Evaluating the Generalizability of Substance Use Treatment Outcomes at One-Month Follow-Ups Using Data From a Random Sample of Discharged Patients at a Single US-Based Treatment Center

There is a dearth of research that evaluates the generalizability of substance use disorder treatment outcomes. We assess one-month follow-up survey results and evaluate generalizability of outcomes for individuals randomly selected after discharge from a non-profit treatment center. The sample consisted of discharged patients (N = 6771). Follow-up measures were: alcohol/drug abstinence, self-reported health, quality of life, peer-recovery group attendance, follow-up treatment receipt, medication use, and craving. Descriptive and logistic regression analyses compared patients who were randomly selected to those not selected, and follow-up respondents to non-respondents; 10 measures (demographic, treatment- and diagnosis-related measures) were used to assess selection and nonresponse. A survey weight was generated that accounted for selection, and response/attrition. Outcomes were estimated using weights, and these estimates were compared to unweighted values. Weighted abstinence rate one month after treatment was 76.4% (95% CI = 74.4%, 78.4%); unweighted rate was 78.2% (95% CI = 76.4%, 79.9%). Eight out of the 10 measures used to assess survey response significantly predicted response. All 6 outcome measures were statistically similar across weighted and unweighted analyses. One-month follow-up results were representative of daily discharged population at a single US-based treatment center. Centers should use sampling designs that ensure generalizability and provide full methodological details when reporting follow-up measures.