Browsing by Subject "Social factors"

Now showing 1 - 3 of 3
  • Thumbnail Image
    Item
    Natural language processing-driven state machines to extract social factors from unstructured clinical documentation
    (Oxford University Press, 2023-04-18) Allen, Katie S.; Hood, Dan R.; Cummins, Jonathan; Kasturi, Suranga; Mendonca, Eneida A.; Vest, Joshua R.; Health Policy and Management, School of Public Health
    Objective: This study sought to create natural language processing algorithms to extract the presence of social factors from clinical text in 3 areas: (1) housing, (2) financial, and (3) unemployment. For generalizability, finalized models were validated on data from a separate health system for generalizability. Materials and methods: Notes from 2 healthcare systems, representing a variety of note types, were utilized. To train models, the study utilized n-grams to identify keywords and implemented natural language processing (NLP) state machines across all note types. Manual review was conducted to determine performance. Sampling was based on a set percentage of notes, based on the prevalence of social need. Models were optimized over multiple training and evaluation cycles. Performance metrics were calculated using positive predictive value (PPV), negative predictive value, sensitivity, and specificity. Results: PPV for housing rose from 0.71 to 0.95 over 3 training runs. PPV for financial rose from 0.83 to 0.89 over 2 training iterations, while PPV for unemployment rose from 0.78 to 0.88 over 3 iterations. The test data resulted in PPVs of 0.94, 0.97, and 0.95 for housing, financial, and unemployment, respectively. Final specificity scores were 0.95, 0.97, and 0.95 for housing, financial, and unemployment, respectively. Discussion: We developed 3 rule-based NLP algorithms, trained across health systems. While this is a less sophisticated approach, the algorithms demonstrated a high degree of generalizability, maintaining >0.85 across all predictive performance metrics. Conclusion: The rule-based NLP algorithms demonstrated consistent performance in identifying 3 social factors within clinical text. These methods may be a part of a strategy to measure social factors within an institution.
  • Thumbnail Image
    Item
    Relation of household income to access and adherence to combination sacubitril/valsartan in heart failure: a retrospective analysis of commercially insured patients
    (American Heart Association, 2022) Johnson, Amber E.; Swabe, Gretchen M.; Addison, Daniel; Essien, Utibe R.; Breathett, Khadijah; Brewer, LaPrincess C.; Mazimba, Sula; Mohammed, Selma F.; Magnani, Jared W.; Medicine, School of Medicine
    Background: Outcomes in heart failure with reduced ejection fraction (HFrEF) are influenced by access and adherence to guideline-directed medical therapy. Our objective was to study the association between annual household income and: (1) the odds of having a claim for sacubitril/valsartan among insured patients with HFrEF and (2) medication adherence (measured as the proportion of days covered). We hypothesized that lower annual household income is associated with decreased odds of having a claim for and adhering to sacubitril/valsartan. Methods: Using the Optum de-identified Clinformatics Data Mart, patients with HFrEF and ≥6 months of enrollment for follow-up (2016-2020) were included. Covariates included age, sex, race, ethnicity, educational attainment, US region, number of prescribed medications, and Elixhauser Comorbidity Index. Prescription for sacubitril/valsartan was defined by the presence of a claim within 6 months of HFrEF diagnosis. Adherence was defined as proportion of days covered ≥80%. We fit multivariable-adjusted logistic regression models and hierarchical logistic regression accounting for covariates. Results: Among 322 007 individuals with incident HFrEF, 135 282 had complete data for analysis. Of the patients eligible for sacubitril/valsartan, 4.7% (6372) had a claim within 6 months of HFrEF diagnosis. Following multivariable adjustment, individuals in the lowest annual income category (<$40 000) were significantly less likely (odds ratio, 0.83 [95% CI, 0.76-0.90]) to have a sacubitril/valsartan claim within 6 months of HFrEF diagnosis than those in the highest annual income category (≥$100 000). Annual income <$40 000 was associated with lower odds of proportion of days covered ≥80% compared with income ≥$100 000 (odds ratio, 0.70 [95% CI, 0.59-0.83]). Conclusions: Lower household income is associated with decreased likelihood of a sacubitril/valsartan claim and medication adherence within 6 months of HFrEF diagnosis, even after adjusting for sociodemographic and clinical factors. Future analyses are needed to identify additional social factors associated with delays in sacubitril/valsartan initiation and long-term adherence.
  • Thumbnail Image
    Item
    Social Determinants of Health and Medication Adherence in Older Adults with Prevalent Chronic Conditions in the United States: An Analysis of the National Health and Nutrition Examination Survey (NHANES) 2009–2018
    (MDPI, 2025-02-07) Adeoye-Olatunde, Omolola A.; Hastings, Tessa J.; Blakely, Michelle L.; Boyd, LaKeisha; Aina, Azeez B.; Sherbeny, Fatimah; Biostatistics and Health Data Science, Richard M. Fairbanks School of Public Health
    Background: The older adult population is rapidly expanding in the United States (US), with a high prevalence of high blood pressure, high cholesterol, and diabetes. Medication nonadherence is prevalent in this population, with less evidence on the influence of social determinants of health (SDoH). Thus, the objective of this study was to identify and prioritize SDoH associated with medication adherence among US older adults with these comorbidities. Method: Using the World Health Organization Commission on Social Determinants of Health and Pharmacy Quality Alliance Medication Access Conceptual Frameworks, publicly available National Health and Nutrition Examination Survey datasets (2009-2018) were cross-sectionally analyzed among respondents aged 65 and older who were diagnosed with study diseases. Data analyses included descriptive statistics, and logistic regression using an alpha level of 0.05. Result: Analyses included 5513 respondents' data. Bivariate analysis revealed significant differences in medication adherence based on several structural (e.g., ethnicity) and intermediary (e.g., disability status) determinants of health. Multivariable analysis revealed significant differences in medication adherence for alcohol consumption (p = 0.034) and usual healthcare place (p = 0.001). Conclusions: The study findings underscore pertinent implications for public health and policy, with specific SDoH being the most likely to affect medication adherence in common chronic conditions among older adults in the US.