To examine the comparative effectiveness of MOUD using within-person fixed effects (case-crossover) models, we derived a subcohort of individuals from the analysis who experienced any drug-related poisoning, defined as emergency department (ED) visit or inpatient hospitalization and received any MOUD. This excluded 40 964 individuals who never received MOUD (naltrexone or buprenorphine) during insurance coverage. We excluded 125 831 individuals who did not experience a drug-related poisoning (outcome of interest) during coverage. We created an observation window (period during which an individual was assessed for drug exposure) limited to 1 year before and after a person’s first (index) event to reduce heterogeneity in observation time, culminating in 12 485 individuals encompassing 7 095 568 person-days.
