Disability Impacts on Opioid Use Treatment

Data analysis was performed from January to September 2022. Analyses were conducted by person-year. After conducting descriptive analyses and χ² tests, we fit generalized estimating equations with robust SEs assuming an exchangeable correlation structure using SAS statistical software version 9.4 (SAS Institute) to account for correlated outcomes of individuals in the data for multiple years. Two-tailed P < .05 was considered to denote statistical significance. Models adjusted for the confounding variables of age, gender, race and ethnicity (identified via database; race and ethnicity were included to isolate their associations with disability to the extent possible), urban residence, SUD other than OUD in the year, mental disorder in the year, eligibility year, and living in an institution for at least 2 months of the year, to be consistent with the literature. Model 1 included any disability as a dichotomous independent variable. Model 2 included each type of disability, also dichotomous with each compared with persons without that disability, as some enrollees have more than 1 type. Full models are in eTables 4, 5, 6, and 7 in Supplement 1. We also modeled type of medication by disability status and type among those receiving buprenorphine or methadone.