Improving treatment completion for young adults with substance use disorder: Machine learning-based prediction algorithms.

Substance use disorder (SUD) treatment completion was intertwined with various factors. However, few studies have explored the intersections of psychosocial and system-related factors with SUD treatment completion, particularly for individuals receiving publicly funded SUD treatment services. This study aimed to examine the intersections of these factors with treatment completion. We analyzed the psycho-social assessment data of 2909 young adults who participated in publicly funded outpatient-based substance use treatments in 2021. The Chi-square Automatic Interaction Detection (CHAID) approach was employed to examine intersections for SUD treatment completion. The analysis highlights the significance of multiple factors and their interactions in predicting SUD treatment completion. The results indicate that SUD treatment outcomes varied based on the level of improvement rates in total actionable items (TAI) improvement rates, underscoring the importance of monitoring individual progress in treatment. Specifically, among young adults with the highest TAI, those residing in rural communities were less likely to complete treatment compared to their urban counterparts. For individuals with TAI improvement rates at the middle level, there was a significant intersection with criminal justice involvement. Within this subgroup, individuals who had both justice system involvement and opioid use disorders had a relatively low SUD treatment completion rate, while those with non-opioid-related SUD exhibited a higher completion rate. The study illustrates the importance of considering multiple factors and their interactions, including TAI improvement rates, family strengths, demographic characteristics, and social determinants, in predicting SUD treatment completion among young adults.