Predictors of co-occurring mental and substance use disorders among clients admitted for substance use treatment

Abstract

Listen

Translate article

Purpose This study aims to determine the predictors of co-occurring mental health and substance use disorders (CMSUD) among patients admitted for substance use disorder treatment. Design/methodology/approach Analysis was done on the 2020 national Treatment Episode Data Set for Admissions (TEDS-A). Among 1,416,357 records, 636,522 with complete client-level data were randomly split into training and validation data sets for cross validation. Findings There were 445,566 and 190,956 records of admissions in the training and validation data, respectively. The predictive model was validated (training area under the curve [AUC] 0.6508, confidence interval [CI] 0.6492–0.6524; validation AUC 0.6522, CI 0.6498–0.6547) and the prediction was accurate (Brier Score 0.23). Those more likely to have co-occurring mental and substance use disorders included younger people compared to those ≥50 y/o, female, being arrested at least twice, unemployment, less substance use in past month, age at first substance use ≤20 years, having had a spouse, having veteran status, being homeless, higher substance use treatment episodes and history of medication-assisted opioid use disorder treatment. Research limitations/implications The lower likelihood of CMSUD in Hispanic, Black and individuals with lower education may be due to inadequate access to healthcare services. Originality/value The identification of predictors of CMSUD will help in the development of efficient screening measures and implementation of preventive and early interventions. Robust statistical analysis was utilized in this research. Despite the exclusion of numerous records with missing data, a substantial study population was still included in the study.