Toward Fairness in Machine Learning Models for Predicting Treatment Retention and Premature Discontinuation in Medication for Opioid Use Disorder

Persistent low retention and completion rates in medications for opioid use disorder (MOUD) have driven the use of machine learning (ML) models to predict retention and identify patients at risk of premature discontinuation. However, the fairness of these models across patient populations remains largely unexplored, raising concerns about their application in treatment decision support. This study systematically assesses algorithmic fairness in ML models for predicting MOUD retention and premature discontinuation and investigates the effectiveness of bias mitigation techniques. Using the cross-sectional Treatment Episode Data Set-Discharges (TEDS-D), which includes treatment episodes for individuals in the U.S. discharged between 2015 and 2019, we trained four ML models to predict premature treatment discontinuation and retention beyond 180 days among individuals receiving outpatient MOUD. We evaluated overall performance and subgroup-level error rates across patient subgroups defined by race, ethnicity, age, and sex, complemented by model explanation analyses. We further assessed pre-processing, in-processing, and post-processing bias mitigation techniques and their effects on both fairness and predictive performance. The models exhibited substantial performance differences across patient subgroups, including overestimation of the likelihood of premature discontinuation for Black patients and of treatment retention beyond 180 days for older patients. Model explanation analyses further identified race and age as influential predictors, but their impacts on model predictions varied substantially across patient subgroups. Bias mitigation strategies reduced specific fairness gaps but often introduced trade-offs, such as increased error rates for other subgroups or reductions in overall predictive performance. These findings demonstrate that ML models for MOUD outcome prediction can exhibit subgroup-level performance gaps even when overall predictive performance appears acceptable and that bias mitigation can reduce, but not fully eliminate, these gaps without trade-offs. By demonstrating the importance of fairness-aware evaluation and transparent reporting of subgroup performance, this study provides practical insights for the responsible and context-sensitive use of ML models for risk stratification and care prioritization in MOUD treatment settings.