Fatal Drug Overdose Risks of Health Care Workers in the United States : A Population-Based Cohort Study
Mark Olfson,Candace M Cosgrove,Melanie M Wall,Carlos Blanco
DOI: https://doi.org/10.7326/M23-0902
Abstract:Background: Despite an unprecedented increase in drug overdose deaths in the United States, the risks faced by U.S. health care workers, who often have access to controlled prescription drugs, are not known. Objective: To estimate risks for drug overdose death among health care workers relative to non-health care workers. Design: Prospective cohort study. Setting: United States. Participants: Health care workers (n = 176 000) and non-health care workers (n = 1 662 000) aged 26 years or older surveyed in 2008 and followed for cause of death through 2019. Measurements: Age- and sex-standardized drug overdose deaths were determined for 6 health care worker groups (physicians, registered nurses, other treating or diagnosing health care workers, health technicians, health care support workers, and social or behavioral health workers) and non-health care workers. Adjusted drug overdose death hazards (and 95% CIs) were also evaluated, with adjustment for age, sex, race/ethnicity, marital status, education, income, urban or rural residence, and region. Results: Approximately 0.07% of our study sample died of a drug overdose during follow-up. Among health care workers, annual standardized rates of drug overdose death per 100 000 persons ranged from 2.3 (95% CI, 0 to 4.8) for physicians to 15.5 (CI, 9.8 to 21.2) for social or behavioral health workers. Compared with those for non-health care workers, the adjusted hazards of total drug overdose death were significantly increased for social or behavioral health workers (adjusted hazard ratio, 2.55 [CI, 1.74 to 3.73]), registered nurses (adjusted hazard ratio, 2.22 [CI, 1.57 to 3.13]), and health care support workers (adjusted hazard ratio, 1.60 [CI, 1.19 to 2.16]), but not for physicians (adjusted hazard ratio, 0.61 [CI, 0.19 to 1.93]), other treating or diagnosing health care workers (adjusted hazard ratio, 0.93 [CI, 0.44 to 1.95]), or health technicians (adjusted hazard ratio, 1.13 [CI, 0.75 to 1.68]). Results were generally similar for opioid-related overdose deaths and unintentional overdose deaths. Limitation: Unmeasured confounding, uncertain validity of cause of death, and one-time assessment of occupation. Conclusion: Registered nurses, social or behavioral health workers, and health care support workers were at increased risk for drug overdose death, suggesting the need to identify and intervene on those at high risk. Primary funding source: National Heart, Lung, and Blood Institute.