Clinical Decision Support and the Naloxone Equity Exception: Insights from an Emergency Department.

INTRODUCTION: Despite rising overdose mortality nationwide, racial and ethnic disparities in naloxone access persist. This study examined whether these disparities remained after adjusting for demographic and clinical risk factors, including an electronic health record-based Clinical Decision Support alert, within an emergency department setting. METHODS: A retrospective cross-sectional study of 10,313 emergency department encounters was conducted over a 2-year period. Multivariable logistic regression was used to assess the association between race or ethnicity and naloxone distribution during the emergency department encounter. The model adjusted for age, illness severity, nursing-documented intravenous drug use risk, and daily morphine milligram equivalents. RESULTS: Documented intravenous drug use risk emerged as the strongest predictor, with patients with this risk having significantly higher odds of naloxone distribution (odds ratio, 1.84; 95% confidence interval [CI], 1.67-2.03). Clinical Decision Support alert activation was also associated with increased odds of distribution (odds ratio, 1.14; 95% CI, 1.04-1.25), indicating that the alert contributed modestly when activated. Minor racial and ethnic differences were observed. Black/African American patients had higher odds of naloxone distribution compared with white patients (odds ratio, 1.08; 95% CI, 0.90-1.30). In contrast, increasing age (odds ratio, 0.99; 95% CI, 0.99-0.99) and higher Emergency Severity Index scores (odds ratio, 0.82; 95% CI, 0.77-0.88) were associated with reduced odds of naloxone distribution. DISCUSSION: Integrated Clinical Decision Support may promote more consistent naloxone distribution, though this study could not determine whether the observed patterns reflected equitable access or other clinical and contextual influences. Strengthening early medication reconciliation and embedding reliable risk indicators into triage workflows may improve the reliability of Clinical Decision Support reliability.