Data were analyzed using PROC SURVEYLOGISTIC in SAS, version 9.4 (SAS v9.4, SAS Institute Inc., Cary, NC). This procedure allowed for incorporating the stratification, clustering (i.e., primary sampling unit), and unequal weighting of the sampling design. Binary logistic regression analysis was used to examine associations between stressful life events (zero or one event vs. two or more events) and gender with transitions in DSM-5 SUD diagnoses (AUD, TUD, CUD, and OUD; new vs. absent and ongoing vs. remitted). Relationships between stressful life events and gender were assessed in terms of odds ratios and were considered significant at p < 0.001. The effects of each variable of interest on any given outcome were interpreted relative to our chosen reference outcome (i.e., zero or one stressful life event, male). Two-way interactions between stressful life events and gender for new vs. absent SUD diagnoses and ongoing vs. remitted SUD diagnoses were performed to investigate whether stress and gender were associated with transitions in DSM-5 SUD diagnoses. Age, race, income, and education were evaluated as potential covariates and were removed from the final models if there was no impact on the pattern of results.
