Browsing by Subject "Health care"

Now showing 1 - 10 of 24
  • Thumbnail Image
    Item
    A large language model for electronic health records
    (Springer Nature, 2022-12-26) Yang, Xi; Chen, Aokun; PourNejatian, Nima; Shin, Hoo Chang; Smith, Kaleb E.; Parisien, Christopher; Compas, Colin; Martin, Cheryl; Costa, Anthony B.; Flores, Mona G.; Zhang, Ying; Magoc, Tanja; Harle, Christopher A.; Lipori, Gloria; Mitchell, Duane A.; Hogan, William R.; Shenkman, Elizabeth A.; Bian, Jiang; Wu, Yonghui; Health Policy and Management, Richard M. Fairbanks School of Public Health
    There is an increasing interest in developing artificial intelligence (AI) systems to process and interpret electronic health records (EHRs). Natural language processing (NLP) powered by pretrained language models is the key technology for medical AI systems utilizing clinical narratives. However, there are few clinical language models, the largest of which trained in the clinical domain is comparatively small at 110 million parameters (compared with billions of parameters in the general domain). It is not clear how large clinical language models with billions of parameters can help medical AI systems utilize unstructured EHRs. In this study, we develop from scratch a large clinical language model—GatorTron—using >90 billion words of text (including >82 billion words of de-identified clinical text) and systematically evaluate it on five clinical NLP tasks including clinical concept extraction, medical relation extraction, semantic textual similarity, natural language inference (NLI), and medical question answering (MQA). We examine how (1) scaling up the number of parameters and (2) scaling up the size of the training data could benefit these NLP tasks. GatorTron models scale up the clinical language model from 110 million to 8.9 billion parameters and improve five clinical NLP tasks (e.g., 9.6% and 9.5% improvement in accuracy for NLI and MQA), which can be applied to medical AI systems to improve healthcare delivery. The GatorTron models are publicly available at: https://catalog.ngc.nvidia.com/orgs/nvidia/teams/clara/models/gatortron_og
  • Thumbnail Image
    Item
    A Qualitative Study of Biomedical Engineering Student Critical Reflection During Clinical Immersion Experiences
    (Springer Nature, 2024) Tabassum, Nawshin; Higbee, Steven; Miller, Sharon; Economics, School of Liberal Arts
    Purpose: Clinical immersion experiences provide engineering students with opportunities to identify unmet user needs and to interact with clinical professionals. These experiences have become common features of undergraduate biomedical engineering curricula, with many published examples in the literature. There are, however, few or no published studies that describe rigorous qualitative analysis of biomedical engineering student reflections from immersion programs. Methods: Fifteen reflection prompts that align with program learning goals were developed and structured based on the DEAL model for critical reflection. Undergraduate participants in a summer immersion program responded to these prompts throughout five weeks of clinical rotations. Data from two summer cohorts of participants (n = 20) were collected, and thematic analysis was performed to characterize student responses. Results: Students reported learning about key healthcare topics, such as medical insurance, access to healthcare (and lack thereof), stakeholder perspectives, and key medical terminology and knowledge. Most reflections also noted that students could apply newly gained medical knowledge to biomedical engineering design. Further, clinical immersion provided students with a realistic view of the biomedical engineering profession and potential areas for future professional growth, with many reflections identifying the ability to communicate with a variety of professionals as key to student training. Some students reflected on conversations with patients, noting that these interactions reinvigorated their passion for the biomedical engineering field. Finally, 63% of student reflections identified instances in which patients of low socioeconomic status were disadvantaged in health care settings. Conclusions: Clinical immersion programs can help close the gap between academic learning and the practical experience demands of the field, as design skills and product development experience are becoming increasingly necessary for biomedical engineers. Our work initiates efforts toward more rigorous analysis of students' reactions and experiences, particularly around socioeconomic and demographic factors, which may provide guidance for continuous improvement and development of clinical experiences for biomedical engineers.
  • Thumbnail Image
    Item
    Acute respiratory failure and the kinetics of neutrophil recovery in pediatric hematopoietic cell transplantation: a multicenter study
    (Springer Nature, 2020-02) Moffet, J.R.; Mahadeo, K.M.; McArthur, J.; Hsing, D.D.; Gertz, S.J.; Smith, L.S.; Loomis, A.; Fitzgerald, J.C.; Nitu, M.E.; Duncan, C.N.; Hall, M.W.; Pinos, E.L.; Tamburro, R.F.; Simmons, R.A.; Troy, J.; Cheifetz, I.M.; Rowan, C.M.; Pediatrics, School of Medicine
    In this multicenter study, we investigated the kinetics of neutrophil recovery in relation to acuity and survival among 125 children undergoing allogeneic hematopoietic cell transplantation (allo-HCT) who required invasive mechanical ventilation (IMV). Recovery of neutrophils, whether prior to or after initiation of IMV, was associated with a significantly decreased risk of death relative to never achieving neutrophil recovery. A transient increase in acuity (by oxygenation index and vasopressor requirements) occurred among a subset of the patients who achieved neutrophil recovery after initiation of IMV; 61.5% of these patients survived to discharge from the intensive care unit (ICU). Improved survival among patients who subsequently achieved neutrophil recovery on IMV was not limited to those with peri-engraftment respiratory distress syndrome. The presence of a respiratory pathogen did not affect the risk of death while on IMV but was associated with an increased length of IMV (p < 0.01). Among patients undergoing HCT who develop respiratory failure and require advanced therapeutic support, neutrophil recovery at time of IMV and/or presence of a respiratory pathogen should not be used as determining factors when counseling families about survival.
  • Thumbnail Image
    Item
    Adolescents' Discussion of Sexual and Reproductive Health Care Topics With Providers: Findings From a Nationally Representative Probability Sample of U.S. Adolescents
    (Elsevier, 2021-03) Hensel, Devon J.; Herbenick, Debby; Beckmeyer, Jonathon J.; Fu, Tsung-chieh; Dodge, Brian; Pediatrics, School of Medicine
    Purpose National practice guidelines encourage providers address sexual and reproductive health (SRH) as part of all clinical encounters with adolescents. Yet, no studies provide nationally representative estimates of how frequently adolescents are screened. Methods Data were adolescent participants (aged 14–17 years; N = 826) in the 2018 National Survey of Sexual Health and Behavior, an online, nationally representative study of sexual health experiences of people in the U.S. SRH variables were: (all no/yes) pregnancy prevention, sexual identity, STD/HIV prevention, sexual difficulties, sexually transmitted infections testing, and sexual activity. We used descriptive statistics and weighted logistic regression (Stata 16.0; all p < .05) to examine differences in the odds of SRH discussion with provider by sexual identity, age, gender, and race/ethnicity. Results The coverage of SRH topics was poor. The most common topic was asking about sexual activity (52.9%), and the least common was being offered a sexually transmitted infection test (21.7%). An adolescent’s sexual identity, race/ethnicity, and age affected the odds of topic screening. Conclusions Health care providers appear to both infrequently and inconsistently address key SRH topics during encounters with young people. Targeted interventions should focus on strengthening the regularity and depth of clinicians’ SRH conversations regardless of adolescent demographic or history.
  • Thumbnail Image
    Item
    African-American Hospitals and Health Care in Early Twentieth Century Indianapolis, Indiana, 1894-1917
    (2016-05) Erickson, Norma B.; Labode, Modupe Gloria; Schneider, William H.; Barrows, Robert G.
    At the end of the nineteenth century, the African-American population of Indianapolis increased, triggering a need for health care for the new emigrants from the South. Within the black population, some individuals pursued medical degrees to become physicians. At the same time, advances in medical treatment—especially surgical operations—shifted the most common site of care from patients’ homes to hospitals. Professionally trained nurses, mostly white, began to replace family members or untrained African-American nurses who previously delivered care to Black patients. Barriers of racial segregation kept both the Black doctors and Black nurses from practicing in the municipal City Hospital in Indianapolis. To remedy this problem, the city's African-American leaders undertook establishing healthcare institutions with nurse training schools during the first few years of the twentieth century. This thesis argues that the healthcare institution-building that occurred in the early twentieth century offered opportunities for the practice of self-help in the Black community. The institutions also created a bridge for Black-white relations because the Black hospitals attracted the support of prominent white leaders. Good health and health care for the sick or injured were necessary to achieve racial uplift, and healthcare consumption became an indicator of social status and economic success. Racially segregated institutions afforded doctors and nurses a chance to increase their expertise and prove they were capable of functioning in the public hospital system. After a decade of working in separate institutions, the Black community prepared to push for full access to the city's tax-supported City Hospital as a civil right.
  • Thumbnail Image
    Item
    Analysis of Hospital Quality Measures and Web-Based Chargemasters, 2019: Cross-sectional Study
    (JMIR, 2021-08-19) Patel, Kunal N.; Mazurenko, Olena; Ford, Eric; Health Policy and Management, Richard M. Fairbanks School of Public Health
    Background: The federal health care price transparency regulation from 2019 is aimed at bending the health care cost curve by increasing the availability of hospital pricing information for the public. Objective: This study aims to examine the associations between publicly reported diagnosis-related group chargemaster prices on the internet and quality measures, process indicators, and patient-reported experience measures. Methods: In this cross-sectional study, we collected and analyzed a random 5.02% (212/4221) stratified sample of US hospital prices in 2019 using descriptive statistics and multivariate analysis. Results: We found extreme price variation in shoppable services and significantly greater price variation for medical versus surgical services (P=.006). In addition, we found that quality indicators were positively associated with standard charges, such as mortality (β=.929; P<.001) and readmissions (β=.514; P<.001). Other quality indicators, such as the effectiveness of care (β=-.919; P<.001), efficient use of medical imaging (β=-.458; P=.001), and patient recommendation scores (β=-.414; P<.001), were negatively associated with standard charges. Conclusions: We found that hospital chargemasters display wide variations in prices for medical services and procedures and match variations in quality measures. Further work is required to investigate 100% of US hospital prices posted publicly on the internet and their relationship with quality measures.
  • Thumbnail Image
    Item
    Artificial Intelligence in Health, Health Care, and Biomedical Science: An AI Code of Conduct Principles and Commitments Discussion Draft
    (National Academy of Sciences, 2024-04-08) Adams, Laura; Fontaine, Elaine; Lin, Steven; Crowell, Trevor; Chung, Vincent C. H.; Gonzalez, Andrew A.; Surgery, School of Medicine
    This commentary presents initial concepts and content that the Steering Committee feel may be important to a draft Code of Conduct framework for use in the development and application of artificial intelligence (AI) in health, health care, and biomedical science.
  • Thumbnail Image
    Item
    Clinician and Visitor Activity Patterns in an Intensive Care Unit Room: A Study to Examine How Ambient Monitoring Can Inform the Measurement of Delirium Severity and Escalation of Care
    (MDPI, 2024-10-14) Nalaie, Keivan; Herasevich, Vitaly; Heier, Laura M.; Pickering, Brian W.; Diedrich, Daniel; Lindroth, Heidi; School of Nursing
    The early detection of the acute deterioration of escalating illness severity is crucial for effective patient management and can significantly impact patient outcomes. Ambient sensing technology, such as computer vision, may provide real-time information that could impact early recognition and response. This study aimed to develop a computer vision model to quantify the number and type (clinician vs. visitor) of people in an intensive care unit (ICU) room, study the trajectory of their movement, and preliminarily explore its relationship with delirium as a marker of illness severity. To quantify the number of people present, we implemented a counting-by-detection supervised strategy using images from ICU rooms. This was accomplished through developing three methods: single-frame, multi-frame, and tracking-to-count. We then explored how the type of person and distribution in the room corresponded to the presence of delirium. Our designed pipeline was tested with a different set of detection models. We report model performance statistics and preliminary insights into the relationship between the number and type of persons in the ICU room and delirium. We evaluated our method and compared it with other approaches, including density estimation, counting by detection, regression methods, and their adaptability to ICU environments.
  • Thumbnail Image
    Item
    Endoscopic versus surgical management for colonic volvulus hospitalizations in the United States
    (Korean Society of Gastrointestinal Endoscopy, 2023) Dahiya, Dushyant Singh; Perisetti, Abhilash; Goyal, Hemant; Inamdar, Sumant; Singh, Amandeep; Garg, Rajat; Cheng, Chin-I; Al-Haddad, Mohammad; Sanaka, Madhusudhan R.; Sharma, Neil; Medicine, School of Medicine
    Background/aims: Colonic volvulus (CV), a common cause of bowel obstruction, often requires intervention. We aimed to identify hospitalization trends and CV outcomes in the United States. Methods: We used the National Inpatient Sample to identify all adult CV hospitalizations in the United States from 2007 to 2017. Patient demographics, comorbidities, and inpatient outcomes were highlighted. Outcomes of endoscopic and surgical management were compared. Results: From 2007 to 2017, there were 220,666 CV hospitalizations. CV-related hospitalizations increased from 17,888 in 2007 to 21,715 in 2017 (p=0.001). However, inpatient mortality decreased from 7.6% in 2007 to 6.2% in 2017 (p<0.001). Of all CV-related hospitalizations, 13,745 underwent endoscopic intervention, and 77,157 underwent surgery. Although the endoscopic cohort had patients with a higher Charlson comorbidity index, we noted lower inpatient mortality (6.1% vs. 7.0%, p<0.001), mean length of stay (8.3 vs. 11.8 days, p<0.001), and mean total healthcare charge ($68,126 vs. $106,703, p<0.001) compared to the surgical cohort. Male sex, increased Charlson comorbidity index scores, acute kidney injury, and malnutrition were associated with higher odds of inpatient mortality in patients with CV who underwent endoscopic management. Conclusion: Endoscopic intervention has lower inpatient mortality and is an excellent alternative to surgery for appropriately selected CV hospitalizations.
  • Thumbnail Image
    Item
    Factors associated with optimal patient outcomes after operative repair of isolated hip fractures in the elderly
    (BMJ, 2020-12-22) deMeireles, Alirio J.; Gerhardinger, Laura; Oliphant, Bryant W.; Jenkins, Peter C.; Cain-Nielsen, Anne H.; Scott, John W.; Hemmila, Mark R.; Sangji, Naveen F.; Surgery, School of Medicine
    Background: Increased time to operative intervention is associated with a greater risk of mortality and complications in adults with a hip fracture. This study sought to determine factors associated with timeliness of operation in elderly patients presenting with an isolated hip fracture and the influence of surgical delay on outcomes. Methods: Trauma quality collaborative data (July 2016 to June 2019) were analyzed. Inclusion criteria were patients ≥65 years with an injury mechanism of fall, Abbreviated Injury Scale (AIS) 2005 diagnosis of hip fracture, and AIS extremity ≤3. Exclusion criteria included AIS in other body regions >1 and non-operative management. We examined the association of demographic, hospital, injury presentation, and comorbidity factors on a surgical delay >48 hours and patient outcomes using multivariable regression analysis. Results: 10 182 patients fit our study criteria out of 212 620 patients. Mean age was 82.7±8.6 years and 68.7% were female. Delay in operation >48 hours occurred in 965 (9.5%) of patients. Factors that significantly increased mortality or discharge to hospice were increased age, male gender, emergency department hypotension, functionally dependent health status (FDHS), advanced directive, liver disease, angina, and congestive heart failure (CHF). Delay >48 hours was associated with increased mortality or discharge to hospice (OR 1.52; 95% CI 1.13 to 2.06; p<0.01). Trauma center verification level, admission service, and hip fracture volume were not associated with mortality or discharge to hospice. Factors associated with operative delay >48 hours were male gender, FDHS, CHF, chronic renal failure, and advanced directive. Admission to the orthopedic surgery service was associated with less incidence of delay >48 hours (OR 0.43; 95% CI 0.29 to 0.64; p<0.001). Discussion: Hospital verification level, admission service, and patient volume did not impact the outcome of mortality/discharge to hospice. Delay to operation >48 hours was associated with increased mortality. The only measured modifiable characteristic that reduced delay to operative intervention was admission to the orthopedic surgery service. Level of evidence: III.