 – 1.07
1.03 – 1.14
0.84 – 1.21
0.92 – 1.11
0.98 – 1.01
0.79 – 1.28
3.44 – 7.59
1.30 – 2.95
1.67 – 2.53
1.11 – 1.20
0.99 – 1.01
1.09 – 1.40

Non-Outpatient Treatment Utilization (y/n)
LogOdds
SE
95% CI
p
Odds
Ratios

Treatment Outcomes (N = 8058)

Table 3.46 Odds Ratios and Associated Statistics for Logistic Regression Models Predicting Outpatient Treatment and NonOutpatient Treatment within Two Years Post-Detention

216

216

217
Table 3.47 Model Summary Statistics for Logistic Regression Predicting Treatment
within Two Years Post-Detention: Cohort One
Model Summary Statistics (N = 5924)
-2 Log
Nagelkerke
Classification
χ2
df
P
2
Likelihood
R
Percentage
Post-Detention Treatment (Y/N)
Stage 1
6992.27
.053
227.53
3
.000
70.3%
Stage 2
6901.19
.074
318.61
8
.000
75.5%
Stage 3
4780.66
.479
2439.13
14
.000
82.1%
Post-Detention Mental Health Treatment (Y/N)
Stage 1
6569.85
.061
254.55
3
.000
73.8%
Stage 2
6496.84
.079
327.59
8
.000
73.9%
Stage 3
3894.46
.570
929.97
14
.000
84.4%
Post-Detention Substance-Related Treatment (Y/N)
Stage 1
3089.69
.030
73.68
3
.000
90.5%
Stage 2
3020.90
.057
142.48
8
.000
92.5%
Stage 3
1870.52
.474
1292.80
14
.000
92.8%
Post-Detention Either/Both Treatment (Y/N)
Stage 1
1677.83
.017
25.09
3
.000
94.5%
Stage 2
1656.05
.032
46.86
8
.000
96.8%
Stage 3
1073.85
.403
629.09
14
.000
96.8%
Post-Detention Outpatient Treatment (Y/N)
Stage 1
6935.32
.052
228.01
3
.000
70.8%
Stage 2
6844.97
.074
318.36
8
.000
71.1%
Stage 3
4769.56
.473
2393.78
14
.000
82.2%
Post-Detention Non-Outpatient Treatment (Y/N)
Stage 1
2743.58
.025
55.17
3
.000
93.7%
Stage 2
2728.85
.031
69.90
8
.000
93.9%
Stage 3
2167.92
.268
639.81
14
.000
93.9%
Note. N = Number; df = Degrees of freedom; χ2 = Chi-square test (comparing model with
predictors to null model with no predictors); p = Significance level. Stage 1 includes the
following demographic predictors: gender, race, and age. Stage 2 includes the following
criminal history variables: number of prior arrests, number of charges at detention entry,
charge severity, length of stay, and violent offender (y/n). Stage 3 includes the following
mental health predictor variables: conduct disorder (y/n), non-conduct mental disorder
(y/n), substance-related disorder (y/n), number of disorders, pre-detention outpatient
treatment (y/n), and pre-detention non-outpatient treatment (y/n).

NA
+
+
+
–
+
+

NA
+
+
+

+

+

+
NA

+

+

+
NA

+

+

–

+

–

+
+
+
+

–

–

–
+

–

+
–
+

+

+

–
–
NA NA

+

+

+

–

+

+

+

+

–

+

–

NA NA
NA NA
+
+ + +
+ –
+
+ + +
– –
–
– – +
+ +
+
+ --+
+

+
NA NA

+

+

+
–
–
+

+

+

+

+

+
+
+

+

–

+
+
–

NA NA
+
+ +

NA NA

+

+

+

+

+

–

+

+

–

–

–

+

–

NA NA

+

–

–

+

+

–

+

+

–

NA NA
NA NA +
+
+ -- +
+ +
+ +
+
+
+ + +
+ +
– + +
+ +
+
+

+
+
NA NA

+

+

+

Note. NA = Not applicable because predictor variable excluded from analysis. Total = Total sample; 1 = Cohort one; 2 = Cohort two; (+) = Increased likelihood;
(–) = Decreased likelihood; RAI = Risk Assessment Inventory; MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition. a N = 8038. b N = 5924.
c
N = 2114.

Predictor Variables
Demographic Variables
Male (y/n)
Black (y/n)
Age
Criminal History Variables
Number of Prior Arrests
Charge Number
Charge Severity
Length of Stay
Violent Offender (y/n)
RAI Scores
Mental Health Variables
MAYSI-2 – Positive Screen (y/n)
Conduct Disorder (y/n)
Non-Conduct Mental Disorder (y/n)
Substance-Related Disorder (y/n)
Number of Disorders
Pre-Detention Outpatient (y/n)
Pre-Detention Non-Outpatient
Treatment (y/n)

Treatment Outcomes
Mental
SubstanceNonAny Treatment
Either/Both
Outpatient
Health
Related
Outpatient
Totala 1b 2c Total 1 2 Total 1 2 Total 1 2 Total 1 2 Total 1 2

Table 3.48 Predictors of Logistic Regression Associated with Increased Likelihood (+) and Decreased Likelihood (–) of PostDetention Treatment Utilization: Total Sample, Cohort One, and Cohort Two

218

218

LogPredictor Variables
Odds
Stage 1: Demographic Variables
Male (y/n)
.542 .078 1.72
1.48 – 2.00 .000
.406
Black (y/n)
-.041 .076 0.96
0.83 – 1.11 .591 -.042
Age
-.033 .023 1.03
0.98 – 1.08 .143 -.016
Stage 2: Criminal History Variables
Number of Prior Arrests
.123 .020 1.13
1.09 – 1.18 .000
.090
Charge Number
.010 .075 1.01
0.87 – 1.17 .891
.017
Charge Severity
.089 .040 1.09
1.01 – 1.08 .024
.092
Length of Stay
-.003 .003 1.00
0.99 – 1.12 .228 -.003
Violent Offender (y/n)
.186 .078 1.21
1.04 – 1.41 .016
.200
Stage 3: Mental Health Variables
Conduct Disorder (y/n)
3.68 .124 9.74
1.14 –20.73 .000
1.834
Non-Conduct Mental Disorder (y/n) 1.71 .115 5.54
4.42 – 6.95 .000
1.929
Substance-Related Disorder (y/n)
.570 .084 1.77
1.50 – 2.09 .000 -.420
Number of Disorders
-.031 .010
0.97
0.95 – 0.99 .002
.008
Pre-Detention Outpatient Treatment
.088 .012 1.09
1.06 – 1.12 .001
.068
(y/n)
Pre-Detention Non-Outpatient
.727 .209 2.07
1.37 – 3.12 .000
.184
Treatment (y/n)
Note. N = Number; SE = Standard error; CI = Confidence interval; p = Significance level.

.000
.831
.032
.242
.004
.000
.001
.000
.422
.000
.000

1.05 – 1.14
0.87 – 1.19
1.00 – 1.19
0.99 – 1.01
1.04 – 1.44
12.10 3.08 – 17.67
8.58 2.00 – 12.01
0.65
0.55 – 0.79
1.01 0.98 – 1.03
1.02 – 1.19
1.07 – 2.19

1.07
1.53

.146
.044
.091
.010
.191
1.53

1.09
1.01
1.09
1.00
1.22

.022
.082
.042
.003
.083

.000
.607
.000

1.50
0.96
0.90

.084
.082
.025

1.27 – 1.78
0.82 – 1.13
0.86 – .94

Treatment Outcomes (N = 5924)
Treatment Utilization (y/n)
Mental Health Treatment Utilization (y/n)
Odds
LogOdds
SE
95% CI
p
SE
95% CI
p
Ratios
Odds
Ratios

Table 3.49 Odds Ratios and Associated Statistics for Logistic Regression Models Predicting Treatment and Mental Health
Treatment within Two Years Post-Detention: Cohort One

219

219

Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
.654 .129 1.92
1.49 – 2.48 .000
Black (y/n)
-.068 .120 0.93
0.78 – 1.18 .934
Age
.206 .037 1.23
1.14 – 1.32 .000
Stage 2: Criminal History Variables
Number of Prior Arrests
.155 .030 1.17
1.10 – 1.24 .000
Charge Number
.004 .199 1.01
0.80 – 1.27 .975
Charge Severity
-.006 .060 0.96
0.88 – 1.12 .920
Length of Stay
-.009 .005 0.99
0.98 – 1.01
.066
Violent Offender (y/n)
-.347 .129 0.71
0.55 – 0.91 .007
Stage 3: Mental Health Variables
Conduct Disorder (y/n)
.691 .170 2.00
1.42 – 2.79 .000
Non-Conduct Mental Disorder (y/n)
-.820 .171 0.44
0.32 – 0.62 .000
Substance-Related Disorder (y/n)
2.130 .906 4.67
1.59 – 7.98 .000
Number of Disorders
-.044 .015 0.96
0.93 – 0.99 .003
Pre-Detention Outpatient Treatment
-.003 .005 1.00
0.98 – 1.01 .622
(y/n)
Pre-Detention Non-Outpatient
.593 .164 1.81
1.32 – 2.50 .000
Treatment (y/n)
Note. N = Number; SE = Standard error; CI = Confidence interval; p = Significance level.
.183
.165
.051
.041
.165
.087
.007
.173
.357
.829
.716
.016
.005
.180

.624
-.087
.167
.069
.108
-.120
-.012
-.189
1.961
1.536
4.93
.008
.011
.150

1.49

1.01

3.02
1.00
3.43
1.01

1.07
1.11
0.92
0.99
0.83

1.87
0.92
1.18

.000
.600
.001
.094
.515
.167
.087
.275
.002
.940
.000
.622
.026
.010

1.30 – 2.67
0.66 – 1.26
1.07 – 1.31
0.98 – 1.16
0.81 – 1.54
0.75 – 1.05
0.97 – 1.00
0.59 – 1.16
1.01 – 5.23
0.98 – 1.04
1.43 – 13.03
0.97 – 1.04
1.00 – 1.02
1.08 – 2.06

Treatment Outcomes (N = 5924)
Substance-Related Treatment Utilization (y/n)
Either/Both Treatment Utilization (y/n)
LogOdds
LogOdds
SE
95% CI
p
SE
95% CI
p
Odds
Ratios
Odds
Ratios

Table 3.50 Odds Ratios and Associated Statistics for Logistic Regression Models Predicting Substance-Related Treatment and
Either/Both Treatment within Two Years Post-Detention: Cohort One

220

220

Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
.531 .078 1.73
1.56 – 1.98 .000
.039
Black (y/n)
.020 .076 1.02
0.88 – 1.18
.792
-.384
Age
.023 .023 0.94
0.91 – 0.98
.005
.038
Stage 2: Criminal History Variables
Number of Prior Arrests
.121 .020 1.13
1.08 – 1.18 .000
.068
Charge Number
-.001 .075 1.00
0.86 – 1.16
.991
.050
Charge Severity
.082 .040 1.09
1.01 – 1.17 .037
.073
Length of Stay
-.002 .003 1.00
0.98 – 1.00
.540
-.009
Violent Offender (y/n)
.186 .077 1.20
1.03 – 1.40 .016
-.072
Stage 3: Mental Health Variables
Conduct Disorder (y/n)
3.730 .126 14.57 3.69 –25.93 .000
1.92
Non-Conduct Mental Disorder (y/n)
1.758 .118 5.80
4.60 – 7.31 .000
.905
Substance-Related Disorder (y/n)
.480 .084 1.62
1.37 – 4.91 .000
.528
Number of Disorders
-.025 .010
0.98 0.95 – 0.99
.013
.118
Pre-Detention Outpatient Treatment
.071 .010 1.07
1.01 – 1.10 .000
.001
(y/n)
Pre-Detention Non-Outpatient
.656 .197 1.93
1.31 – 2.84 .001
.708
Treatment (y/n)
Note. N = Number; SE = Standard error; CI = Confidence interval; p = Significance level.

1.04
0.68
1.04
1.09
1.05
1.08
0.99
0.93
6.80
2.47
1.69
1.13
1.00
2.03

.128
.121
.037
.030
.117
.063
.005
.126
.251
.255
.134
.012
.004
.145

.005
.666
.246
.046
.567
.000
.000
.000
.000

1.03 – 1.16
0.84 – 1.32
0.95 – 1.21
0.98 – 1.00
0.73 – 1.19
4.16 – 11.12
1.50 – 4.07
1.31 – 2.21
1.09 – 1.15

1.52 – 2.70

.000

0.98 – 12.01 .915

.762
.002
.306

0.81 – 1.33
0.54 – 0.86
0.95 – 1.12

Treatment Outcomes (N = 5924)
Outpatient Treatment Utilization (y/n)
Non-Outpatient Treatment Utilization (y/n)
LogOdds
LogOdds
SE
95% CI
p
SE
95% CI
p
Odds
Ratios
Odds
Ratios

Table 3.51 Odds Ratios and Associated Statistics for Logistic Regression Models Predicting Outpatient Treatment and NonOutpatient Treatment within Two Years Post-Detention: Cohort One

221

221

222
Table 3.52 Model Summary Statistics for Logistic Regression Models Predicting
Treatment within Two Years Post-Detention: Cohort Two
Model Summary Statistics (N = 2114)
-2 Log
Nagelkerke
ChiClassification
df
p
2
Likelihood
R
square
Percentage
Post-Detention Treatment (Y/N)
Stage 1
2076.33
.096
119.77
3
.000
63.8%
Stage 2
2016.39
.141
179.70
8
.000
66.2%
Stage 3
1464.93
.360
731.17
14
.000
80.0%
Post-Detention Mental Health Treatment (Y/N)
Stage 1
2073.21
.150
194.80
3
.000
65.3%
Stage 2
2017.28
.189
250.72
9
.000
66.5%
Stage 3
1279.53
.605
988.47
16
.000
80.3%
Post-Detention Substance-Related Treatment (Y/N)
Stage 1
1599.49
.023
23.74
3
.000
80.3%
Stage 2
1572.92
.048
50.30
9
.000
80.3%
Stage 3
885.60
.577
737.68
16
.000
84.5%
Post-Detention Either/Both Treatment (Y/N)
Stage 1
1260.71
.031
27.67
3
.000
86.6%
Stage 2
1226.34
.068
62.05
9
.000
86.6%
Stage 3
907.76
.381
380.63
16
.000
88.7%
Post-Detention Outpatient Treatment (Y/N)
Stage 1
2076.36
.096
119.70
3
.000
63.8%
Stage 2
2016.38
.141
170.67
9
.000
66.2%
Stage 3
1464.92
.488
731.15
16
.000
78.9%
Post-Detention Non-Outpatient Treatment (Y/N)
.000
Stage 1
2743.83
.025
55.17
3
93.3%
Stage 2
2228.55
.031
69.90
9
.000
93.7%
Stage 3
2067.26
.268
630.81
16
.000
93.9%
Note. N = Number; df = Degrees of freedom; χ2 = Chi-square test (comparing model with
predictors to null model with no predictors); p = Significance level. Stage 1 includes the
following demographic predictors: gender, race, and age. Stage 2 includes the following
criminal history variables: number of prior arrests, number of charges at detention entry,
charge severity, length of stay, violent offender (y/n), and Risk Assessment Inventory
score. Stage 3 includes the following mental health predictor variables: Positive screen on
Massachusetts Youth Screening Instrument-2nd Edition (y/n), conduct disorder (y/n), nonconduct mental disorder (y/n), substance-related disorder (y/n), number of disorders, predetention outpatient treatment (y/n), and pre-detention non-outpatient treatment (y/n).

LogPredictor Variables
Odds
Stage 1: Demographic Variables
Male (y/n)
.099 .189 1.10 0.76 – 1.60 .602
-.074 .190 1.08 0.74 – 1.56 .697
Black (y/n)
-.094 .159 0.91 0.67 – 1.24 .554
-.214 .160 0.81 0.59 – 1.11 .182
Age
-.134 .043 0.88 0.90 – 0.95 .002
-.253 .046 0.78 0.71 – 0.85 .000
Stage 2: Criminal History Variables
Number of Prior Arrests
-.009 .021 0.99 0.95 – 1.03 .680
-.011 .022 0.99 0.95– 1.03 .989
Charge Number
-.232 .113 0.79 0.64 – 0.99 .040
-.062 .121 0.94 0.74 – 1.19 .940
Charge Severity
-.082 .070 0.92 0.80 – 1.06 .238
-.091 .073 0.91 0.79 – 1.05 .214
Length of Stay
.010 .005 1.01 1.00 – 1.20 .036
.009 .005 1.01 0.99 – 1.02 .069
Violent Offender (y/n)
.166 .141 1.18 0.89 – 1.56 .237
.053 .149 1.06 0.79 – 1.41 .721
Risk Assessment Inventory Scores
.009 .019 1.01 0.97 – 1.05 .633
-.003 .019 1.00 0.96 – 1.04 .888
Stage 3: Mental Health Variables
MAYSI-2 – Positive Screen (y/n)
.092 .139 1.10 0.94 – 1.44 .506
.069 .149 1.07 0.80 – 1.43 .645
Conduct Disorder (y/n)
1.747 .210 5.74 3.80 – 8.66 .000 2.940 .428 7.37 1.03 – 11.98 .000
Non-Conduct Mental Disorder (y/n) 1.049 .206 2.86 1.91 – 4.27 .000 1.104 .820 4.07 1.02 – 9.09 .003
Substance-Related Disorder (y/n)
1.163 .154 2.20 2.37 – 4.32 .000
-.792 .160 0.45 0.33 – 0.62 .000
Number of Disorders
.256 .038 1.29 1.20 – 1.39 .000
.320 .036 1.38 1.28 – 1.48 .000
Pre-Detention Outpatient (y/n)
.001 .002 1.00 0.99 – 1.04 .725
-.003 .002 1.00 0.99 – 1.01 .255
Pre-Detention Non-Outpatient
-.379 .108 0.68 0.55 – 0.85 .000
-.266 .107 0.77 0.62 – 0.95 .013
Treatment (y/n)
Note. N = Number; SE = Standard error; CI = Confidence interval; p = Significance level; MAYSI-2 = Massachusetts Youth
Screening Instrument-2nd Edition.

Treatment Outcomes (N = 2114)
Treatment Utilization (y/n)
Mental Health Treatment Utilization (y/n)
Odds
LogOdds
SE
95% CI
p
SE
95% CI
p
Ratios
Odds
Ratios

Table 3.53 Odds Ratios and Associated Statistics for Logistic Regression Models Predicting Treatment and Mental Health
Treatment within Two Years Post-Detention: Cohort Two

223

223

Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
.518 .234
1.68 1.06 – 2.65 .002
.648 .252
1.92 1.16 – 3.13
Black (y/n)
-.014 .178
0.99 0.69 – 1.40 .939 -.075 .195
0.93 0.63 – 1.36
Age
.185 .055
1.20 1.08 – 1.34 .001
.225 .061
1.25 1.11 – 1.41
Stage 2: Criminal History Variables
Number of Prior Arrests
-.006 .024
0.99 0.95 – 1.04 .803
.055 .026
1.06 1.01 – 1.11
Charge Number
-.347 .164
0.71 0.51 – 0.98 .035 -.306 .183
0.74 0.51 – 1.05
Charge Severity
-.049 .085
0.95 0.81 – 1.12 .567
.016 .095
1.01 0.84 – 1.22
Length of Stay
.004 .006
1.00 0.99 – 1.02 .493 -.004 .007
1.00 0.96 – 1.09
Violent Offender (y/n)
-.159 .180
0.85 0.60 – 1.21 .377 -.301 .201
0.74 0.50 – 1.29
Risk Assessment Inventory
-.011 .024
0.90 0.95 – 1.04 .640
.003 .025
1.00 0.95 – 1.06
Stage 3: Mental Health Variables
MAYSI-2 – Positive Screen (y/n)
.284 .186
1.32
0.92 – 1.92 .127
.239 .214
1.27 0.84 – 1.93
Conduct Disorder (y/n)
-1.015 .284
0.36
0.21 – 0.63 .000 2.081 .772
2.01 1.05 – 2.23
Non-Conduct Mental Disorder (y/n)
-.841 .310
0.43 0.34 – 0.79 .007 1.532 .033
1.05 0.99 – 1.02
Substance-Related Disorder (y/n)
2.480 .860
3.49 2.94 – 4.17 .000 5.003 .717
5.91 6.53 – 16.93
Number of Disorders
-.010 .022
0.99 0.95 – 1.04 .638
.098 .023
1.10 1.05 – 1.15
Pre-Detention Outpatient Treatment (y/n)
.003 .003
1.00 0.98 – 1.01 .228 -.007 .003
0.99 0.98 – 1.00
Pre-Detention Non-Outpatient Treatment (y/n) -.253 .107
0.78 0.63 – 0.96 .018 -.245 .108
0.78 0.63 – 0.97
Note. N = Number; SE = Standard error; CI = Confidence interval; p = Significance level; MAYSI-2 = Massachusetts Youth Screening
Instrument-2nd Edition.

.264
.000
.090
.000
.000
.009
.023

.033
.095
.869
.592
.135
.922

.010
.700
.000

Treatment Outcomes (N = 2114)
Substance-Related Treatment Utilization (y/n)
Either/Both Treatment Utilization (y/n)
LogOdds
LogOdds
SE
95% CI
p
SE
95% CI
p
Odds
Ratios
Odds
Ratios

Table 3.54 Odds Ratios and Associated Statistics for Logistic Regression Models Predicting Substance- Related Treatment and
Either/Both Treatment within Two Years Post-Detention: Cohort Two

224

224

Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
.099 .189 1.10 0.76 – 1.60 .602 -.227 .217 0.76 0.50 – 1.16 .202
Black (y/n)
-.094 .159 0.91 0.67 – 1.24 .554 -.476 .185 0.62 0.43 – 0.89 .010
Age
-.134 .043 0.88 0.80 – 0.95 .002
.014 .057 1.02 0.91 – 1.13 .807
Stage 2: Criminal History Variables
Number of Prior Arrests
-.009 .021 0.91 0.95 – 1.03 .680
.045 .025 1.05 0.99 – 1.10 .076
Charge Number
-.232 .113 0.79 0.64 – 0.99 .040 -.246 .175 0.78 0.55 – 1.10 .160
Charge Severity
-.082 .070 0.92 0.80 – 1.06 .238 -.087 .090 0.92 0.77 – 1.09 .337
Length of Stay
.010 .005 1.01 1.00 – 1.20 .036
.004 .006 1.00 0.99 – 1.02 .470
Violent Offender (y/n)
.166 .141 1.19 0.89 – 1.55 .237 -.155 .190 0.86 0.59 – 1.24 .416
Risk Assessment Inventory Scores
.009 .019 1.01 0.97 – 1.05 .633 -.006 .024 0.99 0.95 – 1.04 .816
Stage 3: Mental Health Variables
MAYSI-2 – Positive Screen (y/n)
.092 .139 1.10 0.83 – 1.44 .506 .942 .239 2.56 1.61 – 4.09 .000
Conduct Disorder (y/n)
1.746
.20 5.74 3.80 – 8.66 .000 .663 .410 3.94 1.87 – 4.33 .011
Non-Conduct Mental Disorder (y/n)
1.056 .206 2.86 1.91 – 4.27 .000 .379 .438 1.46 0.62 – 3.45 .388
Substance-Related Disorder (y/n)
1.163 .154 3.20 2.37 – 4.32 .000 .860 .194 2.36 1.61 – 3.46 .000
Number of Disorders
.256 .038 1.29 1.19 – 1.39 .000 .221 .035 1.24 1.29 – 1.31 .000
Pre-Detention Outpatient Treatment (y/n)
.001 .002 1.00 0.99 – 1.01 .725 -.005 .003 1.00 0.99 – 1.01 .058
Pre-Detention Non-Outpatient Treatment (y/n) -.379 .108 0.68 0.55 – 0.85 .000 -.183 .095 0.83 0.69 – 1.00 .055
Note. N = Number; SE = Standard error; CI = Confidence interval; p = Significance level; MAYSI-2 = Massachusetts Youth
Screening Instrument-2nd Edition.

Treatment Outcomes (N = 2114)
Outpatient Treatment Utilization (y/n)
Non-Outpatient Treatment Utilization (y/n)
LogOdds
LogOdds
SE
95% CI
p
SE
95% CI
p
Odds
Ratios
Odds
Ratios

Table 3.55 Odds Ratios and Associated Statistics for Logistic Regression Models Predicting Outpatient Treatment and NonOutpatient Treatment within Two Years Post-Detention: Cohort Two

225

225

Participants Censored
Risk
Treatment Treatment User
NonEnteringa
Casesc
Ratioa
Users
Rateb
Users Survivingc
Time Interval
Number
Number
Number
Number
Proportion
Proportion (SD)
0.0 – 6.0 Months
9664
0
9664.00
1582
.16
.84 (.00)
6.01 – 12.0 Months
7112
0
7112.00
292
.04
.71 (.00)
12.01 – 15.0 Months
6820
59
6790.00
236
.03
.68 (.00)
15.01 – 18.0 Months
6525
52
6499.00
165
.03
.66 (.00)
18.01 – 21.0 Months
6308
59
6278.50
190
.03
.64 (.00)
21.01 – 24.0 Months
6059
49
6034.00
155
.03
.63 (.00)
2.0 – 2.5 Years
5855
59
5825.00
142
.02
.60 (.01)
2.51 – 3.0 Years
5473
77
5434.50
279
.05
.58 (.01)
3.01 – 3.5 Years
5132
81
5091.00
265
.05
.54 (.01)
3.51 – 4.0 Years
4786
82
4745.00
221
.05
.51 (.01)
4.01 – 4.5 Years
4483
84
4441.00
207
.05
.49 (.01)
4.51 – 5.0 Years
4159
90
4114.00
207
.05
.46 (.01)
5.01 – 5.5 Years
3862
77
3823.00
163
.04
.44 (.01)
5.51 – 6.0 Years
3622
64
3590.00
167
.05
.42 (.01)
6.01 – 6.5 Years
3391
98
3342.00
125
.04
.41 (.01)
6.51 – 7.0 Years
3168
99
3118.50
127
.04
.39 (.01)
7.01 – 7.5 Years
2942
155
2864.50
121
.04
.37 (.01)
7.51 – 8.0 Years
2666
185
2573.50
103
.04
.36 (.01)
8.01 – 8.5 Years
2378
153
2301.50
74
.03
.35 (.01)
8.51 – 9.0 Years
2151
159
2071.50
89
.04
.33 (.01)
9.01 – 9.5 Years
1903
124
1841.00
71
.04
.32 (.01)
9.51 – 10.0 Years
1708
143
1636.50
68
.04
.31 (.01)
10.01 – 10.5 Years
1497
142
1426.00
68
.05
.29 (.01)
10.51 – 11.0 Years
1287
146
1214.00
46
.04
.28 (.01)
11.01 – 11.5 Years
1095
156
1017.00
45
.04
.27 (.01)
11.51 – 12.0 Years
894
171
808.50
40
.05
.25 (.01)
12.01 – 12.5 Years
683
143
611.50
27
.04
.24 (.01)
12.51 – 13.0 Years
513
121
452.50
22
.05
.23 (.01)
13.01 – 13.5 Years
370
143
298.50
12
.04
.22 (.01)
13.51 – 14.0 Years
216
212
110.00
3
.03
.22 (.01)
Note. SD = Standard deviation. a At start of time interval. b During time interval; c At end of time interval.

Table 3.56 Life Table: Time to First Post-Detention Treatment Utilization
Probability of
Treatmentb
Probability (SD)
.055 (.001)
.010 (.00)
.008 (.00)
.006 (.00)
.007 (.00)
.006 (.00)
.005 (.00)
.005 (.00)
.005 (.00)
.004 (.00)
.005 (.00)
.004 (.00)
.003 (.00)
.003 (.00)
.003 (.00)
.003 (.00)
.003 (.00)
.002 (.00)
.002 (.00)
.002 (.00)
.002 (.00)
.002 (.00)
.002 (.00)
.002 (.00)
.002 (.00)
.002 (.00)
.002 (.00)
.002 (.00)
.002 (.00)
.001 (.00)

Hazard
Rateb
Rate (SD)
.06 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.01 (.00)
.00 (.00)

226

226

Age
Cohorts

Race

Gender

Treatment
Setting

Treatment
Type

Sample

Variables

---2212 (35.5%)
1016 (28.2%)
822 (24.4%)
2203 (38.9%)
590 (24.5%)

1222 (12.6%)
403 (4.2%)
5761 (59.6%)
675 (6.98%)
3850 (63.5%)
2586 (71.8%)
2545 (75.6%)
3464 (61.1%)
1816 (75.5)a

Substance-Related

Either/Both
Outpatient
Non-Outpatient
Male
n = 6062
Female
n = 3602
White
n = 3367
Black
n = 3464
Younger
n = 2406
Mid-Age
n = 3924
Older
n = 3334
1232 (31.4%)
1406 (42.2%)

2692 (68.6%)b
1928 (57.8%)c

--

3228 (33.4%)
--

Number (%)

Non-Users
(Censored)

6436 (66.6%)
4811 (49.8%)

Number (%)

Treatment
Users

Categories/
Groups
Any Treatment
Mental Health

Total Sample
N = 9664

561.0 (58.39)

455.0 (24.98)

591.0 (31.67)

456.0 (23.20)

699.0 (37.64)

1191.0 (41.89)

307.0 (14.20)

685.0 (85.44)
439.0 (17.29)
1528.0 (49.52)

879.0 (56.38)

531.0 (19.28)
446.0 (19.96)

Median (SD)

446.55 – 675.45

406.03 – 503.96

528.93 – 653.07

410.51 – 501.48

625.23 – 772.77

1108.90 – 1273.11

279.18 – 334.82

517.57 – 852.46
405.11 – 472.89
1430.95 – 1625.05

768.49 – 989.51

493.21 – 568.79
406.88 – 485.12

95% CI

Time to First Treatment
(in Days)

15.59,
p < .001

21.08,
p < .001

507.33,
p < .001

198.74,
p < .001

86.65,
p < .001

28.19,
p < .001

23.60,
p < .001

540.08,
p < .001

176.28,
p < .001

83.88,
p < .001

39.97,
p < .001

20.23,
p < .001

454.37,
p < .001

109.87,
p < .001

68.75,
p < .001

Survival Curve Indicators
Breslow/
TaroneLog Rank
Wilcoxon
Ware
2
2
χ,
χ,
χ 2,
p-value
p-value
p-value
----

Table 3.57 Time to First Treatment: Demographic, Mental Health, and Criminal History Groups

227

227

MAYSI-2

No MAYSI-2
4732 (66.3%)a
2400 (33.7%)
992.0 (26.79)
939.48 – 1044.51
n = 7132
Non-positive
1234 (70.2%)b
524 (29.8%)
111.0 (11.28)
88.88 – 133.11
n = 1758
1868.98,
2053.05,
2213.78,
p < .001
p < .001
p < .001
Positive Screen
b
470 (60.7%)
304 (39.3%)
93.0 (6.83)
79.61 – 106.39
n = 774
Mental Disorder
3093 (86.6%)a
478 (13.4%)
516.0 (24.20)
468.55 – 563.44
n = 357
Substance
81.91,
109.19,
102.69,
Disorder Type
627 (94.7%)b
35 (5.3%)
1174.0 (81.28)
1014.69 – 1333.31
n = 662
p < .001
p < .001
p < .001
Comorbid
2716 (97.5%)c
66 (2.4%)
424.0 (19.28)
493.31 – 568.79
n = 2782
No treatment
4574 (63.3%)
2652 (36.7%)
1047.0 (25.47)
997.07 – 1096.92
n = 7226
Pre-Detention
3604.13,
3617.47,
3493.60,
Treatment
p < .001
p < .001
p < .001
Prior Treatment
1862 (76.4%)
576 (23.6%)
46.0 (3.47)
39.19 – 52.81
n = 2438
No
4020 (65.7%)
2102 (34.3%)
580.0 (29.21)
405.03 – 502.97
n = 6122
Violent
28.82,
33.32,
33.93,
Offender
p < .001
p < .001
p < .001
Yes
2416 (68.2%)
1126 (31.8%)
454.0 (24.98)
493.21 – 568.79
n = 3546
No
2577 (58.1%)
1860 (41.9%)
531.0 (19.29)
286.47 – 351.53
n = 4437
170.48,
182.07,
153.46,
Recidivist
p < .001
p < .001
p < .001
Yes
3859 (73.8%)
1368 (26.2%)
318.0 (16.59)
735.79 – 864.20
n = 5227
No
2526 (58.0%)
1832 (42.0%)
536.0 (39.03)
456.51 – 609.49
n = 4358
Re-Detained/
0.04,
1.11,
2.08,
Incarcerated
p = .842
p = .292
p = .149
Yes
3910 (73.7%)
1396 (26.3%)
521.0 (21.59)
488.67 – 573.32
n = 5306
2
Note. n = Number; % = Percentage; χ = Chi-square test; p-value = Significance level; MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition. Different
superscripts indicate significant differences (p ≤ .01) and matching superscripts indicate non-significant differences (p > .01) based upon 2 x 2 chi-square tests between
paired groups.

Table 3.57 Continued.

228

228

229
Table 3.58 Model Summary Statistics for Cox Regression Models Predicting Time to
First Treatment
Model Summary Statistics (N = 8554)

Change Statistics*

Treatment
-2 LL
χ2
df
p
χ2
df
p
Outcomes
Post-Detention Treatment
Stage 1
98306.19
243.42
3
.000
239.08
3
.000
Stage 2
97834.80
767.24
9
.000
471.34
6
.000
Stage 3
91172.09
7807.45
15
.000
6112.12
6
.000
Post-Detention Mental Health Treatment
Stage 1
83658.67
321.58
3
.000
315.95
3
.000
Stage 2
83252.88
760.64
9
.000
404.78
6
.000
Stage 3
76243.71
8364.56
15
.000
7010.12
6
.000
Post-Detention Substance-Related Treatment
Stage 1
38927.16
122.12
3
.000
118.82
3
.000
Stage 2
38664.15
401.15
9
.000
263.08
6
.000
Stage 3
31685.29
7809.60
15
.000
6978.50
6
.000
Post-Detention Either/Both Treatment
Stage 1
25228.87
135.04
3
.000
129.99
3
.000
Stage 2
24992.11
408.50
9
.000
236.77
6
.000
Stage 3
19565.63
6927.60
15
.000
5426.05
6
.000
Post-Detention Outpatient Treatment
95217.64
230.72
3
.000
226.91
3
.000
Stage 1
Stage 2
94740.13
761.92
9
.000
477.10
6
.000
Stage 3
88743.73
7659.04
15
.000
5996.40
6
.000
Post-Detention Non-Outpatient Treatment
Stage 1
27177.58
341.45
3
.000
328.02
3
.000
Stage 2
27072.54
444.90
9
.000
105.07
6
.000
Stage 3
2396.08
53393.14
15
.000
3103.62
6
.000
2
Note. N = Number; -2 LL = -2 Log Likelihood; df = Degrees of freedom; χ = Chi-square
test; p = Significance level. * Results based on the addition of predictors for the individual
Stage, compared to previous Stage. Stage 1 includes the following demographic predictor
variables: gender, race, and age. Stage 2 includes the following criminal history predictor
variables: number of prior arrests, number of charges at detention entry, charge severity,
length of stay, and violent offender (y/n). Stage 3 includes the following mental health
predictor variables: conduct disorder (y/n), non-conduct mental disorder (y/n), substancerelated disorder (y/n), number of disorders, pre-detention outpatient treatment (y/n), and
pre-detention non-outpatient treatment (y/n).

SE
.029
.028
.008
.005
.013
.013
.001
.028
.030
.035
.040

Beta
.045
-.028
-.052
.035
-.009
.041
.001
.107
.170
1.107
1.365

3.03
3.92

1.04
0.99
1.04
1.00
1.11
1.19

1.05
0.97
0.95

HR

p
.117
.305
.000
.000
.494
.002
.323
.000
.000
.000
.000

95% CI
0.99 – 1.11
0.92 – 1.03
0.93 – 0.97
1.02 – 1.06
0.97 – 1.02
1.01 – 1.07
0.99 – 1.01
1.05 – 1.18
1.12 – 1.26
2.83 – 3.24
3.62 – 4.23

.005
.013
.015
.001
.030
.033

.031
.030
.009

SE

1.241 .037
2.30 .054

.033
-.010
.034
.000
.084
.171

-.022
-.040
-.069

Beta

p

1.02 – 1.05
0.97 – 1.02
1.01 – 1.07
0.99 – 1.00
1.03 – 1.15
1.11 – 1.27

.000
.443
.019
.664
.006
.000

0.92 – 1.04 .481
0.91 – 1.02 .175
0.92 – 0.95 .000

95% CI

3.46 3.22 – 3.72 .000
9.97 8.96 – 11.10 .000

1.03
0.99
1.05
1.00
1.09
1.19

0.98
0.96
0.93

HR

First Mental Health Treatment Utilization

Substance-Related Disorder (y/n)
.668 .031 1.95 1.83 – 2.07 .000
.047 .033 1.05 0.98 – 1.12 .156
Number of Disorders
.005 .003 1.01 0.99 – 1.01 .096
.023 .003 1.02 1.01 – 1.03 .000
Pre-Detention Outpatient (y/n)
.414 .040 1.51 1.40 –1 .64 .000
.302 .043 1.35 1.24 – 1.47 .000
Pre-Detention Non-Outpatient (y/n) .096 .053 1.10 0.99 – 1.22 .074
.114 .056 1.12 1.01 – 1.25 .042
Note. N = Number; Beta = Unstandardized beta weight; SE = Standard error; HR = Hazard ratio; CI = Confidence interval; p =
Significance level.

Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
Black (y/n)
Age
Stage 2: Criminal History Variables
Number of Prior Arrests
Charge Number
Charge Severity
Length of Stay
Violent Offender (y/n)
Re-Detained/Incarcerated (y/n)
Stage 3: Mental Health Variables
Conduct Disorder (y/n)
Non-Conduct Mental Disorder (y/n)

First Treatment Utilization

Treatment Outcomes (N = 8554)

Table 3.59 Hazard Ratios and Associated Statistics for Cox Regression Models Predicting Time to First Treatment and First
Mental Health Treatment

230

230

.000
.296
.115
.000
.040

1.11 – 1.33
0.88 – 1.04
0.95 – 1.01
1.02 – 1.05
0.88 – 0.99

1.03
0.94

.044
-.047

.189
-.027
-.011

Beta

.008
.049

.057
.054
.017

SE

1.05
0.95

1.21
0.97
0.99

HR

1.08 – 1.35
0.87 – 1.08
0.96 – 1.02

95% CI

.001
.619
.503

p

.004

.000
.000
.215
.000
.000

.122
.946
.022
.042

.007
.031

.031
-.064

1.22
0.96
0.98

p

95% CI

Charge Severity
.037 .021 1.04
1.00 – 1.08 .071
.040
.026 1.04 0.99 – 1.10
Length of Stay
.000 .002 1.00
1.00 – 1.01 .770
.000
.002 1.00 1.00 – 1.01
Violent Offender (y/n)
.046 .047 1.05
0.96 – 1.15 .326
.130
.057 1.14 1.02 – 1.29
Re-Detained/Incarcerated (y/n)
.179 .047 1.20
1.09 – 1.31 .000
.123
.061 1.13 1.00 – 1.27
Stage 3: Mental Health Variables
Conduct Disorder (y/n)
.686 .054 1.99
1.79 – 2.01 .000
1.020
.068 2.77 2.43 – 3.17
Non-Conduct Mental Disorder (y/n)
-.256 .052 0.77
0.70 – 0.86 .000
1.440
.114 4.22 3.38 – 5.27
Substance-Related Disorder (y/n)
13.188 7.772 3.98 0.13 – 22.09 .090
12.466 10.062 3.84 0.01 – 5.26
Number of Disorders
.017 .004 1.02
1.00 – 1.04 .005
.050
.004 1.05 1.04 – 1.07
Pre-Detention Outpatient Treatment (y/n) .848 .065 2.34
2.05 – 2.65 .000
.928
.082 2.53 2.15 – 2.97
Pre-Detention Non-Outpatient
-.334 .089 0.72
0.60 – 0.85 .000
-.277
.097 0.76 0.63 – 0.92
Treatment (y/n)
Note. N = Number; Beta = Unstandardized beta weight; SE = Standard error; HR = Hazard ratio; CI = Confidence interval; p =
Significance level.

.047
.043
.014

.197
-.045
-.022

HR

.000
.331

SE

Beta

1.03 – 1.06
0.87 – 1.05

Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
Black (y/n)
Age
Stage 2: Criminal History Variables
Number of Prior Arrests
Charge Number

Treatment Outcomes (N = 8554)
First Substance-Related Treatment Utilization
First Either/Both Treatment Utilization

Table 3.60 Hazard Ratios and Associated Statistics for Cox Regression Models Predicting Time to First Substance-Related
Treatment and First Either/Both Treatment

231

231

Beta SE
HR
95% CI
p
Beta
SE
HR
95% CI
Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
.034 .029 1.04
0.98 – 1.10 .245 -.029 .053 0.97 0.88 – 1.08
Black (y/n)
-.009 .028 0.99
0.94 – 1.05 .755 -.293 .052 0.75 0.67 – 0.83
Age
-.054 .009 0.95
0.93 – 0.96 .000 -.072 .016 0.93 0.90 – 0.96
Stage 2: Criminal History Variables
Number of Prior Arrests
.036 .005 1.04
1.03 – 1.05 .000
.027 .009 1.03 1.01 – 1.05
Charge Number
-.011 .014 0.99
0.96 – 1.02 .444 -.103 .053 0.90 0.81 – 1.00
Charge Severity
.035 .014 1.05
1.01 – 1.06 .011
.015 .024 1.02 0.97 – 1.07
Length of Stay
.001 .001 1.00
1.00 – 1.01 .546
.001 .002 1.00 1.00 – 1.01
Violent Offender (y/n)
.110 .029 1.12
1.06 – 1.19 .000
.041 .054 1.04 0.94 – 1.16
Re-Detained/Incarcerated (y/n)
.167 .030 1.18
1.11 – 1.25 .000
.076 .056 1.08 0.97 – 1.20
Stage 3: Mental Health Variables
Conduct Disorder (y/n)
1.137 .035 3.12
2.91 – 3.34 .000
.720 .066 2.06 1.81 – 2.34
Non-Conduct Mental Disorder (y/n)
1.392 .041 4.02
3.71 – 4.36 .000 1.745 .093 5.73 4.77 – 6.88
Substance-Related Disorder (y/n)
.584 .031 1.79
1.69 – 1.91 .000 1.199 .064 3.32 2.93 – 3.76
Number of Disorders
.008 .003 1.01
1.00 – 1.02 .010
.065 .004 1.07 1.06 – 1.08
Pre-Detention Outpatient Treatment (y/n) .432 .041 1.54
1.42 – 1.67 .000
.394 .090 1.48 1.24 – 1.77
Pre-Detention Non-Outpatient Treatment
.062 .054 1.06
0.96 – 1.18 .253
.295 .106 1.34 1.09 – 1.66
(y/n)
Note. N = Number; Beta = Unstandardized beta weight; SE = Standard error; HR = Hazard ratio; CI = Confidence interval; p =
Significance level.

.006

.000
.000
.000
.000
.000

.003
.054
.529
.489
.449
.177

.586
.000
.000

p

Treatment Outcomes (N = 8554)
First Outpatient Treatment Utilization
First Non-Outpatient Treatment Utilization

Table 3.61 Hazard Ratios and Associated Statistics for Cox Regression Models Predicting Time to First Outpatient Treatment
and First Non-Outpatient Treatment

232

232

233
Table 3.62 Model Summary Statistics for Cox Regression Models Predicting Time to
First Treatment: Cohort One
Model Summary Statistics (N = 6203)

Change Statistics*

Treatment
-2 LL
χ2
df
p
χ2
df
p
Outcomes
Post-Detention Treatment
Stage 1
68966.9
269.81
3
.000
266.23
3
.000
Stage 2 68824.49
411.98
9
.000
142.40
6
.000
Stage 3 63876.44
6581.82
15
.000
4948.05
6
.000
Post-Detention Mental Health Treatment
Stage 1 59107.17
327.53
3
.000
323.85
3
.000
Stage 2 58981.98
451.27
9
.000
125.29
6
.000
Stage 3 53176.47
7251.00
15
.000
5805.11
6
.000
Post-Detention Substance-Related Treatment
Stage 1
28108.81
110.65
3
.000
107.97
3
.000
Stage 2
27894.37
234.76
9
.000
124.54
6
.000
Stage 3
22452.60
6368.45
15
.000
5531.76
6
.000
Post-Detention Either/Both Treatment
Stage 1
16719.18
126.81
3
.000
122.63
3
.000
Stage 2
16655.88
189.91
9
.000
63.32
6
.000
Stage 3
12656.03
5386.74
15
.000
3999.51
6
.000
Post-Detention Outpatient Treatment
66409.72
260.81
3
.000
257.87
3
.000
Stage 1
Stage 2
66259.79
410.63
9
.000
149.27
6
.000
Stage 3
61404.10
6483.21
15
.000
4855.96
6
.000
Post-Detention Non-Outpatient Treatment
Stage 1
21990.37
246.83
3
.000
239.98
3
.000
Stage 2
21942.23
294.60
9
.000
48.30
6
.000
Stage 3
19275.46
4452.95
15
.000
2666.76
6
.000
2
Note. N = Number; -2 LL = -2 Log Likelihood; df = Degrees of freedom; χ = Chi-square
test; p = Significance level. * Results based on the addition of predictors for the individual
Stage, compared to previous Stage. Stage 1 includes the following demographic predictor
variables: gender, race, and age. Stage 2 includes the following criminal history predictor
variables: number of prior arrests, number of charges at detention entry, charge severity,
length of stay, and violent offender (y/n). Stage 3 includes the following mental health
predictor variables: conduct disorder (y/n), non-conduct mental disorder (y/n), substancerelated disorder (y/n), number of disorders, pre-detention outpatient treatment (y/n), and
pre-detention non-outpatient treatment (y/n).

b

c

Mental Health Substance-Related

Either/Both

Outpatient

+
+
+

+

+

–
–
–

2

Non-Outpatient

Total
1
2 Total 1 2 Total 1
2 Total
1
2 Total 1
2 Total 1
Predictor Variables
Demographic Variables
Male (y/n)
+
+
+
+
+
+
Black (y/n)
–
–
–
–
–
–
Age
–
–
–
–
–
–
–
–
–
–
–
–
Criminal History Variables
Number of Prior Arrests
+
+
+
+
+
+
+
+
+
+
+
+
Charge Number
–
Charge Severity
+
+
+
+
+
+
Length of Stay
Violent Offender (y/n)
+
+
+
+
+
+
+
+
+
Re-Detained/Incarcerated (y/n)
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
RAI Scores
NA NA
NA NA
NA NA
NA NA
NA NA
NA NA
Mental Health Variables
MAYSI-2 –
NA NA
NA NA
NA NA
NA NA
NA NA
NA NA
Positive Screen (y/n)
Conduct Disorder (y/n)
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
Non-Conduct Mental Disorder (y/n)
+
+
+
+
+
+
–
–
+
+
+
+
+
+
+
+
Substance-Related Disorder (y/n)
+
+
+
+
+
+
+
+
+
Number of Disorders
+
+
+
+
+
+
+
+
+
+
+
+
+
+
Pre-Detention Outpatient
+
+
–
+
+
–
+
+
+
+
–
+
+
–
+
+
Treatment (y/n)
Pre-Detention Non-Outpatient
+
–
–
–
–
–
+
Treatment (y/n)
Note. NA = Not applicable because predictor excluded from analysis; Total = Total Sample; 1 = Cohort One; 2 = Cohort Two; (+) = Increased Risk;
(–) = Decreased Risk; RAI = Risk Assessment Inventory; MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition. a N = 8544. b N = 6203.
c
N = 2341.

a

Any Treatment

Treatment Outcomes

Table 3.63 Predictors of Cox Regression Analyses Associated with Increased Risk (+) and Decreased Risk (–) of Time to First
Treatment: Total Sample, Cohort One, and Cohort Two

234

234

.033
.032
.010
.008
.011
.016

.061
-.023
-.039
.046
-.004
.046

Beta

1.05
0.99
1.05

1.06
0.98
0.96

.064
.479
.000
.000
.726
.003

0.99 – 1.13
0.92 – 1.04
0.94 – 0.98
1.03 – 1.06
0.98 – 1.02
1.02 – 1.08

.043 .009
-.006 .011
.041 .017

.021 .036
-.015 .034
-.048 .010

1.04
0.99
1.04

1.02
0.99
0.95

1.03 – 1.06
0.97 – 1.02
1.01 – 1.08

0.95 – 1.09
0.92 – 1.05
0.93 – 0.97

.000
.603
.015

.548
.654
.000

Length of Stay
-.001 .001 0.99 0.99 – 1.03 .626 -.001 .001 0.99 0.98 – 1.04 .496
Violent Offender (y/n)
.158 .033 1.17 1.10 – 1.25 .000
.142 .035 1.15 1.08 – 1.24 .000
Re-Detained/Incarcerated (y/n)
.134 .034 1.14 1.07 – 1.22 .000
.141 .037 1.15 1.07 – 1.24 .000
Stage 3: Mental Health Variables
Conduct Disorder (y/n)
1.236 .040 3.44 3.18 – 3.72 .000 1.343 .042 3.83 3.53 – 4.16 .000
Non-Conduct Mental Disorder (y/n) 1.611 .046 5.01 4.58 – 5.48 .000 2.608 .063 13.57 11.98 – 15.36 .000
Substance-Related Disorder (y/n)
.631 .036 1.88 1.75 – 2.02 .000 -.013 .039 0.99 0.91 – 1.07 .741
Number of Disorders
-.008 .004 0.99 0.96 – 1.01 .052
.012 .004 1.01 1.00 – 1.02 .001
Pre-Detention Outpatient
.742 .055 2.10 1.89 – 2.34 .000
.650 .057 1.92 1.72 – 2.14 .000
Treatment (y/n)
Pre-Detention Non-Outpatient
-.119 .095 0.89 0.74 – 1.07 .212 -.176 .100 0.84 0.69 – 1.02 .077
Treatment (y/n)
Note. N = Number; Beta = Unstandardized beta weight; SE = Standard error; HR = Hazard ratio; CI = Confidence interval; p =
Significance level.

Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
Black (y/n)
Age
Stage 2: Criminal History Variables
Number of Prior Arrests
Charge Number
Charge Severity

Treatment Outcomes (N = 6203)
First Treatment Utilization
First Mental Health Treatment Utilization
SE
HR
95% CI
p
Beta SE
HR
95% CI
p

Table 3.64 Hazard Ratios and Associated Statistics for Cox Regression Models Predicting Time to First Treatment and First
Mental Health Treatment: Cohort One

235

235

p

.112

.000

.000
.000
.168
.000

.001
.218
.438
.792
.000
.001

.026
.399
.197

First Either/Both Treatment Utilization

Beta
SE
HR
95% CI
p
Beta
SE
HR
95% CI
Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
.149 .051 1.16 1.05 – 1.29 .004
.146
.066 1.16 1.02 – 1.32
Black (y/n)
-.076 .050 0.93 0.84 – 1.02 .128
-.054 .064 0.94 0.84 – 1.08
Age
-.002 .016 0.99 0.97 – 1.03 .920
.026
.020 1.03 0.99 – 1.07
Stage 2: Criminal History Variables
Number of Prior Arrests
.050 .013 1.05 1.02 – 1.08 .000
.060
.018 1.06 1.03 – 1.10
Charge Number
-.065 .034 0.94 0.88 – 1.00 .059
-.084 .068 0.92 0.80 – 1.05
Charge Severity
.036 .024 1.04 0.99 – 1.09 .132
.024
.031 1.02 0.96 – 1.09
Length of Stay
-.001 .002 0.99 0.99 – 1.00 .541
-.001 .002 0.99 0.99 – 1.00
Violent Offender (y/n)
.110 .054 1.12 1.01 – 1.24 .042
.255
.068 1.29 1.13 – 1.48
Re-Detained/Incarcerated (y/n)
.130 .054 1.14 1.02 – 1.27 .016
.060
.018 1.06 1.03 – 1.10
Stage 3: Mental Health Variables
Conduct Disorder (y/n)
.887 .063 2.43 2.15 - 2.75
.000
.991
.079 2.69 2.31 – 3.14
Non-Conduct Mental Disorder (y/n)
-.333 .060 0.72 0.64 – 0.81 .000
2.031 .173 7.62 5.43 – 10.70
Substance-Related Disorder (y/n)
13.323 8.901 1.27 1.20 – 3.74 .003 11.687 8.487 1.02 0.97 – 1.23
Number of Disorders
.016 .005 1.02 1.01 – 1.03 .001
.048
.005 1.05 1.03 – 1.06
Pre-Detention Outpatient
1.362 .089 3.91 3.28 – 4.65 .000
1.276 .120 3.58 2.83 – 4.53
Treatment (y/n)
Pre-Detention Non-Outpatient
-.597 .148 0.55 0.41 – 0.74 .000
-.283 .178 0.75 0.53 – 1.07
Treatment (y/n)
Note. N = Number; Beta = Unstandardized beta weight; SE = Standard error; HR = Hazard ratio; CI = Confidence interval; p =
Significance level.

First Substance-Related Treatment Utilization

Treatment Outcomes (N = 6203)

Table 3.65 Hazard Ratios and Associated Statistics for Cox Regression Models Predicting Time to First Substance-Related
Treatment and First Either/Both Treatment: Cohort One

236

236

Beta
SE
HR
95% CI
p
Beta
SE
HR
95% CI
p
Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
.042 .034
1.04
0.98 – 1.11 .205
. 052 .058 1.05 0.94 – 1.18 .365
Black (y/n)
-.004 .033
0.99
0.93 – 1.06 .895
-.250 .056 0.80 0.70 – 0.87 .000
Age
-.042 .010
0.96
0.94 – 0.97 .000
-.044 .017 0.96 0.93 – 0.99 .010
Stage 2: Criminal History Variables
Number of Prior Arrests
.050 .009
1.05
1.03 – 1.07 .000
.051 .016 1.05 1.02 – 1.09 .001
Charge Number
-.005 .012
0.99
0.97 – 1.02 .659
-.102 .060 0.90 0.80 – 1.02 .091
Charge Severity
.038 .016
1.04
1.01 – 1.07 .019
.043 .027 1.04 0.99 – 1.10 .104
Length of Stay
-.001 .001
0.99
0.99 – 1.01 .359
.002 .002 1.00 0.99 – 1.00 .303
Violent Offender (y/n)
.154 .033
1.17
1.09 – 1.25 .000
.085 .059 1.09 0.97 – 1.22 .151
Re-Detained/Incarcerated (y/n)
.130 .035
1.14
1.06 – 1.22 .000
.037 .061 1.04 0.92 – 1.17 .536
Stage 3: Mental Health Variables
Conduct Disorder (y/n)
1.262 .041
3.53
3.26 – 3.83 .000
.786 .071 2.20 1.91 – 2.52 .000
Non-Conduct Mental Disorder (y/n) 1.655 .048
5.23
4.76 – 5.74 .000 1.877 .102 6.53 5.35 – 7.97 .000
Substance-Related Disorder (y/n)
.526 .037
1.69
1.57 – 1.82 .000 1.178 .070 3.25 2.83 – 3.73 .000
Number of Disorders
-.004 .004
0.99
0.99 – 1.00 .327
.059 .004 1.06 1.05 – 1.07 .000
Pre-Detention Outpatient
.779 .055
2.19
1.96 – 2.43 .000
.816 .111 2.26 1.83 – 2.81 .000
Treatment (y/n)
Pre-Detention Non-Outpatient
-.207 .096
0.81
0.67 – 0.98 .031
.234 .158 1.26 0.93 – 1.72 .138
Treatment (y/n)
Note. N = Number; Beta = Unstandardized beta weight; SE = Standard error; HR = Hazard ratio; CI = Confidence interval; p =
Significance level.

Treatment Outcomes (N = 6203)
First Outpatient Treatment Utilization
First Non-Outpatient Treatment Utilization

Table 3.66 Hazard Ratios and Associated Statistics for Cox Regression Models Predicting Time to First Outpatient Treatment
and First Non-Outpatient Treatment: Cohort One

237

237

238
Table 3.67 Model Summary Statistics for Cox Regression Models Predicting Time to
First Treatment: Cohort Two
Model Summary Statistics (N = 2341)
Change Statistics*
Treatment
-2 LL
χ2
df
p
χ2
df
p
Outcomes
Post-Detention Treatment
Stage 1
13799.56
141.68
3
.000
134.26
3
.000
Stage 2
13665.47
282.98
10
.000
134.09
7
.000
Stage 3
13046.73
1083.78
17
.000
618.74
7
.000
Post-Detention Mental Health Treatment
Stage 1
11248.06
173.40
3
.000
165.89
3
.000
Stage 2
11121.68
307.19
10
.000
126.71
7
.000
Stage 3
10455.84
1123.79
17
.000
665.48
7
.000
Post-Detention Substance-Related Treatment
Stage 1
5180.62
37.56
3
.000
35.21
3
.000
Stage 2
5111.88
108.62
10
.000
68.74
7
.000
Stage 3
4150.01
1093.20
17
.000
961.89
7
.000
Post-Detention Either/Both Treatment
Stage 1
3978.07
83.66
3
.000
76.92
3
.000
Stage 2
3887.77
181.89
10
.000
90.24
7
.000
Stage 3
3079.81
1144.18
17
.000
807.58
7
.000
Post-Detention Outpatient Treatment
13621.94
134.95
3
.000
128.61
3
.000
Stage 1
Stage 2
13494.50
268.75
10
.000
127.43
7
.000
Stage 3
12889.12
1054.11
17
.000
605.48
7
.000
Post-Detention Non-Outpatient Treatment
Stage 1
1864.87
129.56
3
.000
107.24
3
.000
Stage 2
1793.53
208.64
10
.000
71.33
7
.000
Stage 3
1590.84
715.57
17
.000
202.69
7
.000
2
Note. N = Number; -2 LL = -2 Log Likelihood; df = Degrees of freedom; χ = Chi-square
test; p = Significance level. * Results based on the addition of predictors for the individual
Stage, compared to previous Stage. Stage 1 includes the following demographic predictor
variables: gender, race, and age. Stage 2 includes the following criminal history predictor
variables: Risk Assessment Inventory scores, number of prior arrests, number of charges at
detention entry, charge severity, length of stay, and violent offender (y/n). Stage 3 includes
the following mental health predictor variables: Positive Screen on Massachusetts Youth
Screening Instrument-2nd Edition Positive Screen, conduct disorder (y/n), non-conduct
mental disorder (y/n), substance-related disorder (y/n), number of disorders, pre-detention
outpatient treatment (y/n), and pre-detention non-outpatient treatment (y/n).

Beta
Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
-.108
.086 0.90 0.76 – 1.06 .211 -.135 .093 0.87 0.73 – 1.05 .149
Black (y/n)
-.199
.072 0.82 0.71 – 0.94 .006 -.224 .079 0.80 0.68 – 0.93 .005
Age
-.123
.025 0.88 0.84 – 0.93 .000 -.162 .028 0.85 0.81 – 0.90 .000
Stage 2: Criminal History Variables
Number of Prior Arrests
.015
.008 1.02 1.00 – 1.03 .067
.014 .009 1.01 1.00 – 1.03 .115
Charge Number
-.079
.059 0.92 0.82 – 1.04 .180 -.065 .065 0.94 0.82 – 1.06 .316
Charge Severity
-.011
.034 0.99 0.93 – 1.06 .754
.013 .038 1.01 0.94 – 1.09 .737
Length of Stay
.001
.002 1.00 1.00 – 1.01 .489
.001 .002 1.00 1.00 – 1.01 .673
Violent Offender (y/n)
.089
.069 1.10 0.96 – 1.25 .197
.060 .076 1.06 0.92 – 1.23 .432
Re-Detained/Incarcerated (y/n)
.281
.078 1.33 1.14 – 1.54 .000
.286 .088 1.33 1.12 – 1.58 .001
Risk Assessment Inventory Scores
.001
.009 1.00 0.98 – 1.02 .892 -.001 .010 1.00 0.98 – 1.02 .957
Stage 3: Mental Health Variables
MAYSI-2 - Positive Screen (y/n)
.021
.071 1.02 0.89 – 1.17 .768
.055 .081 1.06 0.90 – 1.24 .498
Conduct Disorder (y/n)
.658
.088 1.93 1.62 – 2.30 .000 1.017 .106 2.76 2.25 – 3.40 .000
Non-Conduct Mental Disorder (y/n)
.741
.104 2.10 1.71 – 2.57 .000 1.445 .143 4.24 3.21 – 5.61 .000
Substance-Related Disorder (y/n)
.745
.072 2.11 1.83 – 2.43 .000
.100 .078 1.11 0.95 – 1.30 .203
Number of Disorders
.060
.008 1.06 1.05 – 1.08 .000
.072 .008 1.07 1.06 – 1.09 .000
Pre-Detention Outpatient (y/n)
-.308
.091 0.74 0.62 – 0.88 .001 -.458 .103 0.63 0.52 – 0.77 .000
Pre-Detention Non-Outpatient
-.035
.092 0.97 0.81 – 1.16 .701
.011 .099 1.01 0.83 – 1.23 .915
Treatment (y/n)
Note. N = Number; Beta = Unstandardized beta weight; SE = Standard error; HR = Hazard ratio; CI = Confidence interval; p =
Significance level; MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition.

Treatment Outcomes (N = 2341)
First Treatment Utilization
First Mental Health Treatment Utilization
SE
HR
95% CI
p
Beta
SE
HR
95% CI
p

Table 3.68 Hazard Ratios and Associated Statistics for Cox Regression Models Predicting Time to First Treatment and First
Mental Health Treatment: Cohort Two

239

239

First Either/Both Treatment Utilization

Beta
SE
HR
95% CI
p
Beta
SE
HR
95% CI
p
Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
.419
.170
1.52 1.09 – 2.12 .014
.474
.186
1.61 1.12 – 2.31 .011
Black (y/n)
-.174
.118
0.84 0.67 – 1.06 .140
-.323
.129
0.72 0.56 – 0.93 .012
Age
-.066
.043
0.94 0.86 – 1.02 .127
-.115
.049
0.89 0.81 – 0.98 .019
Stage 2: Criminal History Variables
Number of Prior Arrests
.008
.013
1.01 0.98 – 1.04 .541
.020
.014
1.02 0.99 – 1.05 .145
Charge Number
-.115
.103
0.89 0.73 – 1.09 .266
.031
.115
1.03 0.82 – 1.29 .786
Charge Severity
-.050
.055
0.95 0.85 – 1.06 .362
.016
.064
1.02 0.90 – 1.15 .797
Length of Stay
.003
.003
1.00 1.00 – 1.01 .367
-.003
.004
0.99 0.98 – 1.02 .401
Violent Offender (y/n)
-.019
.117
0.98 0.78 – 1.23 .870
-.034
.134
0.97 0.74 – 1.26 .801
Re-Detained/Incarcerated (y/n)
.277
.123
1.32 1.04 – 1.68 .025
.430
.146
1.54 1.16 – 2.05 .003
Risk Assessment Inventory
.002
.016
1.00 0.97 – 1.03 .918
.025
.017
1.03 0.99 – 1.06 .150
Stage 3: Mental Health Variables
MAYSI-2 Positive Screen (y/n)
.077
.117
1.08 0.86 – 1.36 .511
.004
.139
1.00 0.76 – 1.32 .980
Conduct Disorder (y/n)
-.207
.133
0.81 0.63 – 1.05 .118
.964
.210
2.62 1.74 – 3.96 .000
Non-Conduct Mental Disorder (y/n)
.146
.143
1.16 0.88 – 1.53 .306
.606
.202
1.83 1.23 – 2.72 .003
Substance-Related Disorder (y/n)
12.882 19.047 3.54 0.02 – 6.43 .499 12.138 20.986 8.92 0.06 – 12.39 .563
Number of Disorders
.057
.014
1.06 1.03 – 1.09 .000
.100
.012
1.11 1.08 – 1.13 .000
Pre-Detention Outpatient (y/n)
-.132
.134
0.88 0.67 – 1.14 .324
-.353
.157
0.70 0.52 – 0.96 .025
Pre-Detention Non-Outpatient
-.504
.162
0.60 0.44 – 0.83 .002
-.438
.168
0.66 0.46 – 0.90 .009
Treatment (y/n)
Note. N = Number; Beta = Unstandardized beta weight; SE = Standard error; HR = Hazard ratio; CI = Confidence interval; p =
Significance level; MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition.

First Substance-Related Treatment Utilization

Treatment Outcomes (N = 2341)

Table 3.69 Hazard Ratios and Associated Statistics for Cox Regression Models Predicting Time to First Substance-Related
Treatment and First Either/Both Treatment: Cohort Two

240

240

Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
-.094 .087 0.91 0.77 – 1.08 .283 -.502 .201
0.61 0.41 – 0.90 .012
Black (y/n)
-.153 .073 0.86 0.74 – 0.99 .036 -.868 .185
0.42 0.29 – 0.60 .000
Age
-.129 .026 0.88 0.84 – 0.92 .000 -.254 .071
0.78 0.68 – 0.89 .000
Stage 2: Criminal History Variables
Number of Prior Arrests
.013 .008 1.01 0.99 – 1.03 .110
.066 .018
1.07 1.03 – 1.10 .000
Charge Number
-.087 .060 0.92 0.82 – 1.03 .145 -.080 .160
0.92 0.68 – 1.26 .619
Charge Severity
-.009 .034 0.99 0.93 – 1.06 .784 -.151 .091
0.86 0.72 – 1.03 .099
Length of Stay
.001 .002 1.00 0.99 – 1.01 .530
.004 .005
1.00 0.99 – 1.01 .457
Violent Offender (y/n)
.098 .069 1.10 0.96 – 1.26 .155
.097 .188
1.10 0.76 – 1.60 .604
Re-Detained/Incarcerated (y/n)
.262 .078 1.30 1.12 – 1.52 .001
.205 .223
1.23 0.79 – 1.90 .358
Risk Assessment Inventory Scores
.000 .009 1.00 0.98 – 1.02 .998 -.019 .023
0.98 0.94 – 1.03 .405
Stage 3: Mental Health Variables
MAYSI-2 Positive Screen (y/n)
.004 .072 1.00 0.87 – 1.16 .955
.519 .243
1.68 1.04 – 2.70 .033
Conduct Disorder (y/n)
.671 .089 1.96 1.64 – 2.33 .000
.398 .268
1.49 0.88 – 2.52 .138
Non-Conduct Mental Disorder (y/n) .728 .105 2.07 1.69 – 2.54 .000 1.355 .395
3.88 1.79 – 8.41 .001
Substance-Related Disorder (y/n)
.711 .072 2.03 1.77 – 2.35 .000 1.172 .221
3.23 2.09 – 4.98 .000
Number of Disorders
.062 .008 1.06 1.05 – 1.08 .000
.105 .014
1.11 1.08 – 1.14 .000
Pre-Detention Outpatient (y/n)
-.303 .091 0.74 0.63 – 0.88 .001 -.484 .267
0.62 0.37 – 1.04 .069
Pre-Detention Non-Outpatient
-.049 .093 0.95 0.79 – 1.14 .602 -.074 .234
0.93 0.59 – 1.50 .751
Treatment (y/n)
Note. N = Number; Beta = Unstandardized beta weight; SE = Standard error; HR = Hazard ratio; CI = Confidence interval; p =
Significance level; MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition.

Treatment Outcomes (N = 2341)
First Outpatient Treatment Utilization
First Non-Outpatient Treatment Utilization
Beta
SE
HR
95% CI
p
Beta
SE
HR
95% CI
p

Table 3.70 Hazard Ratios and Associated Statistics for Cox Regression Models Predicting Time to First Outpatient Treatment
and First Non-Outpatient Treatment: Cohort Two

241

241

Participants
Survivingc

Probability of
Terminationb

Time Interval
Number
Number
Proportion Proportion (SD) Probability (SD)
0 – 3.0 Months
6234
1656
.27
.73 (.01)
.089 (.002)
3.01 – 6.0 Months
4578
518
.11
.65 (.01)
.028 (.001)
6.01 – 9.0 Months
4060
373
.09
.59 (.01)
.020 (.001)
9.01 – 12.0 Months
3687
297
.08
.54 (.01)
.016 (.001)
12.01 – 15.0 Months
3390
260
.08
.50 (.01)
.014 (.001)
15.01 – 18.0 Months
3130
222
.07
.47 (.01)
.012 (.001)
18.01 – 21.0 Months
2908
204
.07
.43 (.01)
.011 (.001)
21.01 – 24.0 Months
2704
184
.07
.40 (.01)
.010 (.001)
2.01 – 2.5 Years
2520
164
.07
.38 (.01)
.009 (.00)
2.51 – 3.0 Years
2213
270
.12
.31 (.01)
.007 (.00)
3.01 – 3.5 Years
1943
216
.11
.28 (.01)
.006 (.00)
3.51 – 4.0 Years
1727
216
.13
.24 (.01)
.006 (.00)
4.01 – 4.5 Years
1511
201
.13
.21 (.01)
.005 (.00)
4.51 – 5.0 Years
1310
163
.12
.18 (.00)
.004 (.00)
5.01 – 5.5 Years
1147
142
.12
.16 (.00)
.004 (.00)
5.51 – 6.0 Years
1005
146
.15
.14 (.00)
.004 (.00)
6.01 – 6.5 Years
859
160
.19
.11 (.00)
.004 (.00)
6.51 – 7.0 Years
699
129
.18
.09 (.00)
.003 (.00)
7.01 – 7.5 Years
570
134
.24
.07 (.00)
.004 (.00)
7.51 – 8.0 Years
436
131
.30
.05 (.00)
.004 (.00)
8.01 – 8.5 Years
305
143
.47
.03 (.00)
.004 (.00)
8.51 – 9.0 Years
162
162
1.00
.00 (.00)
.004 (.00)
Note. SD = Standard deviation. a At start of time interval. b During time interval; c At end of time interval.

Number of
Number of Termination
Participantsa Terminatorsc
Rateb

Table 3.71 Life Table: Time to Termination from Outpatient Treatment

Rate (SD)
.10 (.00)
.04 (.00)
.03 (.00)
.03 (.00)
.03 (.00)
.02 (.00)
.02 (.00)
.02 (.00)
.02 (.00)
.02 (.00)
.02 (.00)
.02 (.00)
.02 (.00)
.02 (.00)
.03 (.00)
.03 (.00)
.03 (.00)
.03 (.00)
.04 (.00)
.06 (.00)
.10 (.00)
.33 (.00)

Hazard Rateb

242

242

MAYSI-2
Positive
Screen

Age Cohorts

Race

Gender

Treatment
Type

Total Sample

Variables

Categories/
Groups
Any Treatment
N = 9664
Mental Health
SubstanceRelated
Either/Both
Male
n = 6062
Female
n = 3602
White
n = 3367
Black
n = 3464
Younger
n = 2406
Mid-Age
n = 3924
Older
n = 3334
No MAYSI-2
n = 7132
Non-positive
n = 1758
Positive Screen
n = 774
639 (26.5%)

1767 (73.5%)a

543 (30.9%)
309 (39.9%)

1215 (19.5%)b
465 (60.1%)c

2578 (36.1%)

2303 (40.6%)

3364 (59.4%)

4554 (63.9%)a

914 (27.1%)

2453 (72.9%)

1480 (44.4%)

1090 (30.2%)

2512 (69.8%)

1854 (55.6%)c

2340 (38.6%)

3722 (61.4%)

1311 (33.4%)

--

574 (9.2%)c

2613 (66.6%)b

--

--

4399 (70.6%)a
1261(20.2%)b

3430 (35.5%)

Number (%)

Non-Users
(Censored
Cases)

6234 (64.5%)

Number (%)

Outpatient
Users

398.00 (18.71)

244.00 (21.56)

552.00 (23.68)

275.00 (16.92)

438.00 (21.01)

783.00 (36.96)

369.00 (17.20)

639.00 (32.15)

746.00 (37.79)

356.99 (14.13)

758.00 (84.68)

273.00 (23.86)

496.00 (18.34)

459.00 (15.25)

Median (SD)

361.34 – 434.66

201.74 – 286.26

505.59 – 598.41

241.83 – 308.17

396.82 – 479.18

710.56 – 855.44

335.30 – 402.70

575.98 – 702.02

671.90 – 820.07

328.30 – 383.71

592.03 – 923.97

226.24 – 319.76

460.05 – 531.94

429.11 – 488.89

95% CI

Time to Termination (in Days)

84.88,
p < .001

151.75,
p < .001

70.25,
p < .001

135.66,
p < .001

96.18,
p < .001

--

154.39,
p < .001

128.77,
p < .001

72.37,
p < .001

165.85,
p < .001

97.40,
p < .001

--

255.16,
p < .001

71.99,
p < .001

54.44,
p < .001

144.72,
p < .001

89.23,
p < .001

--

Survival Curve Indicators
Breslow/
TaroneLog Rank
Wilcoxon
Ware
χ 2,
χ2,
χ 2,
p-value
p-value
p-value

Table 3.72 Time to Termination from Outpatient Treatment: Demographic, Mental Health, and Criminal History Groups

243

243

Mental Disorder
3023 (84.7%)a
548 (15.3%)
287.00 (15.14)
257.32 – 316.68
n = 3571
Disorder
Substance-Related
538 (81.3%)b
124 (3.6%)
56.00 (4.80)
46.60 – 65.41
Type
n = 662
1197.91,
1185.34,
1044.54,
Comorbid
p < .001
p < .001
p < .001
2673 (96.2%)c
109 (3.8%)
943.00 (30.75)
882.73 – 1003.27
n = 2782
No treatment
4399 (60.9%)
2827 (39.1%)
431.00 (2163)
388.60 – 473.40
Pren = 7226
4.72,
11.32,
31.85,
Detention
p < .001
p < .001
p < .001
Treatment
Treatment
1835 (75.3%)
603 (24.7%)
503.00 (19.10)
465.56 – 540.44
n = 2438
No
3872 (63.2%)
2250 (36.8%)
431.00 (18.51)
394.72 – 467.28
n = 6122
Violent
12.85,
17.34,
20.77,
Offender
p < .001
p < .001
p < .001
Yes
2362 (66.7%)
1180 (33.3%)
519.00 (28.88)
462.41 – 575.59
n = 3542
No
1821 (55.8%)
1442 (44.2%)
342.00 (19.98)
302.85 – 381.15
n = 3263
42.11,
34.84,
22.70,
Recidivist
p < .001
p < .001
p < .001
Yes
4413 (69.0%)
1986 (31.0%)
544.00 (19.10)
506.56 – 581.44
n = 6401
No
2439 (56.0%)
1919 (44.0%)
343.00 (19.99)
303.81 – 382.19
Re-Detained/ n = 4358
37.74,
31.90,
21.08,
Incarcerated
p < .001
p < .001
p < .001
Yes
3795 (71.5%)
1511 (28.5%)
532.00 (18.85)
505.05 – 578.95
n = 5306
Note. n = Number; % = Percentage; χ2 = Chi-square test; p-value = Significance level; MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition.
Different superscripts indicate significant differences (p ≤ .01) and matching superscripts indicate non-significant differences (p > .01) based upon 2 x 2 chisquare tests between paired groups.

Table 3.72 Continued.

244

244

245
Table 3.73 Model Summary Statistics for Cox Regression Models Predicting Time to
Termination from Outpatient Treatment
Model Summary Statistics (N = 5791)
Treatment
-2 LL
χ2
df
p
Outcomes
Termination from Outpatient Treatment
Stage 1
85491.09
352.67
3
.000
Stage 2

85363.46

507.83

9

.000

Change Statistics*
χ2

df

p

353.28

3

.000

127.63

6

.000

Stage 3
84518.45
1303.67
15
.000
845.01
6
.000
Termination from Mental Health Outpatient Treatment
Stage 1
57535.00
204.17
3
.000
204.47
3
.000
Stage 2
57438.39
318.43
9
.000
96.68
6
.000
Stage 3
56760.72
1056.42
15
.000
677.57
6
.000
Termination from Substance-Related Outpatient Treatment
Stage 1
13594.70
154.62
3
.000
156.29
3
.000
Stage 2
13569.51
180.62
9
.000
25.19
6
.000
Stage 3
13489.14
244.58
14
.000
80.36
5
.000
Termination from Either/Both Outpatient Treatment
Stage 1
5458.00
27.83
3
.000
27.58
3
.000
Stage 2
5429.33
61.27
9
.000
28.77
6
.000
Stage 3
5287.74
216.68
15
.000
141.98
6
.000
2
Note. N = Number; -2 LL = -2 Log Likelihood; df = Degrees of freedom; χ = Chi-square
test; p = Significance level. * Results based on the addition of predictors for the individual
Stage, compared to previous Stage. Stage 1 includes the following demographic predictor
variables: gender, race, and age. Stage 2 includes the following criminal history predictor
variables: number of prior arrests, number of charges at detention entry, charge severity,
length of stay, and violent offender (y/n). Stage 3 includes the following mental health
predictor variables: conduct disorder (y/n), non-conduct mental disorder (y/n), substancerelated disorder (y/n), number of disorders, pre-detention outpatient treatment (y/n), and
pre-detention non-outpatient treatment (y/n).

Beta
Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
.295 .029
1.34
1.27 – 1.42 .000
.226 .035 1.25 1.17 – 1.34
Black (y/n)
.034 .028
1.04
0.98 – 1.09 .222
.050 .033 1.05 0.99 – 1.12
Age
.063 .009
1.06
1.04 – 1.08 .000
.045 .010 1.05 1.03 – 1.07
Stage 2: Criminal History Variables
Number of Prior Arrests
.040 .005
1.04
1.03 – 1.05 .000
.044 .006 1.05 1.03 – 1.06
Charge Number
-.006 .008
0.99
0.98 – 1.01 .445
-.011 .010 0.99 0.96 – 1.01
Charge Severity
.041 .014
1.04
1.01 – 1.07 .003
.028 .017 1.03 0.99 – 1.06
Length of Stay
.000 .001
1.00
0.99 – 1.00 .704
.000 .001 1.00 0.99 – 1.00
Violent Offender (y/n)
-.061 .029
0.94
0.89 – 0.99 .034
-.061 .034 0.94 0.88 – 1.01
Re-Detained/Incarcerated (y/n)
.037 .029
1.04
0.98 – 1.20 .207
.064 .034 1.07 0.99 – 1.14
Stage 3: Mental Health Variables
Conduct Disorder (y/n)
.177 .034
1.19
1.12 – 1.28 .000
.139 .043 1.15 1.06 – 1.25
Non-Conduct Mental Disorder (y/n)
-.520 .041
0.59
0.55 – 0.65 .000
-.878 .064 0.42 0.37 – 0.47
Substance-Related Disorder (y/n)
-.107 .031
0.90
0.85 – 0.95 .001
-.063 .039 0.96 0.87 – 1.10
Number of Disorders
-.060 .004
0.94
0.91 – 0.97 .000
-.063 .004 0.94 0.93 – 0.96
Pre-Detention Outpatient Treatment (y/n) .414 .038
1.51
1.40 – 1.63 .000
.460 .045 1.58 1.45 – 1.73
Pre-Detention Non-Outpatient
.232 .055
1.26
1.13 – 1.40 .000
.281 .064 1.32 1.17 – 1.50
Treatment (y/n)
Note. N = Number; Beta = Unstandardized beta weight; SE = Standard error; HR = Hazard ratio; CI = Confidence interval; p =
Significance level.

.000

.001
.000
.111
.000
.000

.000
.277
.097
.906
.071
.061

.000
.135
.000

Treatment Outcomes (N = 5791)
Termination from
Termination from Mental Health
Outpatient Treatment
Outpatient Treatment
SE
HR
95% CI
p
Beta
SE
HR
95% CI
p

Table 3.74 Hazard Ratios and Associated Statistics for Cox Regression Models Predicting Time to Termination from Outpatient
Treatment and from Mental Health Outpatient Treatment

246

246

p
Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
.546
.068 1.73
1.51 – 1.97 .000
.151 .105
1.16
0.95 – 1.42 .150
Black (y/n)
.068
.065 1.07
0.94 – 1.22 .301
-.057 .095
0.95
0.78 – 1.14 .550
Age
.111
.020 1.12
1.08 – 1.16 .000
.122 .030
1.13
1.07 – 1.20 .000
Stage 2: Criminal History Variables
Number of Prior Arrests
.034
.009 1.04
1.03 – 1.05 .000
.033 .014
1.03
1.01 – 1.06 .020
Charge Number
.146
.063 1.16
1.02 – 1.31 .021
.070 .092
1.07
0.90 – 1.28 .447
Charge Severity
.078
.031 1.08
1.02 – 1.15 .012
.081 .049
1.08
0.99 – 1.20 .096
Length of Stay
.001
.002 1.01
0.99 – 1.01 .785
-.004 .004
0.99
0.99 – 1.01 .341
Violent Offender (y/n)
-.020
.066 0.98
0.86 – 1.12 .759
-.003 .099
1.00
0.82 – 1.21 .978
Re-Detained/Incarcerated (y/n)
.051
.067 1.05
0.92 – 1.20 .447
-.132 .098
0.88
0.72 – 1.06 .178
Stage 3: Mental Health Variables
Conduct Disorder (y/n)
-.072
.081 0.93
0.79 – 1.09 .377
.169 .119
1.18
0.94 – 1.49 .155
Non-Conduct Mental Disorder (y/n) -.148
.081 0.86
0.74 – 1.01 .068
-.930 .176
0.40
0.28 – 0.55 .000
Substance-Related Disorder (y/n)
-.058
.011 0.94
0.92 – 0.96 .000
1.493 1.014 4.45 0.61 – 12.51 .141
Number of Disorders
-.058
.011 0.94
0.91 – 0.96 .000
-.067 .009
0.94
0.93 – 0.96 .000
Pre-Detention Outpatient
.293
.096 1.34
1.11 – 1.62 .002
.636 .127
1.89
1.47 – 2.42 .000
Treatment (y/n)
Pre-Detention Non-Outpatient
.118
.144 1.13
0.85 – 1.49 .414
.139 .169
1.15
0.83 – 1.60 .410
Treatment (y/n)
Note. N = Number; Beta = Unstandardized beta weight; SE = Standard error; HR = Hazard ratio; CI = Confidence interval; p =
Significance level.

Treatment Outcomes (N = 5791)
Termination from Substance-Related
Termination from Either/Both
Outpatient Treatment
Outpatient Treatment
Beta
SE
HR
95% CI
p
Beta
SE
HR
95% CI

Table 3.75 Hazard Ratios and Associated Statistics for Cox Regression Models Predicting Time to Termination from SubstanceRelated Outpatient Treatment and from Either/Both Outpatient Treatment

247

247

248
Table 3.76 Model Summary Statistics for Cox Regression Models Predicting Time to
Termination from Outpatient Treatment: Cohort One
Model Summary Statistics (N = 4484)
Change Statistics*
Treatment
-2 LL
χ2
df
p
χ2
df
p
Outcomes
Termination from Outpatient Treatment
Stage 1
59379.69
119.50
3
.000
119.04
3
.000
Stage 2
59340.99
159.41
9
.000
38.69
6
.000
Stage 3
58818.68
664.30
15
.000
522.38
6
.000
Termination from Mental Health Outpatient Treatment
Stage 1
40506.09
77.50
3
.000
77.02
3
.000
Stage 2
40484.73
99.45
9
.000
21.32
6
.002
Stage 3
40027.27
642.83
15
.000
457.62
6
.000
Termination from Substance-Related Outpatient Treatment
Stage 1
9475.39
72.63
3
.000
72.27
3
.000
Stage 2
9455.30
92.08
9
.000
20.04
6
.003
Stage 3
9406.62
132.01
14
.000
48.63
5
.000
Termination from Either/Both Outpatient Treatment
Stage 1
3135.98
4.90
3
.175
4.90
3
.176
Stage 2
3126.26
14.59
9
.104
9.71
6
.137
Stage 3
3061.33
86.45
14
.000
64.25
5
.000
2
Note. N = Number; -2 LL = -2 Log Likelihood; df = Degrees of freedom; χ = Chi-square
test; p = Significance level. * Results based on the addition of predictors for the individual
Stage, compared to previous Stage. Stage 1 includes the following demographic predictor
variables: gender, race, and age. Stage 2 includes the following criminal history predictor
variables: number of prior arrests, number of charges at detention entry, charge severity,
length of stay, and violent offender (y/n). Stage 3 includes the following mental health
predictor variables: conduct disorder (y/n), non-conduct mental disorder (y/n), substancerelated disorder (y/n), number of disorders, pre-detention outpatient treatment (y/n), and predetention non-outpatient treatment (y/n).

–
+
+

–
+

NA NA
+
+
–
–

+

+

NA
+
–
–
–
+

+

+

+

NA NA

–
–
–
+
+

+

+

+

2c

NA

–

+

+

1b

–
-–
+
+

--

+

+

+

NA
–
–
–
–
+

NA

+
+
+

+

+

–
–

NA

NA

+

+

+

–
–

+

+

–

+

–
+

–

–

–

NA NA

NA NA

+

+

–

–

+

+

Note. NA = Not applicable because predictor excluded from analysis; Total = Total Sample; 1 = Cohort One; 2 = Cohort Two; (+) = Increased Risk; (–) =
Decreased Risk; RAI = Risk Assessment Inventory; MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition. a N = 5791. b N = 4484. c N = 1307.

Predictor Variables
Totala
Demographic Variables
Male (y/n)
+
Black (y/n)
Age
+
Criminal History Variables
Number of Prior Arrests
+
Charge Number
Charge Severity
+
Length of Stay
Violent Offender (y/n)
–
Re-Detained/Incarcerated (y/n)
RAI Scores
NA
Mental Health Variables
MAYSI-2 – Positive Screen (y/n)
NA
Conduct Disorder (y/n)
+
Non-Conduct Mental Disorder (y/n)
–
Substance-Related Disorder (y/n)
–
Number of Disorders
–
Pre-Detention Outpatient Treatment (y/n)
+
Pre-Detention Non-Outpatient Treatment (y/n) +

Any Outpatient

Termination from Outpatient Treatment
Mental Health
Substance-Related
Either/Both
Outpatient
Outpatient
Outpatient
Total 1
2
Total
1
2
Total 1
2

Table 3.77 Predictors in Cox Regression Analyses Associated with Increased Risk (+) and Decreased Risk (–) of Time to
Termination from Outpatient Treatment: Total Sample, Cohort One, and Cohort Two

249

249

Treatment Outcomes (N = 4484)
Termination from Mental Health
Termination from Outpatient Treatment
Outpatient Treatment
Beta
SE
HR
95% CI
p
Beta
SE
HR
95% CI

Hazard Ratios and Associated Statistics for Cox Regression Models Predicting Time to Termination from Outpatient
Treatment and from Mental Health Outpatient Treatment: Cohort One

p
Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
.238
.034 1.27
1.19 – 1.36 .000
.170
.041
1.19 1.09 – 1.29 .000
Black (y/n)
-.027
.032 0.97
0.91 – 1.04 .403
-.008
.038
0.99 0.92 – 1.07 .836
Age
.009
.010 1.01
0.99 – 1.03 .346
-.017
.012
0.98 0.96 – 1.01 .156
Stage 2: Criminal History Variables
Number of Prior Arrests
.037
.009 1.04
1.02 – 1.06 .000
.024
.010
1.35 1.01 – 1.55 .017
Charge Number
-.009
.009 0.99
0.97 – 1.01 .338
-.013
.011
0.99 0.97 – 1.01 .246
Charge Severity
.016
.016 1.02
0.98 – 1.05 .318
-.001
.020
0.99 0.96 – 1.04 .973
Length of Stay
-.002
.001 0.99
0.99 – 1.01 .098
-.001
.001
1.00 0.99 – 1.00 .346
Violent Offender (y/n)
-.070
.033 0.93
0.87 – 0.99 .036
-.056
.039
0.95 0.88 – 1.02 .155
Re-Detained/Incarcerated (y/n)
.053
.034 1.06
0.90 – 1.13 .117
.068
.040
1.00 0.99 – 1.16 .089
Stage 3: Mental Health Variables
Conduct Disorder (y/n)
.147
.039 1.16
1.07 – 1.25 .000
.101
.049
1.11 1.01 – 1.22
.039
Non-Conduct Mental Disorder (y/n) -.478
.048 0.62
0.56 – 0.68 .000
-.914
.076
0.40 0.35 – 0.47
.000
Substance-Related Disorder (y/n)
-.090
.037 0.91
0.85 – 0.98 .015
.005
.048
1.01 0.92 – 1.10
.913
Number of Disorders
-.057
.004 0.95
0.94 – 0.96 .000
-.063
.005
0.94 0.93 – 0.96
.000
Pre-Detention Outpatient
.183
.055 1.20
1.08 – 1.34 .001
.227
.062
1.25 1.11 – 1.41
.000
Treatment (y/n)
Pre-Detention Non-Outpatient
.128
.096 1.14
0.94 – 1.37 .185
.125
.112
1.13 0.91 – 1.41
.265
Treatment (y/n)
Note. N = Number; Beta = Unstandardized beta weight; SE = Standard error; HR = Hazard ratio; CI = Confidence interval; p =
Significance level.

Table 3.78

250

250

.326

.444
.000
.881
.000
.270

.136
.722
.420
.007
.739
.469

.535
.632
.015

Treatment Outcomes (N = 4484)
Termination from Substance-Related
Termination from Either/Both
Outpatient Treatment
Outpatient Treatment
Beta SE
HR
95% CI
p
Beta
SE
HR
95% CI
p

Hazard Ratios and Associated Statistics for Cox Regression Models Predicting Time to Termination from SubstanceRelated Outpatient Treatment and from Either/Both Outpatient Treatment: Cohort One

Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
.468 .076 1.60 1.38 – 1.85 .000
.079 .127 1.08 0.84 – 1.29
Black (y/n)
.014 .076 1.01 0.87 – 1.18 .854
-.058 .120 0.94 0.75 – 1.29
Age
.065 .023 1.07 1.02 – 1.12 .004
.091 .037 1.10 1.02 – 1.18
Stage 2: Criminal History Variables
Number of Prior Arrests
.064 .019 1.07 1.03 – 1.11 .001
.051 .034 1.055 0.98 – 1.13
Charge Number
.139 .073 1.15 0.99 – 1.33 .056
.047 .131 1.05 0.81 – 1.35
Charge Severity
.067 .036 1.07 0.99 – 1.15 .063
.053 .065 1.05 0.93 – 1.20
Length of Stay
-.002 .003 0.99
0.96– 1.00 .358
-.014 .005 0.98 0.97 – 0.99
Violent Offender (y/n)
-.081 .079 0.92 0.79 – 1.08 .303
.041 .123 1.04 0.82 – 1.33
Re-Detained/Incarcerated (y/n)
.089 .078 1.09 0.94 – 1.27 .253
-.092 .128 0.91 0.71 – 1.17
Stage 3: Mental Health Variables
Conduct Disorder (y/n)
-.085 .097 0.92 0.76 – 1.11 .383
-.116 .151 0.89 0.66 – 1.20
Non-Conduct Mental Disorder (y/n)
-.126 .094 0.88 0.73 – 1.06 .178
-1.228 .281 0.29 0.17 – 0.51
Substance-Related Disorder (y/n)
-.105 .037 0.90 0.84 – 0.97 .007
.047 .023 1.007 0.92 – 1.11
Number of Disorders
-.050 .012 0.95 0.93 – 0.98 .000
-.051 .011 0.95 0.93 – 0.97
Pre-Detention Outpatient Treatment (y/n) .117 .137 1.13 0.86 – 1.47 .395
.266 .241 1.31 0.81 – 2.10
Pre-Detention Non-Outpatient Treatment
.110 .238 1.12 0.70 – 1.78 .643
.342 .348 1.41 0.71 – 2.78
(y/n)
Note. N = Number; Beta = Unstandardized beta weight; SE = Standard error; HR = Hazard ratio; CI = Confidence interval; p =
Significance level.

Table 3.79

251

251

252
Table 3.80 Model Summary Statistics for Cox Regression Models Predicting Time to
Termination from Outpatient Treatment: Cohort Two
Model Summary Statistics (N = 1307)
Treatment
-2 LL
χ2
df
p
Outcomes
Termination from Outpatient Treatment
Stage 1
16954.88
361.69
3
.000
Stage 2
16940.98
376.48
10
.000
Stage 3
16572.96
710.42
17
.000
Termination from Mental Health Outpatient Treatment
Stage 1
10783.55
205.82
3
.000
Stage 2
10777.94
211.89
10
.000
Stage 3
10556.31
411.29
17
.000
Termination from Substance-Related Outpatient Treatment
Stage 1
2510.60
99.70
3
.000
Stage 2
2503.55
104.73
10
.000
Stage 3
2444.63
154.39
16
.000
Termination from Either/Both Outpatient Treatment
Stage 1
1344.52
38.23
3
.175
Stage 2
1330.78
53.73
10
.000

Change Statistics*
χ2

df

p

361.90
13.86

3
7

.000
.053

368.23

7

.000

203.35
5.61
221.63

3
7
7

.000
.586
.000

104.30
7.05
58.91

3
7
6

.000
.423
.000

38.18
13.74

3
7

.176
.056

Stage 3
1279.61
97.26
16
.000
51.17
6
.000
2
Note. N = Number; -2 LL = -2 Log Likelihood; df = Degrees of freedom; χ = Chi-square
test; p = Significance level. * Results based on the addition of predictors for the
individual Stage, compared to previous Stage. Stage 1 includes the following
demographic predictor variables: gender, race, and age. Stage 2 includes the following
criminal history predictor variables: Risk Assessment Inventory scores, number of prior
arrests, number of charges at detention entry, charge severity, length of stay, and violent
offender (y/n). Stage 3 includes the following mental health predictor variables: Positive
screen on Massachusetts Youth Screening Instrument-2nd Edition, conduct disorder (y/n),
non-conduct mental disorder (y/n), substance-related disorder (y/n), number of disorders,
pre-detention outpatient treatment (y/n), and pre-detention non-outpatient treatment (y/n).

Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
.122 .073 1.13 0.98 – 1.21 .095
.170 .084 1.19 1.01 – 1.40
Black (y/n)
.107 .062 1.11 0.98 – 1.26 .083
.119 .076 1.13 0.97 – 1.31
Age
.311 .019 1.37 1.31 – 1.42 .000
.267 .022 1.63 1.25 – 1.36
Stage 2: Criminal History Variables
Number of Prior Arrests
.016 .007 1.02 1.00 – 1.03 .020
.019 .009 1.02 1.00 – 1.04
Charge Number
.087 .055 1.09 0.98 – 1.21 .114
.059 .066 1.06 0.93 – 1.21
Charge Severity
.087 .029 1.09 1.03 – 1.16 .003
.053 .036 1.05 0.98 – 1.13
Length of Stay
.003 .002 1.00
0.99– 1.01 .057
.001 .002 1.00 0.99 – 1.01
Violent Offender (y/n)
-.049 .059 0.95 0.85 – 1.07 .402 -.043 .070 0.96 0.83 – 1.10
Re-Detained/Incarcerated (y/n)
-.011 .064 0.99 0.87 – 1.12 .859
.040 .074 1.04 0.89 – 1.20
Risk Assessment Inventory Scores
-.012 .008 0.98 0.97 – 1.00 .117 -.018 .009 0.98 0.97 – 0.99
Stage 3: Mental Health Variables
MAYSI-2 - Positive Screen (y/n)
-.070 .061 0.93 0.83 – 1.05 .255 -.036 .076 0.96 0.83 – 1.12
Conduct Disorder (y/n)
-.096 .073 0.91 0.79 – 1.05 .186
.080 .103 1.08 0.88 – 1.32
Non-Conduct Mental Disorder (y/n)
-.629 .087 0.53 0.45 – 0.63 .000 -.803 .132 0.45 0.35 – 0.58
Substance-Related Disorder (y/n)
-.384 .059 0.68 0.61 – 0.77 .000 -.403 .075 0.67 0.58 – 0.78
Number of Disorders
-.093 .008 0.91 0.90 – 0.93 .000 -.084 .010 0.92 0.90 – 0.94
Pre-Detention Outpatient (y/n)
.174 .072 1.19 1.03 – 1.27 .016
.312 .093 1.37 1.14 – 1.64
Pre-Detention Non-Outpatient Treatment (y/n) .272 .077 1.31 1.13 – 1.53 .000
.338 .091 1.43 1.18 – 1.68
Note. N = Number; Beta = Unstandardized beta weight; SE = Standard error; HR = Hazard ratio; CI = Confidence interval; p =
Significance level; MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition.

.635
.440
.000
.000
.000
.001
.000

.034
.371
.143
.529
.537
.595
.045

.044
.116
.000

Treatment Outcomes (N = 1307)
Termination from Mental Health
Termination from Outpatient Treatment
Outpatient Treatment
Beta SE
HR
95% CI
p
Beta SE
HR
95% CI
p

Table 3.81 Hazard Ratios and Associated Statistics for Cox Regression Models Predicting Time to Termination from Outpatient
Treatment and from Mental Health Outpatient Treatment: Cohort Two

253

253

Predictor Variables
Stage 1: Demographic Variables
Male (y/n)
.442 .206 1.56
1.04 – 2.33 .032
.264 .267 1.30 0.77 – 2.20
Black (y/n)
-.039 .145 0.96
0.72 – 1.28 .789
-.373 .200 0.69 0.47 – 1.02
Age
.388 .049 1.47
1.34 – 1.62 .000
.415 .067 1.52 1.33 – 1.73
Stage 2: Criminal History Variables
Number of Prior Arrests
.014 .014 1.01
0.99 – 1.04 .346
-.029 .021 0.97 0.93 – 1.01
Charge Number
.106 .139 1.11
0.85 – 1.46 .444
-.118 .146 0.88 0.68 – 1.18
Charge Severity
.091 .069 1.01
0.96 – 1.26 .189
.148 .088 1.16 0.98 – 1.38
Length of Stay
.008 .004 1.01
0.99 – 1.02 .079
.012 .005 1.02 1.01 – 1.03
Violent Offender (y/n)
-.011 .141 0.98
0.75 – 1.30 .938
-.160 .217 0.85 0.56 – 1.31
Re-Detained/Incarcerated (y/n)
.028 .144 1.03
0.78 – 1.37 .845
.178 .185 1.19 0.83 – 1.72
Risk Assessment Inventory Scores
-.018 .022 0.98
0.94 – 1.03 .409
.024 .023 1.02 0.98 – 1.07
Stage 3: Mental Health Variables
MAYSI-2 Positive Screen (y/n)
.176 .143 1.19
0.90 – 1.58 .219
-.104 .186 0.90 0.63 – 1.30
Conduct Disorder (y/n)
-.301 .161 0.74
0.54 – 1.02 .062
-.044 .301 0.96 0.53 – 1.73
Non-Conduct Mental Disorder (y/n)
-.188 .178 0.83
0.58 – 1.17 .290
-.654 .255 0.52 0.32 – 0.86
Substance-Related Disorder (y/n)
.392 2.00 0.78
0.44 – 0.93 .001
-.256 1.25 1.50 0.04 – 6.61
Number of Disorders
-.116 .027 0.89
0.85 – 0.94 .000
-.105 .025 0.90 0.86 – 0.94
Pre-Detention Outpatient (y/n)
.050 .172 1.05
0.75 – 1.47 .773
.387 .201 1.47 0.99 – 2.19
Pre-Detention Non-Outpatient Treatment (y/n) .261 .227 1.30
0.83 – 2.02 .250
-.188 .236 0.83 0.52 – 1.32
Note. N = Number; Beta = Unstandardized beta weight; SE = Standard error; HR = Hazard ratio; CI = Confidence interval; p =
Significance level; MAYSI-2 = Massachusetts Youth Screening Instrument-2nd Edition.

Treatment Outcomes (N = 1307)
Termination from Substance-Related
Termination from Either/Both
Outpatient Treatment
Outpatient Treatment
Beta SE
HR
95% CI
p
Beta
SE
HR
95% CI

Table 3.82 Hazard Ratios and Associated Statistics for Cox Regression Models Predicting Time to Termination from SubstanceRelated Outpatient Treatment and from Either/Both Outpatient Treatment: Cohort Two

254

.575
.883
.010
.595
.000
.054
.425

.156
.419
.093
.022
.463
.337
.313

.323
.061
.000

p

254

255

FIGURES

255

Figure 3.1

Survival Curve: Total Sample. Survival curve for time (in days) to first

treatment from detention release (i.e., day 0) to treatment utilization among all
participants (N = 9664). Cum Survival = Cumulative survival, or total proportion of
participants surviving without treatment (i.e., non-users). Survival function is the
survival curve for participants who utilized treatment. Censored refers to participants
who were lost due to attrition (i.e., death or missing data) or remained non-users until
the end of the study time frame.

256

Figure 3.2

Survival Curve: Treatment Type. Survival curves for time (in days) to

first treatment from detention release (i.e., day 0) to treatment utilization among
treatment users (N = 6437), divided by treatment type. Cum Survival = Cumulative
survival, or total proportion of participants surviving without treatment (i.e., non-users).
Survival functions refer to the survival curves for participants who utilized treatment.

257

Figure 3.3

Survival Curve: Treatment Setting. Survival curves for time (in days) to

first treatment from detention release (i.e., day 0) to treatment utilization among
treatment users (N = 6437), divided by treatment setting. Cum Survival = Cumulative
survival, or total proportion of participants surviving without treatment (i.e., non-users).
Survival functions refer to the survival curves for participants who utilized treatment.

258

Figure 3.4

Survival Curve: Gender. Survival curves for time (in days) to first

treatment from detention release (i.e., day 0) to treatment utilization among all
participants (N = 9664), divided by gender. Cum Survival = Cumulative survival, or
total proportion of participants surviving without treatment (i.e., non-users). Survival
functions refer to the survival curves for participants who utilized treatment. Censored
refers to participants who were lost due to attrition (i.e., death or missing data) or
remained non-users to the end of the study time frame.

259

Figure 3.5

Survival Curve: Race. Survival curves for time (in days) to treatment

from detention release (i.e., day 0) to treatment utilization among all participants (N =
9034), divided by racial groups. Cum Survival = Cumulative survival, or total proportion
of participants surviving without treatment (i.e., non-users). Survival functions refer to
the survival curves for participants who utilized treatment. Censored refers to
participants who were lost due to attrition (i.e., death or missing data) or remained nonusers until the end of the study time frame.

260

Figure 3.6

Survival Curve: Age Cohorts. Survival curves for time (in days) to first

treatment from detention release (i.e., day 0) to treatment utilization among all
participants (N = 9664), divided by age cohorts. Cum Survival = Cumulative survival,
or total proportion of participants surviving without treatment (i.e., non-users). Survival
functions refer to the survival curves for participants who utilized treatment. Censored
refers to participants who were lost due to attrition (i.e., death or missing data) or
remained non-users until the end of the study time frame.

261

Figure 3.7

Survival Curve: MAYSI-2. Survival curves for time (in days) to first

treatment from detention release (i.e., day 0) to treatment utilization among all
participants (N = 9664), divided by MAYSI-2 groups. MAYSI-2 = Massachusetts
Youth Screening Instrument-2nd Edition. Cum Survival = Cumulative survival, or total
proportion of participants surviving without treatment (i.e., non-users). Survival
functions refer to the survival curves for participants who utilized treatment. Censored
refers to participants who were lost due to attrition (i.e., death or missing data) or
remained non-users until the end of the study time frame.

262

Figure 3.8

Survival Curve: Disorder Type. Survival curves for time (in days) to first

treatment from detention release (i.e., day 0) to treatment utilization among participants
with disorders (N = 7015), divided by disorder type. Cum Survival = Cumulative
survival, or total proportion of participants surviving without treatment (i.e., non-users).
Survival functions refer to the survival curves for participants who utilized treatment.
Censored refers to participants who were lost due to attrition (i.e., death or missing data)
or remained non-users until the end of the study time frame.

263

Figure 3.9

Survival Curve: Pre-Detention Treatment. Survival curves for time (in

days) to first treatment from detention release (i.e., day 0) to treatment utilization among
all participants (N = 9664), divided by pre-detention treatment groups. Cum Survival =
Cumulative survival, or total proportion of participants surviving without treatment (i.e.,
non-users). Survival functions refer to the survival curves for participants who utilized
treatment. Censored refers to participants who were lost due to attrition (i.e., death or
missing data) or remained non-users until the end of the study time frame.

264

Figure 3.10

Survival Curve: Violent Offender. Survival curves for time (in days) to

first treatment from detention release (i.e., day 0) to treatment utilization among all
participants (N = 9664), divided by violent offender groups. Cum Survival =
Cumulative survival, or total proportion of participants surviving without treatment (i.e.,
non-users). Survival functions refer to the survival curves for participants who utilized
treatment. Censored refers to participants who were lost due to attrition (i.e., death or
missing data) or remained non-users until the end of the study time frame.

265

Figure 3.11

Survival Curve: Recidivism. Survival curves for time (in days) to first

treatment from detention release (i.e., day 0) to treatment utilization among all
participants (N = 9664), divided by recidivism groups. Cum Survival = Cumulative
survival, or total proportion of participants surviving without treatment (i.e., non-users).
Survival functions refer to the survival curves for participants who utilized treatment.
Censored refers to participants who were lost due to attrition (i.e., death or missing data)
or remained non-users until the end of the study time frame.

266

Figure 3.12

Survival Curve: Re-Detained. Survival curves for time (in days) to first

treatment from detention release (i.e., day 0) to treatment utilization among all
participants (N = 9664), divided by re-detained/incarcerated groups. Cum Survival =
Cumulative survival, or total proportion of participants surviving without treatment (i.e.,
non-users). Survival functions refer to the survival curves for participants who utilized
treatment. Censored refers to participants who were lost due to attrition (i.e., death or
missing data) or remained non-users until the end of the study time frame.

267

Figure 3.13

Survival Curve for Treatment Termination: Total Sample. Survival

curve for time (in days) until termination from outpatient treatment, from date of first
outpatient service to date of final outpatient service among total sample (N = 9664).
Cum Survival = Cumulative survival, or total proportion of participants surviving
without treatment (i.e., non-users). Survival functions refer to the survival curves for
participants who utilized treatment. Censored refers to participants who were lost due to
attrition (i.e., death or missing data) or remained non-users until the end of the study
time frame.

268

Figure 3.14

Survival Curve for Treatment Termination: Treatment Type. Survival

curves for time (in days) until termination from outpatient treatment, from date of first
outpatient service to date of final outpatient service among outpatient treatment users (N
= 6234). Survival curves are divided by treatment type. Cum Survival = Cumulative
survival, or total proportion of participants surviving without treatment (i.e., non-users).
Survival functions refer to the survival curves for participants who utilized treatment.

269

Figure 3.15

Survival Curve for Treatment Termination: Gender. Survival curves for

time (in days) until termination from outpatient treatment, from date of first outpatient
service to date of final outpatient service among total sample (N = 9664). Survival
curves are divided by gender. Cum Survival = Cumulative survival, or total proportion
of participants surviving without treatment (i.e., non-users). Survival functions refer to
the survival curves for participants who utilized treatment. Censored refers to
participants who were lost due to attrition (i.e., death or missing data) or remained nonusers until the end of the study time frame.

270

Figure 3.16

Survival Curve for Treatment Termination: Race. Survival curves for

time (in days) until termination from outpatient treatment, from date of first outpatient
service to date of final outpatient service among total sample (N = 9664). Survival
curves are divided by racial groups. Cum Survival = Cumulative survival, or total
proportion of participants surviving without treatment (i.e., non-users). Survival
functions refer to the survival curves for participants who utilized treatment. Censored
refers to participants who were lost due to attrition (i.e., death or missing data) or
remained non-users until the end of the study time frame.

271

Figure 3.17

Survival Curve for Treatment Termination: Age Cohorts. Survival

curves for time (in days) until termination from outpatient treatment, from date of first
outpatient service to date of final outpatient service among total sample (N = 9664).
Survival curves are divided by age cohorts. Cum Survival = Cumulative survival, or
total proportion of participants surviving without treatment (i.e., non-users). Survival
functions refer to the survival curves for participants who utilized treatment. Censored
refers to participants who were lost due to attrition (i.e., death or missing data) or
remained non-users until the end of the study time frame.

272

Figure 3.18

Survival Curve for Treatment Termination: MAYSI-2. Survival curves

for time (in days) until termination from outpatient treatment, from date of first outpatient
service to date of final outpatient service among total sample (N = 9664). MAYSI-2 =
Massachusetts Youth Screening Instrument-2nd Edition. Survival curves are divided by
MAYSI-2 groups. Cum Survival = Cumulative survival, or total proportion of
participants surviving without treatment (i.e., non-users). Survival functions refer to the
survival curves for participants who utilized treatment. Censored refers to participants
who were lost due to attrition (i.e., death or missing data) or remained non-users until the
end of the study time frame.

273

Figure 3.19

Survival Curve for Treatment Termination: Disorder Type. Survival

curves for time (in days) until termination from outpatient treatment, from date of first
outpatient service to date of final outpatient service among total sample (N = 9664).
Survival curves are divided by disorder type. Cum Survival = Cumulative survival, or
total proportion of participants surviving without treatment (i.e., non-users). Survival
functions refer to the survival curves for participants who utilized treatment. Censored
refers to participants who were lost due to attrition (i.e., death or missing data) or
remained non-users until the end of the study time frame.

274

Figure 3.20

Survival Curve for Treatment Termination: Prior Treatment. Survival

curves for time (in days) until termination from outpatient treatment, from date of first
outpatient service to date of final outpatient service among total sample (N = 9664).
Survival curves are divided by pre-detention treatment groups. Cum Survival =
Cumulative survival, or total proportion of participants surviving without treatment (i.e.,
non-users). Survival functions refer to the survival curves for participants who utilized
treatment. Censored refers to participants who were lost due to attrition (i.e., death or
missing data) or remained non-users until the end of the study time frame.

275

Figure 3.21

Survival Curve for Treatment Termination: Violent Offender. Survival

curves for time (in days) until termination from outpatient treatment, from date of first
outpatient service to date of final outpatient service among total sample (N = 9664).
Survival curves are divided by violent offender groups. Cum Survival = Cumulative
survival, or total proportion of participants surviving without treatment (i.e., non-users).
Survival functions refer to the survival curves for participants who utilized treatment.
Censored refers to participants who were lost due to attrition (i.e., death or missing data)
or remained non-users until the end of the study time frame.

276

Figure 3.22

Survival Curve for Treatment Termination: Recidivism. Survival

curves for time (in days) until termination from outpatient treatment, from date of first
outpatient service to date of final outpatient service among total sample (N = 9664).
Survival curves are divided by recidivism groups. Cum Survival = Cumulative survival,
or total proportion of participants surviving without treatment (i.e., non-users). Survival
functions refer to the survival curves for participants who utilized treatment. Censored
refers to participants who were lost due to attrition (i.e., death or missing data) or
remained non-users until the end of the study time frame.

277

Figure 3.23

Survival Curve for Treatment Termination: Re-Detained. Survival

curves for time (in days) until termination from outpatient treatment, from date of first
outpatient service to date of final outpatient service among total sample (N = 9664).
Survival curves are divided by violent offender groups. Cum Survival = Cumulative
survival, or total proportion of participants surviving without treatment (i.e., non-users).
Survival functions refer to the survival curves for participants who utilized treatment.
Censored refers to participants who were lost due to attrition (i.e., death or missing data)
or remained non-users until the end of the study time frame.

APPENDIX

278

APPENDIX

Table A.1. Codebook for Variables, Variable Type, and Levels

Variable

Variable Type

General
Follow-up time-points

Categorical

Time cohort

Categorical

Inclusion Criteria

Categorical

Include: 2-Year Followup
Demographics Domain
Date of Birth
Date of Death
Gender

Categorical

Race/Ethnicity

Categorical

Race

Categorical

Continuous
Continuous
Categorical

Definition/Levels
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
1.
2.
0.
1.
0.
1.

Pre-detention
During detention
Day of release from detention
1 month post detention
3 months post-detention
6 months post-detention
1 year post-detention
1.5 years post-detention
2 years post-detention
Anytime post-detention
Cohort one: 1998 – 2005
Cohort two: 2006 – 2011
Exclude, does not met criteria
Include, meets criteria
Exclude for 2-year follow-up analysis
Include for 2-year follow-up analysis

Month and year of birthdate
Month and year of death
0. Female
1. Male
0. White
1. Black
2. Hispanic
3. Other
0. White
1. Black

279
Table A.1 Continued.
Gender X Race

Categorical

Age Cohort

Categorical

Age
Insurance

Continuous
Categorical

Criminal History Domain
Year of Detention
Age at First Contact

Continuous
Continuous

Detention Entry
Detention Release

Continuous
Continuous

Length of Stay
Prior Referral

Continuous
Categorical

Prior Referrals

Continuous

Number of Charges
Charge Level

Continuous
Categorical

Charge Class

Categorical

Charge Severity

Categorical

1. White males
2. White females
3. Black males
4. Black females
5. Other males
6. Other females
1. Younger: Age 11-13
2. Mid-Age: Age 14-15
3. Older: Age 16-18
Actual age of youth at detention entry
1. Medicaid
2. Private insurance
3. No insurance/Self-pay
4. Unknown/Missing
Year of first detention
Age at first contact with juvenile justice
system
Date (day, month, year) of detention entry
Date (day, month, year) of detention
release
Time of detention stay (in days)
0. No
1. Yes, prior contact
Number of referrals/contacts with juvenile
justice system prior to first detention stay
Total number of charges for participant
1. Status offense/probation violation
2. Misdemeanor
3. Felony
1. A
2. B
3. C
4. D
5. S
1. Status offense/Probation violation
2. Non-violent misdemeanor
3. Violent misdemeanor
4. Non-violent felony
5. Violent felony

280
Table A.1 Continued
Charge Type

Categorical

Reason for Detention
Stay

Categorical

Reason for Detention
Release

Categorical

1. Severe (murder, manslaughter,
kidnapping)
2. Conduct-related (battery, assault,
recklessness, intimidation)
3. Weapons (possession, use, or no
license for weapons)
4. Property (theft, robbery, arson)
5. Drug/Alcohol (possession, dealing,
intoxication)
6. Sexual (rape, molestation, indecent
exposure, sexual battery)
7. Trespass/Loiter (trespass, conversion,
entry)
8. Driving-related (no license, reckless
driving, failure to stop)
9. Fraud (fraud, identity deception,
impersonation, false reporting)
10. Resist arrest (flee, escape, resist,
interfere with arrest)
11. Warrant arrest
12. Modification/Violation (probation,
runaway, curfew, home detention)
13. Other
14. Unknown/Missing
1. Awaiting action (pre-adjudication,
post-adjudication)
2. Court-ordered detention
3. Arrest (outright arrest, new arrest)
4. Warrant arrest/Failure to appear
5. Probation violation
6. Other
7. Missing/Unknown
1. Released to home detention
2. Community adjustment
3. Prison, county jail, other jurisdiction
4. Released to parent/guardian
5. Released to emergency shelter or
Department of Child Services
6. Released for unspecified placement
7. Released for treatment, hospital, or
diagnostic center
8. Other
9. Unknown/Missing

281
Table A.1 Continued
Risk Assessment
Inventory (RAI) Date
Risk Assessment
Inventory Score
Risk Assessment
Inventory: Risk Level

Continuous

Violent Offender

Categorical

Recidivism

Categorical

Recidivism: 2-Year
Follow-up

Categorical

Recidivism Timing

Categorical

Number of Recidivism
Events
Number of Recidivism
Events: 2-Year Followup
Recidivism Rate

Continuous

Recidivism before
Treatment

Categorical

Re-Detention

Categorical

Continuous
Categorical

Continuous

Continuous

Date (day, month, year) in which
participant completed RAI
Total score on Risk Assessment Inventory
1. Low-risk: Low-risk: 0-5 for males; 0-7
for females
2. Medium-risk: 6-12 for males; 8-13 for
females
3. High-risk: ≥13 for males; ≥14 for
females
0. No
1. Yes, at least one violent offense
0. No
1. Yes, re-arrest following release from
detention
0. No
1. Yes, re-arrest within two years of
detention release
0. No recidivism
1. Recidivism within 3 months
2. Recidivism within 4-6 months
3. Recidivism within 6 months-2 years
4. Recidivism after 2 years
Number of recidivism events following
release from detention
Number of recidivism events within two
years of detention release
Number of recidivism event divided by
total number of detentions
0. No, not re-arrested and/or utilized
treatment
1. No, recidivism after first treatment
service
2. Yes, recidivism before first treatment
service
0. No
1. Yes, placed in detention for a second
time following release from first
detention stay

282
Table A.1 Continued
Re-Detention: 2-Year
Follow-up

Categorical

Timing of 2nd Detention

Categorical

Re-Detained before
Treatment

Categorical

Number of Detentions

Continuous

Number of Detentions_
2 Years

Continuous

Incarceration in Prison

Categorical

Incarceration: 2-Year
Follow-up

Categorical

Number of
Incarcerations
Number of
Incarcerations: 2-Year
Follow-up
Re-detained/
Incarcerated

Continuous

Re-detained/
Incarcerated: 2-Year
Follow-up

Categorical

Continuous

Categorical

0. No
1. Yes, placed in detention for second
within two years of detention release
0. No 2nd Detention
1. 2nd detention stay within 1 day
2. 2nd detention stay within 1 month
3. 2nd detention stay within 2 months
4. 2nd detention stay within 3 months
5. 2nd detention stay within 4-6 months
6. 2nd detention stay within 7-12 months
7. 2nd detention Stay within 1-2 years
8. 2nd detention Stay after 2 years
0. No, not re-detained or no treatment
1. No, re-detained after first treatment
service
2. Yes, re-detained before first treatment
service
Number of detentions (excluding first
detention stay)
Number of detentions within two years of
detention release (excluding first detention
stay)
0. No
1. Yes, incarcerated in prison
0. No
1. Yes, incarcerated in prison within two
years of detention release
Number of prison stays during study time
frame
Number of prison stays within two years
of detention release
0. No
1. Yes, placed in second detention stay
and/or incarcerated in prison following
release from detention
0. No
1. Yes, placed in second detention stay
and/or incarcerated in prison within
two years post-detention

283
Table A.1 Continued
Timing of Incarceration
in Prison

1. No Incarceration in prison
2. Incarceration with 24 hours
3. Incarceration within 1 day-1 month
4. Incarceration within 1-2 months
5. Incarceration within 2-3 months
6. Incarceration within 4-6 months
7. Incarceration within 7-12 months
8. Incarceration within 1-2 years
9. Incarceration after 2 years
Mental Health and Substance-Related Needs/Diagnoses Domain
MAYSI-2 Score
Continuous
Total number of items endorsed
Alcohol/Drug Use
Continuous
Number of items endorsed within subscale
Subscale
(out of 8 items)
Angry/Irritable Subscale
Continuous
Number of items endorsed within subscale
(out of 9 items)
Depressed/Anxious
Continuous
Number of items endorsed within subscale
Subscale
(out of 9 items)
Somatic Complaints
Continuous
Number of items endorsed within subscale
Subscale
(out of 6 items)
Suicidal Ideation
Continuous
Number of items endorsed within subscale
Subscale
(out of 5 items)
Traumatic Experiences
Continuous
Number of items endorsed within subscale
Subscale
(out of 5 items)
Thought Disturbances
Continuous
Number of items endorsed within subscale
Subscale
(out of 5 items)
Caution or Warning on
Categorical
0. No
Alcohol/Drug Use
1. Yes, score falls within caution or
Subscale
warning range
Caution or Warning on
Categorical
0. No
Angry/Irritable Subscale
1. Yes, score falls within caution or
warning range
Caution or Warning on
Categorical
0. No
Depressed/Anxious
1. Yes, score falls within caution or
Subscale
warning range
Caution or Warning on
Categorical
0. No
Somatic Complaints
1. Yes, score falls within caution or
Subscale
warning range
Caution or Warning on
Categorical
0. No
Suicidal Ideation
1. Yes, score falls within caution or
Subscale
warning range

284
Table A.1 Continued
Caution or Warning on
Traumatic Experiences
Subscale
Caution or Warning on
Thought Disturbances
Subscale
Caution Subscales

Categorical

Warning Subscales

Continuous

MAYSI-2 - Positive
Screen

Categorical

MAYSI-2 Date

Continuous

MAYSI-2 Length
Mental Disorders
Disorder Prevalence

Continuous
Categorical
Continuous

Mental Disorder
Prevalence

Continuous

Substance-Related
Prevalence
Comorbidity Prevalence

Continuous

Date of Diagnosis

Continuous

Timing of Diagnosis

Categorical

Disorder (Y/N)

Categorical

Mental Disorder (Y/N)

Categorical

Conduct Disorder (Y/N)

Categorical

Non-Conduct Mental
Disorder (Y/N)

Categorical

Categorical

Continuous

Continuous

0. No
1. Yes, score falls within caution or
warning range
0. No
1. Yes, score falls within caution or
warning range
Number of subscales in caution range
(out of 7 subscales)
Number of subscales in warning range
(out of 7 subscales)
0. No
1. Yes, positive screen (i.e., warning/
caution on Suicidal Ideation and/or
warning/caution on ≥2 subscales)
Date (day, month, year) in which
participant completed MAYSI-2
Minutes to complete MAYSI-2
ICD-9 codes for different disorders
Percent of sample with any diagnosis
(ICD-9 codes)
Percent of sample with mental disorder
codes (excluding substance-related
disorders)
Percent of Sample with substance-related
codes (excluding mental disorders)
Percent of sample meeting criteria for both
types of disorders
Month and year of each disorder (i.e., date
in which ICD-9 code first appears in
medical records)
1. Pre-detention
2. During detention
3. Post-detention
0. No disorder
1. Disorder
0. No disorder
1. Mental disorder
0. No
1. Yes, conduct disorder
0. No
1. Yes, mental disorder (other than
conduct disorder)

285
Table A.1 Continued
Substance-Related
Disorder (Y/N)
Comorbid for both
Disorders (Y/N)
Disorder Type

Categorical

Number of Disorders
Disorder Intensity

Continuous
Categorical

Number of Mental
Disorders
Mental Disorder
Intensity

Continuous

Number of SubstanceRelated Disorders
Substance-Related
Disorder Intensity

Continuous

Mental Disorder Type

Categorical

Substance-Related Type

Categorical

Categorical
Categorical

Categorical

Categorical

0. No Disorder
1. Substance-related disorder
0. No
1. Comorbid
1. Only mental health disorder(s)
2. Only substance-related disorder(s)
3. Comorbid for both disorders
Total number of disorders for participant
1. 1 disorder
2. 2-3 disorders
3. 4-5 disorders
4. 6-7 disorders
5. ≥8 disorders
Total number of mental disorders for
participant
0. No disorders
1. 1 disorder
2. 2-3 disorders
3. 4-5 disorders
4. ≥6 disorders
Total number of substance-related
disorders for participant
0. No disorders
1. 1 disorder
2. 2-3 disorders
3. ≥4 disorders
1. Behavior-related disorder (conduct
disorder, oppositional defiant disorder)
2. Attention deficit-hyperactivity
disorder
3. Mood disorder (depression, bipolar)
4. Anxiety disorder (generalized,
obsessive-compulsive disorder)
5. Adjustment or somatization disorder
6. Psychosis-related disorder
7. Other disorder (eating, sleep, feeding,
personality)
1. Cannabis-related (abuse, dependence)
2. Alcohol-related (abuse, dependence)
3. Drug-related (abuse, dependence)
4. Other (using during pregnancy, fetal
alcohol)

286
Table A.1 Continued
Treatment Services Domain
Treatment Date
Continuous
Treatment Timing

Categorical

Treatment Type

Categorical

Treatment Timing in
Days
Number of Treatment
Types

Continuous

Pre-Detention Treatment

Categorical

During Detention
Treatment

Categorical

Type of User

Categorical

Treatment Setting

Categorical

Treatment Setting Final
Number of Treatment
Settings

Categorical

Categorical

Categorical

Date (day, month, year) of treatment
utilization
1. Pre-detention
2. During detention
3. Post-detention
1. Mental health treatment
2. Substance-related treatment
3. Either/Both treatment
Days from detention release to treatment
utilization
0. No types/Non-user
1. One type
2. Two types
0. Three types
0. No treatment
1. Yes, at least one treatment service
prior to detention
0. No treatment
1. Yes, at least one treatment service
during first detention stay
0. Non-User
1. Only mental health
2. Only substance-related
3. Only either/both
4. Mental health and substance-related
5. Mental health and either/both
6. Substance-related and either/both
0. All three types
1. Outpatient
2. Inpatient stay
3. Emergency department visit
1. Outpatient
2. Non-outpatient
0. Non-user
1. One setting
2. Two settings
3. Three settings

287
Table A.1 Continued
Treatment User Setting

Categorical

Outpatient Service
Intensity

Categorical

Non-Outpatient Service
Intensity

Categorical

Dropout from
Outpatient Treatment

Categorical

Reason for Dropout

Categorical

Dropout Session

Continuous

Outpatient Treatment
Gap

Categorical

Reason for Treatment
Gap

Categorical

Pre-Detention Treatment
– Mental Health

Categorical

0.
1.
2.
3.
4.

Non-user
Only outpatient
Only non-outpatient
Outpatient and inpatient
Outpatient and emergency department
visit
5. Outpatient, inpatient, and emergency
department visit
0. Non-user: No sessions
1. Low: 1-2 sessions
2. Low-to-moderate: 3-7 sessions
3. Moderate: 8-12 sessions
4. Moderate-to-high: 13-17 Sessions
5. High: 18-22 sessions
6. Extreme: ≥23 sessions
0. Non-user: No services
1. Low: 1 stay/visit
2. Low-to-moderate: 2 stays/visits
3. Moderate: 3-4 stays/visits
4. High: 5-6 stays/visits
5. Extreme: ≥7 stays/visits
0. No
1. Yes, dropped out within 3 outpatient
sessions of same treatment type
0. No dropout
1. Detention/Prison
2. Inpatient
3. No clear reason
Number of outpatient sessions obtained
until dropout
0. No
1. Yes, at least one gap between
outpatient sessions
0. No gaps
1. Detention/Prison
2. Inpatient
3. No clear reason
0. No
1. Yes, at least one mental health
treatment service prior to detention

288
Table A.1 Continued
Pre-Detention Treatment
– Substance-Related

Categorical

Pre-Detention Treatment
– Either/Both

Categorical

Pre-Detention Treatment
– Outpatient

Categorical

Pre-Detention Treatment
– Non-Outpatient

Categorical

0. No
1. Yes, at least one substance-related
treatment service prior to detention
0. No
1. Yes, at least one either/both treatment
service prior to detention
0. No
1. Yes, at least one outpatient treatment
service prior to detention
0. No
1. Yes, at least one non-outpatient
treatment service prior to detention
Number of treatment services prior to
detention
Number of mental health treatment
services prior to detention

Number of PreContinuous
Detention Treatment
Number of PreContinuous
Detention Treatment –
Mental Health
Number of PreContinuous
Number of substance-related treatment
Detention Treatment –
services prior to detention
Substance-Related
Number of PreContinuous
Number of either/both treatment services
Detention Treatment –
prior to detention
Either/Both
Number of PreContinuous
Number of outpatient treatment services
Detention Treatment –
prior to detention
Outpatient
Number of PreContinuous
Number of non-outpatient treatment
Detention Treatment –
services prior to detention
Non-Outpatient
*** Follow coding rules from the previous 12 variables for treatment utilization (for
three types and two settings) and number of treatment services (for three types and two
settings) for During Detention Treatment and for Post-Detention Treatment

289

VITA

289

VITA

Laura Morgan White
EDUCATION
August 2015

Doctor of Philosophy, Clinical Psychology
Indiana University-Purdue University Indianapolis, IN
GPA: 4.0, Summa Cum Laude
Advisors: John McGrew, PhD, and Matthew C. Aalsma, PhD
Dissertation: Mental Health and Substance-Related Treatment
Utilization, Dropout, and Continuity of Care among Detained
Adolescents: A 14-Year Longitudinal Study

May 2012

Master of Science, Clinical Psychology
Indiana University-Purdue University Indianapolis, IN
GPA: 4.0; Summa Cum Laude
Advisor: John McGrew, PhD
Thesis: Parents Served by Assertive Community Treatment: A
Needs- Based Assessment

May 2008

Bachelor of Science, Psychology
Loyola University Chicago, IL
Major: Psychology; Minor: Bioethics
GPA: 4.0; Summa Cum Laude
Advisor: R. Scott Tindale, PhD
Honors Thesis: The Use of Reasoning in resource Allocation
Decisions

290
May 2008

Bachelor of Arts, Communication
Loyola University Chicago, IL
Major: Communications
Summa Cum Laude; GPA: 3.96
Advisor: Elizabeth Lozano, PhD
Independent Study: Female Communication: A Qualitative
Analysis of Communication between College Female First Year
and Fourth Year Students

HONORS AND AWARDS
2015

Chief Intern
Primary Children’s Hospital Internship, Salt Lake City, UT

2014

Clinical Psychology Graduate Student Clinical Award
Indiana University-Purdue University Indianapolis, Department of
Psychology, Indianapolis, IN

2013

New Investigator Award, Nominee
One of five nationally recognized scholars for excellence in
research
Society for Adolescent Health and Medicine 2014 Annual Meeting,
Austin, TX

2013

First Place, Student Poster Competition
Indiana Psychological Association 2013 Fall Conference,
Indianapolis, IN

2013

Diversity Award, Student Poster Competition
Indiana Psychological Association 2013 Fall Conference,
Indianapolis, IN

291
2013

Graduate Student Research Award, Honorable Mention
Awarded annually to one graduate student for outstanding research
Indiana University-Purdue University Indianapolis, Department of
Psychology, Indianapolis, IN

2012

First Place, Student Poster Competition
Indiana Psychological Association 2012 Fall Conference,
Indianapolis, IN

2009 – 2014

Dean’s List, Purdue School of Science
Indiana University-Purdue University Indianapolis, IN

2008

Outstanding Achievement in Communication, School of
Communication
Loyola University Chicago, IL

2008

Outstanding Achievement in Science, College of Arts and Sciences
Loyola University Chicago, IL

2004 – 2008

Dean’s List, College of Arts and Sciences
Loyola University Chicago, IL

2008

Phi Beta Kappa, National Honor Society, United States

2008

Alpha Sigma Nu, National Jesuit Honor Society, United States

GRANTS
2014

Educational Enhancement Grant, $500
Graduate and Professional Student Government,
Indiana University-Purdue University Indianapolis, IN

292
2013

School of Science Graduate Student Council Travel Award, $500
Indiana University-Purdue University Indianapolis, IN

2012

School of Science Graduate Student Council Travel Award, $500
Indiana University-Purdue University Indianapolis, IN

FELLOWSHIPS AND ASSISTANTSHIPS
2013 – 2014

Pre-Doctoral Research Assistantship, Student Academic
Appointment Indiana University School of Medicine, Department
of Pediatrics, Section of Adolescent Medicine, Indianapolis, IN

2012 – 2013

Leadership Education in Adolescent Health Fellowship, Psychology
Fellow
Indiana University School of Medicine, Department of Pediatrics,
Section of Adolescent Medicine, Indianapolis, IN

2010 – 2012

Research Assistantships
Indiana University-Purdue University Indianapolis, Department of
Psychology, Indianapolis, IN

2009 – 2010

University Fellowship
Indiana University-Purdue University Indianapolis, Department of
Psychology, Indianapolis, IN

CLINICAL EXPERIENCE
August 11, 2014 –

Predoctoral Intern in Clinical Psychology

August 7, 2015

Primary Children’s Hospital Internship, Salt Lake City, UT
APA-Accredited Psychology Internship Training Program
Director of Training: Bruce Poulsen, PhD

293
First Rotation:

Psychology Intern (1000 hours, 63 children/adolescents, 48
families)

August 2014 –

Union Park Psychiatry and Counseling Services, Cottonwood
Heights, UT

February 2015

Setting: Outpatient Clinic
Hours: 45-55 hours/week
Supervisor: Matthew Wenner, PhD

Duties: Manage a weekly caseload of 16-20 direct hours with children/adolescents (ages
3-22) presenting with psychological concerns (e.g., depression, anxiety, mood
dysregulation, suicidality, substance abuse, behavior/conduct problems, trauma, social
skills, feeding/eating concerns, sleep problems, poor attention/concentration, chronic
medical conditions, autism) and psychosocial stressors (e.g., parental divorce, domestic
violence, involvement in the juvenile justice system). Conduct intakes with
children/adolescents and parents/guardians using structured interview protocols and brief
screeners. Administer and score psychological assessments, projective tests, and selfreport measures. Perform differential diagnosis. Utilize measures to assess functioning,
monitor symptoms, and track progress. Research treatment literature to determine best
practices. Coordinate with clients, parents/guardians, and supervisors to develop treatment
plans based on empirically-supported interventions. Conduct weekly (or biweekly)
individual therapy (CBT orientation with elements of family systems, developmental,
biopsychosocial). Utilize motivational interviewing, empirically-supported interventions
(e.g., CBT, DBT, ACT), and manualized evidence-based treatments (e.g., Coping Cat,
Seeking Safety, Parenting with Love and Limits). Develop and implement behavioral
reward systems and behavior modification plans. Conduct Structural Family Therapy,
parent training, psychoeducation, and feedback sessions. Collaborate/consult with
multidisciplinary team of psychologists, social workers, psychiatrists, and advanced
practice registered nurses. Consult and coordinate treatment across various systems (e.g.,
school, juvenile justice, foster care). Give referrals for medication management and
community resources. Write treatment summaries and updates for juvenile justice and
school systems. Participate in weekly staff meetings and monthly multidisciplinary case

294
consultations. Document client encounters and write progress notes within electronic
medical records system. Complete daily insurance billing sheets. Participate in weekly
individual supervision, group supervision, and peer supervision with other interns. Present
case conferences and review therapy sessions in supervision using video/audio recordings.
Attend four different weekly didactic seminars.
Second Rotation

Psychology Intern (1000 hours, 5-8 adolescents and families)

February 2014 –

Wasatch Canyons Campus, Taylorsville, UT

August 2015

Setting: Residential Unit; Day Treatment Program
Hours: 45-55 hours/week
Supervisor: Bruce Poulsen, PhD (Training Director)

Duties: Manage a weekly caseload of 3-4 adolescents (ages 13-18) with significant
psychiatric concerns (e.g., Major Depressive Disorder, Trauma-Related Disorders,
Obsessive Compulsive and Related Disorders, Bipolar Spectrum Disorders, Anxiety
Disorders, Schizophrenia Spectrum/Psychotic Disorders, Conduct Disorders, Feeding and
Eating Disorders, and Substance-Related Disorders). Learn daily schedule/operations of
Residential Unit and Day Treatment Program. Manage referrals from outside providers.
Coordinate referrals/transitions to more intensive treatment facilities (e.g., inpatient,
hospital). Complete one new comprehensive psychological evaluation with full battery of
assessments (e.g., academic, intelligence, personality, behavior, memory, attention,
adaptive functioning) and report per week. Provide weekly therapy (CBT orientation).
Develop and implement behavioral reward systems and behavior modification plans.
Conduct family therapy, psychoeducation, feedback session, and parent training.
Facilitate or co-facilitate weekly groups, such as a Mood Group, Sexual Issues Group,
and DBT. Collaborate with interdisciplinary team of providers. Document client
encounters and write progress notes within electronic medical records system. Attend
weekly staff meetings and grand rounds. When necessary, contact insurance companies
to discuss/ensure coverage for clients. Participate in weekly two-hour individual, group,
and peer supervision. Supervise a pre-doctoral graduate student. Present case conferences

295
and review therapy sessions in supervision using recordings of sessions. Attend four
different weekly didactic seminars covering a range of psychological topics.
August 2012 –

Practicum Student (310 hours, 90 adults)

June 2013

St. Vincent Primary Care Clinic, Indianapolis, IN
Setting: Integrated Primary Care Clinic
Hours: 12-15 hours/week
Supervisor: Thomas Barbera, PhD, HSPP

Duties: Received training in the Behavioral Health Consultant Model of Integrated
Primary Care. Provided behavioral health consultation as part of an integrated treatment
team of physicians, medical residents, social workers, psychiatrists, and pharmacists.
Managed a weekly caseload of 7-10 direct hours with adults (ages 17-88) presenting with
medical and/or mental health concerns (e.g., depression, anxiety, sleep disturbances,
obesity, chronic pain, diabetes, anger management, headaches, organ transplantation,
somatic complaints, attention, smoking cessation, and substance use). Conducted weekly
intake interviews and brief individual cognitive-behavioral therapy. Used motivational
interviewing, behavior modification, and relaxation training to promote health-related
behavior changes. Administered brief assessments to track symptoms and treatment
progress. Created and distributed educational handouts. Attended weekly Brown Bag
meetings with medical residents. Documented all patient encounters within electronic
medical record system. Participated in weekly individual supervision, staff meetings, and
grand rounds. Attended monthly group supervision with practicum students.
August 2012 –

Clinical Psychology Fellow (470 hours, 22 adolescents)

August 2013

Marion County Superior Probation Department, Juvenile Services
Division, Indianapolis, IN
Setting: Juvenile Detention Center
Hours: 20 hours/week
Supervisor: Matthew Aalsma, PhD, HSPP

296
Duties: Completed one weekly psychological or psycho-educational evaluation and
assessment report for the juvenile justice system. Clients included detained adolescents
(ages 12-17) awaiting trial for misdemeanor and/or felony charges. Evaluations included
structured clinical interviews with adolescents, interviews with parents/guardians, and the
administration of psychological assessments and self-report surveys measuring
intelligence, academic functioning, executive functioning, processing speed, mood,
behavior, personality, thought disturbances, substance use, and criminality. Scored and
interpreted testing results. Wrote 21 integrated reports detailing behavioral observations,
testing results, mental health diagnoses, treatment recommendations, and treatment plans
for multiple systems of care (e.g., home, school, legal). Provided feedback to
parents/guardians, schools, and probation officers. Collaborated with interdisciplinary
team of mental health providers, social workers, probation officers, judges, lawyers,
physicians, and teachers. Spoke at local community meetings about juvenile justiceinvolved adolescents. Attended weekly didactic seminars about adolescent health.
Supervised undergraduate assistants. Participated in weekly individual supervision and
monthly group supervision.
August 2011 –

Practicum Student (298 hours, 35 children/adolescents, 5 families)

March 2012

Beacon Psychology Service, LLC, Carmel, IN
Setting: Private Practice
Hours: 14-16 hours/week
Supervisor: Jennifer Horn, PhD, HSPP

Duties: Conducted one weekly psychological, psycho-educational, or neuropsychological
evaluation with children/adolescents (ages 2-18) presenting at a private practice.
Administered battery of assessments to measure intelligence, academic functioning,
language, executive functioning, visual-motor coordination, motor skills, processing
speed, memory, learning, attention/concentration, mood, behavior, personality, autism,
thought disturbances, and substance use. Scored and interpreted testing results.
Collaborated with supervisor to perform differential diagnosis, design behavioral
modification plans, and plan treatment. Wrote 12 integrated reports detailing behavioral

297
observations, testing results, diagnoses, and treatment recommendations. Conducted
intake interviews with parents and assisted with feedback sessions. Performed day-long
school observations of children with Autism Spectrum Disorders and submitted reports to
school system. Gained experience with creating Individualized Education Programs and
attending Individualized Education Programs case conferences. Co-facilitated a weekly
Social Skills Group with six males (ages of 15-18) with Autism Spectrum Disorders.
Participated in weekly individual supervision and monthly group supervision.
January 2011 –

Practicum Student (282 hours, 22 adolescents, 8 families)

May 2011

Larue D. Carter Memorial Hospital, Indianapolis, IN
Setting: Inpatient Unit
Hours: 16-18 hours/week
Supervisors: John Spanke, PhD, HSPP and Angela Neese, PhD

Duties: Managed a weekly caseload of 2-4 female adolescents (ages 13-18) with
significant psychiatric concerns (e.g., Mood Disorders, Anxiety Disorders, Behavior
Disorders, Trauma-Related Disorders, Autism Spectrum Disorders, Intellectual
Disabilities, Borderline Personality Disorder, Substance-Related Disorders, and/or
Schizophrenia Spectrum/Psychotic Disorders). Performed daily mental status exams.
Administered and scored intellectual and personality measures. Provided weekly
individual therapy and milieu therapy, using cognitive-behavioral therapy techniques,
mindfulness, relaxation training, psychoeducation, behavioral reward systems, goalsetting, motivational interviewing, coping cards, and social skills training. Facilitated
weekly Process Group and biweekly Anger Management Group. Provided
recommendations and consultation to systems of care (e.g., juvenile justice, school).
Collaborated with interdisciplinary team of psychologists, psychiatrists, recreational
therapists, medical residents, nurses, social workers, gatekeepers, and hospital staff.
Attended weekly staff meetings, chart review meetings, and treatment update meetings
with clients. Participated in weekly individual supervision and peer supervision.

298
May 2010 –

Psychology Practicum Student (438 hours, 75 adults)

December 2011

Richard L. Roudebush Veterans Administration Medical Center
Indianapolis, IN
Setting: Day Treatment Program; Inpatient Unit
Hours: 16-20 hours/week
Supervisor: Paul Lysaker, PhD, HSPP

Duties: Managed a weekly caseload of 3-4 veterans (ages 18-90) with Severe Mental
Illness (i.e., Schizophrenia Spectrum/Psychotic Disorders, Bipolar Disorder, and/or
Major Depressive Disorder). Provided weekly individual psychotherapy within an
inpatient unit and/or the Psychosocial Rehabilitation and Recovery Center (PRRC),
utilizing elements of meta-cognitive therapy, cognitive-behavioral therapy, and
interpersonal therapy. Facilitated a weekly Process Group, Social Skills Group, and
Recovery Group. Facilitated a weekly Dual Diagnosis Group with veterans diagnosed
with Severe Mental Illness and Substance-Related Disorders. Co-facilitated a weekly
Seeking Safety Group with veterans diagnosed with Post-traumatic Stress Disorder.
Obtained personal narratives from veterans. Provided psychoeducation. Collaborated
with veterans to create and implement treatment plans. Administered assessments to
measure intelligence, executive functioning, theory of mind, processing speed, memory,
learning, and attention. Wrote integrated reports detailing behavioral observations, testing
results, treatment plans, and recommendations for educational and/or vocational
programs. Served as member of an interdisciplinary team. Documented veteran
encounters and wrote progress notes within electronic medical records system.
Participated in weekly staff meetings, chart reviews, and group supervision.

COMMUNITY CLINICAL EXPERIENCE
January 2010 –

Graduate Student Clinician (108 hours, 28 adults)

October 2010

Midtown Community Mental Health Center, Indianapolis, IN
Setting: Community Mental Health Center
Hours: 3-6 hours per/week
Supervisors: Alan McGuire, PhD and Michelle Salyers, PhD

299
Duties: Co-facilitated weekly Illness Management and Recovery (IMR) and Process
Group for adults with Severe Mental Illness (i.e., Schizophrenia Spectrum Disorders or
Bipolar Spectrum Disorders). Provided psychoeducation and reviewed materials from
IMR treatment manual. Collaborated with clients to set and achieve personal goals.
Participated in training workshops and consultation calls related to IMR. Audio recorded
group sessions and reviewed audiotapes to rate fidelity of sessions using the IMR
Treatment Integrity Scale (IT-IS). Attended biweekly meetings and supervision.
August 2008 –

Children’s Counselor (1800 hours, 102 children, 44 families)

July 2009

Catherine’s Hearth Homeless Support Center, Baltimore, MD
Setting: Community Mental Health Center
Hours: 40-50 hours per/week
Supervisor: Dorothy Dobbyn, MSW

Duties: Served as children’s programming coordinator and counselor at a support center
for homeless families. Clients included homeless children/adolescents (>18 years)
presenting with a variety of problems (e.g., depression, anxiety, trauma/neglect,
intellectual disabilities, learning disabilities, conduct problems, malnutrition/feeding
problems, and sleep disturbances). Created and implemented educational, physical, and
art activities. Created and implemented behavioral reward systems. Assisted children
with school assignments, studying for tests, and completing homework. Co-facilitated
weekly family therapy sessions. Recorded daily attendance. Made weekly reminder
phone calls to families regarding appointments and activities at the center. Participated in
monthly supervision with other AmeriCorps volunteers.

SUPERVISION OF OTHER STUDENTS
August 2014 –

Peer Supervisor

Present

Primary Children’s Hospital Internship
Wasatch Canyons Campus, Taylorsville, UT
Hours: 2 hours per/week
Supervisor: Bruce Poulsen, PhD

300
Duties: Participate in weekly supervision meetings with other interns. Review ethical,
legal, and cultural issues related to clinical work. Discuss current clinical cases,
placements, and other issues related to the internship. Present cases and seek/provide
feedback, resources, and recommendations.
August 2012 –

Peer Supervisor

May 2013

Seminar in Teaching Psychology, Indiana University-Purdue
University Indianapolis, Department of Psychology, Indianapolis,
IN
Hours: 4 hours/month
Supervisor: John Guare, PhD, HSPP

Duties: Attended monthly training meetings about different theories/approaches to
supervision. Provided biweekly peer supervision and mentoring with a graduate student
completing his/her first practicum placement. Role played different clinical scenarios and
helped student learn appropriate administration of assessments. Offered support and
resources to address difficult intervention and/or assessment cases.
August 2012 –

Peer Coach

July 2013

Leadership Education in Adolescent Health, Indiana University
School of Medicine, Section of Adolescent Medicine, Indianapolis,
IN
Hours: 4 hours/month
Supervisors: Matthew Aalsma, PhD, HSPP and James Hall, PhD,
LCSW

Duties: Participated in monthly mentoring sessions with a social work graduate student.
Reviewed and offered feedback on supervisee’s research, clinical cases, and course work.
Obtained feedback and suggestions from supervisee. Discussed professional development
and career trajectory. Offered support and resources to address difficult cases involving
children/adolescents with complex mental and/or medical histories.

301
ASSESSMENTS ADMINISTERED
 Achenbach System of Empirically Based Assessment
 Achenbach Youth Self-Report, Ages 11-18
 Adaptive Behavior Assessment System-2
 Adult ADHD Self-Report Scale, Symptom Checklist
 Anxiety Disorder Symptom Rating Scale
 Attention-Deficit Hyperactivity Symptom Checklist
 Autism Diagnostic Observation Scale
 Autism Spectrum Rating Scales, including Short Form and Full Form
 Beck Anxiety Inventory
 Beck Depression Inventory-II
 Behavior Assessment System for Children-2, including Self-Report, Parent Rating
Scales, and Teacher Rating Scales
 Bell-Lysaker Emotion Recognition Task
 Beery Test of Visual Motor Integration
 Brown Attention-Deficit Disorder Scales, including Ages 3-12, Ages 12-18, Parent
Forms, and Teacher Forms
 California Verbal Learning Test-2
 California Verbal Learning Test-Children’s
 Child Behavior Checklist for Ages 6-18, including Parent Report Form and Teacher
Report Form
 Children’s Depression Inventory-II, including Self-Report and Parent Report
 Children’s Color Trails Test
 Children's Measure of Obsessive-Compulsive Symptoms
 Children’s Memory Scale
 Children's Yale-Brown Obsessive Compulsive Scale
 Clinical Evaluation of Language Fundamentals-4
 Columbia Suicide Severity Rating Scale
 Comprehensive Test of Phonological Processing-2

302
 Conners 3 ADHD Index, including Self-Report, Parent Report, and Teacher Report
 Conners Adult ADHD Rating Scales, including Short Version and Long Version
 Conners’ Kiddie Continuous Performance Test-2 and Continuous Performance Test-2
 Connors’ Rating Scale
 Delis-Kaplan Executive Function System
 Depression Symptom Rating Scale
 Developmental Test of Visual-Motor Integration-6
 Developmental Test of Visual Perception-2
 Devereux Scales of Mental Disorders
 Differentiation of Self Inventory-Short Form
 Edinburgh Postnatal Depression Scale
 Hinting Task
 Generalized Anxiety Disorder-7 item scale
 Kaufman Brief Intelligence Test-2
 Learning Disabilities Diagnostic Inventory
 Massachusetts Youth Screening Instrument-2
 Mental Status Examination
 Minnesota Multiphasic Personality Inventory-2
 Minnesota Multiphasic Personality Inventory -Adolescent
 Mood Disorder Questionnaire
 Mood Regulation Symptom Rating Scale
 Multidimensional Anxiety Scale for Children-2
 NEO-Personality Inventory-Revised
 Ohio State University TBI Identification Method
 Oral and Written Language Scale
 Panic Attack Scale
 Panic Disorder Severity ScaleSelf-Report Form
 Patient Health Questionnaire-9, Adult and Adolescent Version

303
 Prodromal Questionnaire-Brief Version
 PTSD Symptom Scale
 Reynolds Adolescent Depression Scale
 Revised Children’s Manifest Anxiety Scale
 Roberts Apperception Test for Children-2
 Rorschach Inkblot Test
 Rotter Incomplete Sentences Blanks
 Screen for Child Anxiety Related Disorders
 Self-Reflection and Insight Scale
 Sentence Completion Test
 Sentence Completion Test-Adolescent
 Sheehan Disability Scale
 Social Language Development Test Elementary
 Social Responsiveness Scale-2nd Edition Parent Rating Scale for School Age; for
Preschool Version
 Stressful Life Events Screening Questionnaire
 Structured Clinical Interview for DSM-IV
 Substance Abuse Subtle Screening Inventory-Adolescent 2
 Test of Everyday Attention for Children
 Test Observation Form for Ages 2-18
 Test of Word Reading Efficiency-2
 Test of Written Language-4
 Thematic Apperception Test
 Trail Making Test
 Trauma Symptom Checklist for Children
 Vanderbilt ADHD Diagnostic Parent Rating Scale, and Teacher Rating Scale
 Wechsler Adult Intelligence Scale-IV
 Wechsler Individual Achievement Test-III
 Wechsler Intelligence Scale for Children-IV (Training in WISC-V)

304
 Wechsler Preschool and Primary Scale of Intelligence-III
 Wender Utah Rating Scale
 Wide Range Achievement Test-4
 Wide Range Assessment of Memory and Learning-II
 Wisconsin Card Sorting Test
 Woodcock-Johnson III Tests of Achievement
 Woodcock-Johnson III Tests of Cognitive Abilities
 Youth Level of Service/Case Management Inventory

RESEARCH EXPERIENCE
August 2012 –

Study Coordinator; Predoctoral Research Fellow

July 2014

Juvenile Justice Laboratory, Indiana University School of Medicine,
Department of Pediatrics, Section of Adolescent Medicine,
Indianapolis, IN
Hours: 20-25 hours per/week
Grants/Studies: Health Care Access for Delinquent Youth
(HRSA/MCHB R40MC08721); Arrested Youth: Epidemiological,
Mental Health Care, and Cost-Effectiveness Research to Guide
Public Health Partnerships (HRSA/MCHB R40 MC 21514-01);
Youth Personal Responsibility Education Program (Indiana Family
Health Council)
Supervisor: Matthew Aalsma, PhD, HSPP

Duties: Served as member of research laboratory. Recruited participants for studies by
visiting high schools, sharing study information, answering questions, and obtaining
contact information. Visited the homes of participants, obtained written consent, and
provided reimbursement. Administered standardized surveys to juvenile justice-involved
adolescents and parole officers. Conducted standardized interviews with adolescents
participating in sexual education classes and their parents. Completed and submitted IRB
protocols. Managed longitudinal, statewide datasets. Conducted data analysis, including
data coding, data entry, and statistical analyses (e.g., qualitative analysis, grounded

305
theory analysis, structural equation modeling, regression, path analysis, survival
analysis). Prepared literature reviews. Contributed to conceptualization and design of new
studies. Assisted with writing and submitting grants. Presented research via posters and
oral presentations. Submitted and published research manuscripts. Participated in weekly
lab meetings, conference calls, and individual supervision. Supervised undergraduate
research assistants.
June 2013 –

Dissertation Research

August 2015

Chair: John H. McGrew, PhD Co-Chair: Matthew C. Aalsma,
PhD
Title: Mental Health and Substance-Related Treatment Utilization,
Dropout Rates, and Continuity of Care among Detained
Adolescents: A 14-Year Longitudinal Study

Abstract: Although approximately 60%-80% of detained adolescents have a psychiatric
disorder, little is known about their utilization of mental health and substance-related
treatment services upon release from detention. Given that treatment can potentially
reduce symptomology and recidivism, the study examined detained adolescents’ postdetention treatment utilization and longitudinal patterns of use. Data were abstracted
from the electronic juvenile justice records and medical records of 9664 detained
adolescents (62.7% male; 34.8% White, 65.2% Black; 72.6% with disorder) with
Medicaid coverage held in a Midwestern detention center at some time during 19982011. A series of statistical tests (e.g., chi-square, ANOVA, logistic regression, KaplanMeier survival analyses, and Cox regression) were conducted to identify group
differences in treatment utilization during the 14-year follow-up period. Following
detention release, approximately 66.2% of adolescents were re-arrested and 54.9% were
re-detained/incarcerated. Treatment utilization within two years post-detention was
36.7%; 31.4% obtained mental health treatment, 10.4% obtained substance-related
treatment, 36.0% obtained outpatient treatment, and 6.2% obtained non-outpatient
treatment. Among treatment users, 22.5% dropped out of treatment within 1-3 sessions
and 40.6% experienced gaps (>45 days) between treatment services. Treatment

306
utilization was significantly higher among males, White (vs. Black) adolescents, younger
adolescents, violent (vs. non-violent) offenders, recidivists (vs. non-recidivists), and
adolescents with mental disorders (vs. substance-related disorders). Variables associated
with increased likelihood of post-detention treatment included: male gender, psychiatric
disorder(s), pre-detention arrest(s), charge severity, violent offender, incarceration, and
pre-detention treatment; age and Black race were associated with decreased likelihood of
treatment. As one of the only longitudinal studies to examine treatment utilization among
detained adolescents upon community reentry, findings suggest limited service
utilization, as well as treatment gaps and disparities. Future research should focus on the
treatment needs of detained adolescents, factors associated with disparities, and
programs/policies to ensure consistent identification, referral, and connection to care for
detained adolescents.
January 2012 –

Preliminary Examination

May 2012

Chair: John H. McGrew, PhD Co-chair: Matthew C. Aalsma, PhD
Title: Treatment Utilization among the Detained Adolescent
Population: A Meta-Analysis of Treatment Prevalence and
Moderators

Abstract: Each year, about 330,000 adolescents in the US are arrested and detained in
juvenile justice facilities. As many as 60-80% of these detained adolescents (DAs) meet
criteria for a mental or substance-related disorder, compared to only 15-20% of other
adolescents. Despite the high prevalence of mental health disorders, research is limited
regarding this population’s access to and use of treatment. In fact, the prevalence of
treatment utilization among DAs is not known, with utilization rates varying widely from
3% to 76% across studies. A meta-analysis was therefore conducted to review and
synthesize the current literature regarding mental health and substance use treatment
utilization among DAs. A total of 32 studies of 34 samples totaling 21,620 adolescents
were coded and meta-analyzed using the Hunter and Schmidt approach. Mean samplesize weighted effect sizes were calculated and heterogeneity of effect sizes was examined
via Cochran Q chi-square tests, meta-regressions, and I2 indices to identify significant

307
moderators. Prevalence effect sizes (P) were low, with main effect sizes of P=26.1
(CI=18.0-0.36.2) for substance use services, P=29.6 (CI= 21.5-39.2) for mental health
services, and P=38.1 (CI=28.5-48.7) for either service. The moderator analysis identified
the following significant moderators of treatment utilization: gender, race/ethnicity,
mental health, offender severity, treatment timing, treatment setting, study location, and
study measure. Odd ratio effect sizes revealed males (OR=0.63, CI=0.52-0.78), racial
minorities (OR=0.43, CI=0.33-0.56), and ethnic minorities (OR=0.53, CI=0.37-0.75)
were significantly less likely to receive treatment. Overall, findings showed low service
utilization and treatment disparities. Future research should focus on the treatment needs
of DAs, improving services within juvenile justice facilities, and enacting programs for
consistent identification and connection to care for all DAs.

Summer 2010

Research Consultant; Study Coordinator

Summer 2011

Richard L. Roudebush Veterans Administration Medical Center

Summer 2012

Health Services Research & Development, Indianapolis, IN

Summer 2013

Hours: 10-20 hours per/week
Grants/Studies: Illness Management and Recovery (IMR)
Treatment Integrity Scale Validation and Leadership Intervention
Development (NIMH R21MH096835-01, $1,097,262); Process
Evaluation of Illness Management and Recovery in VA Mental
Health Services (VA HSR&D RRP 11-017, $154,200)
Supervisors: Alan McGuire, PhD and Michelle P. Salyers, PhD

Duties: Former coordinator and contact person for listed grants. Completed and submitted
IRB documents. Aided in recruitment of participants. Created web-based survey and
interview protocol for measuring implementation of IMR within treatment sites across
the US. Uploaded expert survey into online survey system and managed data retrieval.
Conducted phone interviews with VA staff. Reviewed interview transcripts and assisted
with creation of qualitative codebook. Coded interviews. Assisted with writing,
reviewing, and submitting grants. Presented findings at conferences and co-authored

308
published papers. Scheduled research meetings, prepared agenda, and took
minutes during research meetings. Supervised undergraduate research assistants.
January 2010 –

Thesis Research

2014

Chair: John H. McGrew, PhD
Title: Parents Served by Assertive Community Treatment: A
Needs-Based Assessment

Abstract: Assertive Community Treatment (ACT) is an evidence-based practice for
individuals with severe mental illness. Although studies estimate at least half of all
people with severe mental illness are parents, little is known about ACT policies and
treatment services for parents. Thus, this study utilized a mixed-methods design to
evaluate treatment services for parents with Schizophrenia Spectrum /Psychotic
Disorders being treated with Assertive Community Treatment (ACT). The purpose of the
study was to 1) estimate the prevalence of parent consumers, 2) identify current ACT
policies and practices for treating parent consumers, 3) and examine the perspective of
parent consumers. Quantitative and qualitative data were collected and analyzed via 2
separate studies. In study 1, 82 ACT providers from 76 teams in the US and Canada were
surveyed. Providers estimated 21.6% of ACT consumers were parents. Less than half of
providers (46.3%) reported asking consumers about parental status and only 20.7%
belonged to teams with special programs for parent consumers. The majority of providers
(75.6%) reported negative or mixed attitudes about parents with severe mental illness. In
study 2, 17 parents with severe mental illness were interviewed. All parents identified
positive aspects of parenting and most (76.5%) identified negative aspects of parenting,
including 15 parents (88.2%) who lost custody of their children. Most parents with young
children (77.8%) reported unmet parenting needs and moderate satisfaction (3.78 of 5)
with ACT services. Findings revealed significant attitude differences between providers
and parents, gaps in services, and evidence that ACT may not be effective for treating
mentally ill parents with dependent children.

309
August 2009 –

Research Assistant

May 2012

ACT Center of Indiana; Indiana University-Purdue University
Indianapolis, IN
Hours: 20 hours/week
Grants/Studies: Center of Excellence on Implementing EvidenceBased Practice (VA HSR&D, $1,200,000); Recovery Oriented
Assertive Community Treatment (NIMH 1 R24 MH074 670,
$2,505,813); Limited ACT Fidelity, Training, and Technical
Assistance (DMHA grant, $103,824)
Supervisors: John McGrew, PhD and Michelle P. Salyers, PhD

Duties: Co-coordinator for listed grants. Co-facilitated training seminars with mental
health providers within Indiana. Created, submitted, and managed IRB protocols.
Contacted and recruited Assertive Community Treatment (ACT) teams in Indiana.
Conducted fidelity assessments of ACT teams using the Dartmouth Assertive Community
Treatment Scale. Managed SPSS datasets. Wrote over 40 ACT fidelity reports to
different ACT teams outlining fidelity scores and recommendations for better
implementation of ACT. Drafted technical reports for the Department of Mental Health
and Addictions. Presented findings at national conferences. Co-authored peer-reviewed
papers. Attended weekly lab meetings and individual supervision.
August 2006 –

Research Assistant

May 2008

Social Psychology Laboratory, Loyola University Chicago, IL
Hours: 10-12 hours/ week
Supervisor: R. Scott Tindale, PhD

Duties: Ran group experiments involving undergraduate students engaging in resource
allocation and decision-making tasks. Reviewed videotapes of experiments and rated
tapes according to fidelity scales. Reviewed participants’ qualitative responses to surveys
and created qualitative coding scheme. Coded responses according to scheme. Collected,
entered, and analyzed data in SPSS. Presented research at conferences. Attended trainings
and weekly laboratory meetings.

310
PEER REVIEWED PUBLICATIONS
1. White, L. M., Aalsma, M. C., Holloway, E. A., Adams, E. L., & Salyers, M. P.
(2015). Job-related burnout among juvenile probation officers: Implications for
mental health stigma and competency. Psychological Services.
2. Aalsma, M. C., White, L.M., Lau, K. S. L., Perkins, A. J., Monohan, P. O., &
Grisso, T. (2015). Behavioral health care needs, detention-based care and criminal
recidivism at community re-entry from juvenile detention: A multi-site survival
curve analysis. American Journal of Public Health, (0), e1-e7
3. White, L. M., McGrew, J. H., Salyers, M. P., & Firmin, R. L. (2014). Assertive
community treatment for parents with serious mental illnesses: A comparison of
“parent-sensitive” assertive community treatment teams versus other teams.
Psychiatric Rehabilitation Journal, 37(3), 251-260.
4. McGuire, A. B., Luther, L., White, D.A., White, L. M., McGrew, J.H, & Salyers, M.
P. (2014). The “critical” elements of illness management and recovery: Comparing
methodological approaches. Administration and Policy in Mental Health and Mental
Health Services Research, 1-10.
5. Bonfils, K. A., Adams, E. L., Firmin, R. L., White, L. M., & Salyers, M. P. (2014).
Parenthood and severe mental illness: Relationships with recovery. Psychiatric
Rehabilitation Journal, 37(3), 186-193.
6. White, L. M., McGrew, J. H., & Salyers, M. P. (2013). Parents served by assertive
community treatment: Parenting needs, services, and attitudes. Psychiatric
Rehabilitation, 36(1), 22-27.
7. White, L. M. & McGrew, J. H. (2013). Parents served by assertive community
treatment: Prevalence, treatment services, and provider attitudes. The Journal of
Behavioral Health Services & Research, 2, 1-16.
8. McGuire, A. B., White, D. A., White, L. M., & Salyers, M. P. (2013).
Implementation of illness management and recovery in the VA: An online survey.
Psychiatric Services, 36(4), 264-271.

311
9. McGrew, J. H., White, L. M., Stull, L. G., & Wright-Berryman, J. (2013). A
comparison of self-reported and phone-based fidelity for assertive community
treatment (ACT): A pilot study in Indiana. Psychiatric Services, 64(3), 272-276.
10. McGrew, J. H., White, L. M., & Stull, L. G. (2013). Self-assessed fidelity: Proceed
with caution: In reply. Psychiatric Services, 64(4), 394-395.
11. McGuire, A. B., Stull, L. G., Mueser, K., Santos, M., Mook, A., Nicksic, C., Rose,
N., White, L. M., & Salyers, M. P. (2012). Development and reliability of a measure
of clinician competence in providing illness management and recovery. Psychiatric
Services, 63(8), 772-778.

Published Abstracts
1. White, L.M., Aalsma, M.C., Monahan, P.O., & Perkins, A. (2014). Adolescents
Involved in the Juvenile Justice System: Epidemiologic Study of Trends from 19992011. Journal of Adolescent Health, 54(2), S143-44.
2. White, L. M. & Aalsma, M. C. (2013). Mental health screenings in juvenile
detention centers: Predictors of recidivism and mental healthcare utilization among
detained adolescents with mental illness. Journal of Adolescent Health, 52(2), S1112.
3. White, L. M., Aalsma, M. C., McGrew, J. H., Salyers, M. P., & McGrew (2014).
Mental health service utilization among detained adolescents: A meta-analysis of
prevalence and potential moderators of service utilization. Journal of Adolescent
Health, 52(2), S1-S2.

Unpublished Technical Reports
1. McGrew, J. H, & White, L. M. (2012). Final report to department of mental health
and addictions: DACTS fidelity assessments of assertive community treatment teams
in Indiana. Unpublished technical report.
2. McGrew, J. H, & White, L. M. (2011). Final report to department of mental health
and addictions: Summary of self-report versus phone-based fidelity assessments for
2010-2011. Unpublished technical report.

312
Works under Review
1. White, L. M., Lau, K. S. & Aalsma, M. C. (under review). Mental health needs of
detained adolescents: Predictors of mental health treatment utilization and
recidivism. The Journal of the American Academy of Psychiatry and the Law.
2. White, L. M., Aalsma, M. C., Salyers, M. P. & McGrew, J. H. (under review).
Mental health service utilization among detained adolescents: A meta-analysis of
prevalence and potential moderators. Journal of the American Academy of Child and
Adolescent Psychiatry.

Works in Progress
1. McGuire, A. B., Kukla, M., White, D. A., White, L. M., & Salyers, M. P. Factors
affecting implementation of an evidence-based practice in the VA: Illness
management and recovery.

ORAL PRESENTATIONS AT SCIENTIFIC MEETINGS
1. Wenner, M. V., & White, L. M. (December 2014). Treatment of Anxiety Disorders.
Presentation at Provider Training Seminar at Union Park Psychiatry and Counseling
Services. Cottonwood Heights, UT.
2. Aalsma, M. C., & White, L. M. (May 2014). Job-related burnout among juvenile
probation officers. Presentation at Department of Corrections Annual Juvenile
Justice Conference. Indianapolis, IN.
3. White, L. M., McGrew, J. H., Salyers, M. P., & Aalsma, M. C. (March 2014).
Mental health service utilization among detained adolescents: A meta-analysis of
prevalence and moderators. Presentation at Society of Adolescent Health and
Medicine 2014 Annual Meeting, Austin, TX.
4. White, L. M., & Aalsma, M. C. (May 2013). Detained adolescents: Mental illness,
service utilization, and recidivism. Presentation at Midwest Psychological
Association 85th Annual Meeting, Chicago, IL.
5. White, L. M., & Aalsma, M. C. (March 2013). Mental health screenings in juvenile
detention centers: Predictors of recidivism and mental healthcare utilization among

313
detained adolescents with mental illness. Presentation at Society of Adolescent
Health and Medicine 2013 Annual Meeting, Atlanta, GA.
6. McGrew, J. H., & White, L. M. (May 2012). An innovative approach to measuring
ACT model fidelity. Workshop at 28th Annual Assertive Community Treatment
Association Conference, Boston, MA.
7. White, L. M., McGrew, J., H., & Salyers, M. (May 2012). ACT team policies and
practices for the treatment of parent consumers: Provider and consumer perspectives.
Workshop at 28th Annual Assertive Community Treatment Association Conference,
Boston, MA.
8. McGrew, J. H., White, L. M., & Rollins, A. (July 2011). DACTS fidelity results of
assertive community treatment teams in Indiana. Presentation at Department of
Mental Health and Addiction (DMHA) Annual Fidelity Meeting, Indianapolis, IN.

POSTER PRESENTATIONS AT SCIENTIFIC MEETINGS
1. White, L. M., Aalsma, M. C., Perkins, A. J., & Monahan, P. O. (April 2014).
Adolescents involved in the juvenile justice system: Epidemiologic study of trends
from 1999-2011. Poster presentation at 4th Annual Maternal Child and Health
Bureau Conference and Poster Session, Indiana University School of Science,
Indianapolis, IN.
2. White, L. M., Aalsma, M. C., Perkins, A. J., & Monahan, P. O. (March 2014).
Adolescents involved in the juvenile justice system: Epidemiologic study of trends
from 1999-2011. Poster presentation at Society of Adolescent Health and Medicine
2014 Annual Meeting, Austin, TX.
3. White, L. M., Aalsma, M. C., Perkins, A. J., & Grisso, T. (March 2014). The impact
of mental health treatment on criminal recidivism among detained adolescents: A
survival curve analysis. Poster presentation at 15th Society for Research on
Adolescence Biennial Meeting, Austin, TX.
4. White, L. M., McGrew, J. H., Salyers, M. P., & Aalsma, M. C. (November 2013).
Mental health service utilization among detained adolescents: A meta-analysis.
Poster presentation at 2013 Indiana Psychological Association Fall Conference and

314
Annual Meeting, Indianapolis, IN
5. White, L. M., McGrew, J. H., Salyers, M. P., & Aalsma, M. C. (October 2013).
Mental health service utilization among detained adolescents: A meta-analysis of
prevalence and moderator factors. Poster presentation at Institute of Psychiatric
Services 65th Annual Meeting: Transforming Psychiatric Practice, Reforming Health
Care, Philadelphia, PA.
6. White, L. M., McGrew, J. H., & Salyers, M. P. (May 2013). Parents served by
assertive community treatment: Needs, services, and attitudes. Poster presentation at
Midwest Psychological Association 85th Annual Meeting, Chicago, IL.
7. White, L. M., & Aalsma, M. C. (April 2013). Detained adolescents: Mental illness,
service utilization, and recidivism. Poster presentation at 3rd Annual Maternal Child
and Health Bureau Poster Session, Indiana University School of Science,
Indianapolis, IN.
8. White, L. M., & Aalsma, M. C. (April 2013). Mental health screenings in juvenile
detention centers: Predictors of mental healthcare utilizations and recidivism. Poster
presentation at 17th Annual PhD Spring Symposium 2013, Indiana University School
of Social Work, Indiana University-Purdue University Indianapolis, Indianapolis,
IN.
9. White, L. M., & Aalsma, M. C. (April 2013). Mental health screenings in juvenile
detention centers: Predictors of mental healthcare utilizations and recidivism. Poster
presentation at 2013 IUPUI Research Day, Indiana University-Purdue University
Indianapolis, Indianapolis, IN.
10. White, L. M., & Aalsma, M. C. (February 2013). Detained adolescents: Mental
illness, service utilization, and recidivism. Poster presentation at 24th Annual Joseph
T. Taylor Symposium, Indiana University-Purdue University Indianapolis,
Indianapolis, IN.
11. White, L. M., & Aalsma, M. C. (October 2012). Mental health screenings in
juvenile detention centers: Predictors of mental healthcare utilizations and
recidivism. Poster presentation at 2012 Indiana Psychological Association Fall
Conference and Annual Meeting, Carmel, IN.

315
12. White, L. M., McGrew, J. H., & Salyers, M. P. (April 2012). Parents served by
assertive community treatment: A needs-based assessment. Poster presentation at
2012 IUPUI Research Day, Indiana University-Purdue University Indianapolis,
Indianapolis, IN.
13. McGuire, A. B., Stull, L. G., Mueser, K., Santos, M., Mook, A., Nicksic, C., Rose,
N., White, L. M., & Salyers, M. P. (March 2011). The illness management and
recovery treatment integrity scale: Development and reliability. Poster presentation at
4th Annual NIH Conference on the Science of Dissemination and Implementation,
Bethesda, MD.

TEACHING EXPERIENCE
Summer 2011

Seminar in Teaching Psychology
Indiana University-Purdue University Indianapolis, Department of
Psychology, Indianapolis, IN

2009 – 2010

Graduate Teaching Assistant
Psychology 499, Honors Research Seminar
Indiana University-Purdue University Indianapolis, Department of
Psychology, Indianapolis, IN

2008 – 2009

Substitute Instructor
Psychology 104, Psychology as a Social Science
Indiana University-Purdue University Indianapolis, Department of
Psychology, Indianapolis, IN

2004 – 2008

Private Tutor
Biology, Mathematics, and Statistics
Purdue University Calumet, Hammond, IN

316
2005 – 2007

Elementary School Substitute Teacher
Kindergarten to 4th grade
Merrillville Community School Corporation, Merrillville, IN

LEADERSHIP AND COMMUNITY SERVICE
2010 – 2014

Treasurer, Psychology Graduate Student Organization
Indiana University-Purdue University Indianapolis, IN

2006 – 2008

Mentor and Tutor
Inspired Youth Tutoring, Chicago, IL

2006 – 2008

Student Volunteer
Midwest Psychological Association, Chicago, IL

2007 – 2008

Vice-President and Chair of Community Service
National Society of Collegiate Scholars, Loyola University
Chicago, IL

2004 – 2008

Vice President, Lambda Bi Eta, Beta Rho Chapter
National Honor Society in Communication, Loyola University
Chicago, IL

PROFESSIONAL ASSOCIATIONS AND ACTIVITIES
2012 – 2014

Student Member, Neuropsychology Graduate Student Organization

2011 – Present

Student Affiliate, Indiana Psychological Association

2011 – Present

American Psychological Association, Professional Affiliate
o Division 35: Society for the Psychology of Women
o Division 41: American Psychology: Law Society

317
o Division 53: Society of Clinical Child and Adolescent
Psychology
o Division 54: Society of Pediatric Psychology
2009 – Present

Reviewer for Journal of Behavioral Health Services and Research
Ad Hoc Review for Psychiatric Services, Health Psychology, and
Journal of Adolescent Health

2009 – Present

Assistance with grants, including writing, editing, and reviewing
grants

2006 – Present

Student Affiliate, Midwestern Psychological Association

PROFESSIONAL SEMINARS
August 2014 –

Family Therapy Seminar

June 2015

Program: Primary Children's Hospital Psychology Internship
Location: Wasatch Canyons Campus, Taylorsville, UT

Description: Weekly seminar covering different theories, approaches, research, and
practices related to family therapy. Based on the live observation team approach, the
seminar provides instruction in all aspects of family therapy, including live observation
of family therapy sessions, co-facilitation of family therapy sessions, review of recorded
session materials and clinical notes, and group supervision with trained experts in
Structured Family Therapy and Functional Family Therapy.
August 2014 –

Psychotherapy Seminar

June 2015

Program: Primary Children's Hospital Psychology Internship
Location: University of Utah Neuropsychiatric Institute, Salt Lake
City, UT

Description: Weekly training seminar for pre-doctoral interns and psychiatry residents
from the University of Utah that addresses the diagnosis, assessment, and treatment of

318
children, adolescents, and families. Professionals provide specialized training in
evidenced-based treatments, including Cognitive Behavioral Therapy, Dialectical
Behavior Therapy, and Acceptance and Commitment Therapy. Clinical cases are
presented with video or audio recordings, followed by discussion, review, and
recommendations for these cases.
August 2014 –

Didactic Training Seminar

June 2015

Program: Primary Children's Hospital Psychology Internship
Location: Varied

Description: Weekly seminar covering a myriad of topics related to Child and Adolescent
Psychology. Presenters include providers at Primary Children’s Hospital, communitybased psychologists, private practitioners, and academic faculty at the University of Utah.
Some seminars are held with psychology interns from other local APA-accredited
psychology internship training programs, including Utah State Hospital, The University
of Utah Neuropsychiatric Institute, and the Veterans Administration Salt Lake City
Health Care System.
December 2014 –

Neuropsychology Seminar

February 2015

Program: Primary Children's Hospital Psychology Internship
Location: Primary Children’s Hospital, Salt Lake City, UT

Description: Weekly seminar focused on topics within neuropsychology, including the
role of neuro-psychology on an interdisciplinary treatment team, issues related to working
within a medical setting, current neuropsychological assessments, neuroimaging,
neuropsychological disorders, and evidence-based treatments.
2009 – 2014

Proseminar in Clinical Psychology
Program: Clinical Psychology PhD Program
Location: Indiana University-Purdue University Indianapolis,
Department of Psychology, Indianapolis, IN

Description: Weekly professional development course covering advanced clinical topics,
including professional ethics, consultation, multicultural competence, case conferences,

319
clinical skills, evidence-based practices and interventions, research methods,
publications, supervision, teaching, internships, job opportunities, licensure, and grant
writing.
2013 – 2014

Supervision Seminar
Program: Teaching in Psychology
Location: Indiana University-Purdue University Indianapolis,
Department of Psychology, Indianapolis, IN

Description: Biweekly course covering topics related to supervision within the field of
psychology, including different supervision theories/approaches, ethical and legal issues,
challenges to receiving and/or providing supervision, case studies, and evidence-based
supervision.
2012 – 2013

Seminar in Adolescent Health and Medicine
Program: Leadership Education in Adolescent Health
Location: Indiana University School of Medicine Department of
Pediatrics, Section of Adolescent Medicine, Indianapolis, IN

Description: Weekly professional development course covering topics related to
adolescent health and medicine in the fields of psychology, psychiatry, social work,
nutrition, nursing, and medicine.
2012 – 2013

Indiana Inter-Professional Leadership Learning Collaborative
Program: Leadership Education in Neurodevelopmental
Disabilities
Location: Child Development Center, Riley Hospital for Children,
Indianapolis, IN

Description: Biweekly half-day training workshop for professionals in psychology,
psychiatry, social work, nutrition, nursing, and medicine. Training topics included
developing leadership skills, engaging in advocacy, public policy, and working within
interdisciplinary teams.

320
WORKSHOPS
November 2014

Introduction to Wechsler Intelligence Scale for Children-V
Presenter: Megan Frye, MS, LMHW, Pre-Doctoral Intern
Location: Primary Children’s Center for Counseling, Taylorsville,
UT

October 2014

Motivational Interviewing
Presenter: Kelly Lundberg, PhD, Psychologist, Associate
Professor, Department of Psychiatry, University of Utah
Location: Assessment & Referral Services, Salt Lake City, UT

October 2014

Psychological and Neuropsychological Sequelae of Pediatric
Cancer
Presenter: Paul Colte, PsyD, MSCP, Pediatric Psychologist,
Department of Pediatric Hematology/Oncology and Bone Marrow
Transplant, Primary Children's Hospital
Location: Primary Children’s Hospital, Salt Lake City, UT

September 2014

Autism Diagnostic Observation Schedule-2 Training
Presenter: Stephanie Creekpaum, PsyD, Child and Adolescent
Psychologist, Primary Children’s Center for Counseling
Location: Primary Children’s Center for Counseling, Taylorsville,
UT

April 2014

Biofeedback
Presenter: Eric Scott, PhD, Assistant Professor of Clinical
Psychology, Department of Psychiatry, Indiana University School
of Medicine
Location: Indiana University-Purdue University Indianapolis, IN

321
March 2014

Mechanisms of Mindfulness for Therapeutic Change
Presenter: Linda Brown, PhD, Behavioral Health & Social Service
Provider, Private Practice in Bloomington, Indiana
Location: Indiana University-Purdue University Indianapolis, IN

October 2013

Consultation in Psychology: End-of-Life Services
Presenter: Susan Hickman, PhD, Associate Professor, Co-Director,
Melvin and Bren Simon Cancer Center, Indiana University School
of Nursing
Location: Indiana University-Purdue University Indianapolis, IN

August 2013

Meta-Analysis Workshop
Presenter: Noel Card, PhD, Associate Professor, University of
Arizona
Location: Indiana University-Purdue University Indianapolis, IN

April 2013

Self-Hypnosis Training for Chronic Main Management
Presenter: Mark Jensen, PhD, Professor and Vice Chair of
Research, Department of Rehabilitation Medicine, University of
Washington
Location: Indiana University-Purdue University Indianapolis, IN

January 2013

Consultation: Evidenced-Based Practices and Fidelity Assessment
Presenter: Angie Rollins, PhD, Research Scientist, IN
University-Purdue University Indianapolis, ACT Center of Indiana
Location: Indiana University-Purdue University Indianapolis, IN

December 2012

Illness Management and Recovery
Presenter: Alan McGuire, PhD, Research Scientist, Richard L.
Roudebush Veterans Administration Medical Center

322
Location: Richard L. Roudebush Veterans Administration Medical
Center, Indianapolis, IN

August 2012

Introduction to Structural Equation Modeling
Presenter: Gregory R. Hancock, PhD, Professor and Director of
Center for Integrated Latent Variable Research, University of
Maryland
Location: Indiana University-Purdue University Indianapolis, IN

October 2011

Overview of Mental Health and Behavioral Consultation
Presenter: Lisa Ruble, PhD, Associate Professor of Educational
Psychology, University of Kentucky
Location: Indiana University-Purdue University Indianapolis, IN

August 2011

Sort & Sift, Think & Shift: Multidimensional Qualitative Analysis
Presenter: Raymond Maietta, PhD, Research Talk, Inc.
Location: Indiana University-Purdue University Indianapolis, IN

July 2011

Fidelity Assessment of Recovery-Oriented Interventions
Presenters: Michelle Salyers, PhD, Associate Professor of
Psychology, Indiana University-Purdue University Indianapolis
Alan McGuire, PhD, Research Scientist, Roudebush Veterans
Administration Medical Center, Indianapolis, IN
Location: Richard L. Roudebush Veterans Administration Medical
Center, Indianapolis, IN

June 2011

Motivational Interviewing
Presenters: Hea-Won Kim, PhD Associate Professor of Social
Work, Indiana University-Purdue University Indianapolis

323
Angela Rollins, PhD, Assistant Research Professor, Indiana
University-Purdue University Indianapolis
Location: Indiana University-Purdue University Indianapolis, IN

May 2011

Group Schema Therapy
Presenter: Joan Farrell, PhD, Adjunct Professor, Indiana University
School of Medicine
Location: Indiana University-Purdue University Indianapolis, IN

January 2011

Introduction to Grounded Theory A Social Constructionist
Approach
Presenter: Kathy Charmaz, PhD, Professor of Sociology and
Director of Faculty Writing Program, Sonoma State University
Location: Cary, NC

October 2008

Evidence Based Practices in Children’s Mental Health
Presenter: National Alliance on Mental Illness, Baltimore Chapter
Location: Baltimore, MD