High-Risk Medication Prescription Receipt in the Military Health System: Elucidating Patterns and Substance Use Outcomes among Military Service Members - Summary/Abstract In 2015, high-risk medication (HRM) receipt was identified by the Army Surgeon General as requiring programmatic mitigation and monitoring efforts due to its potential impact on quality of life, mortality, and military separation. Soon after, the Defense Health Agency (DHA) adopted and disseminated the following HRM criteria: (1) ≥1 opioid prescription and ≥3 prescriptions for other medications in the past 30 days; (2) ≥3 emergency department visits each involving opioids within a 12-month period, with ≥1 visit in the past 30 days; and (3) ≥4 psychotropic prescriptions in the past 30 days from the following medication classes: opioids, antidepressants, anxiolytics, sleep medications, anticonvulsants, stimulants, and neuroleptics/major tranquilizers. Historical research has focused primarily on prescription opioids alone; one study identified HRM criteria prevalence in 2016 among active duty service members and found it was also associated with 12- month incidence of suicide ideation. No study has evaluated the longitudinal patterns of HRM receipt, the predictive qualities (e.g., discrimination, calibration, performance) of the DHA-HRM criteria, or predictive fairness. As a result, targeted, data-driven policy and programming efforts are stymied. Therefore, this study will leverage the pre-existing Substance Use and Psychological Injury Combat (SUPIC) Study Next-Generation cohort, which includes comprehensive longitudinal healthcare records from over 1.66 million active duty, National Guard, and Reserve service members from 2016-2021 to address these questions. The investigators from Brandeis University, Boston University School of Public Health, and the Uniformed Services University have had extensive experience in evaluating longitudinal medical and administrative records from prior SUPIC studies. The Specific Aims of this study are to: (1) identify the prevalence, temporal patterns and potential variation of HRM receipt; and describe HRM composition (e.g., classes, combinations) per DHA-HRM criteria; and (2) evaluate whether HRM criteria based on data-driven, enhanced, and granular prescription pattern measures can more accurately and fairly identify service members who experience substance use outcomes relative to current DHA-HRM criteria. Results from this study will inform future data-driven risk mitigation policies and programming, especially those regarding prescription monitoring tools and requirements.
