The Intersectionality of Factors Predicting Co-occurring Disorders: A Decision Tree Model

Abstract

Individuals with co-occurring psychiatric and substance use disorders (COD) face challenges, including accessing treatment, accurate diagnoses, and effective treatment for both disorders. This study aimed to develop a COD prediction model by examining the intersectionality of COD with race/ethnicity, age, gender identity, pandemic year, and behavioral health needs and strengths. Individuals aged 18 or older who participated in publicly funded behavioral health services (N = 22,629) were selected. Participants completed at least two Adult Needs and Strengths Assessments during 2019 and 2020, respectively. A chi-squared automatic interaction detection (CHAID) decision tree analysis was conducted to identify patterns that increased the likelihood of having COD. Among the decision tree analysis predictors, Involvement in Recovery emerged as the most critical factor influencing COD, with a predictor importance value (PIV) of 0.46. Other factors like Legal Involvement (PIV = 0.12), Decision-Making (PIV = 0.12), Parental/Caregiver Role (PIV = 0.11), Other Self-Harm (PIV = 0.10), and Criminal Behavior (PIV = 0.09) had progressively lower PIVs. Age, gender, race/ethnicity, and pandemic year did not show statistically significant associations with COD. The CHAID decision tree analysis provided insights into the dynamics of COD. It revealed that legal involvement played a crucial role in treatment engagement. Individuals with legal challenges were less likely to be involved in treatment. Individuals with COD displayed more complex behavioral health needs that significantly impaired their functioning compared to individuals with psychiatric disorders to inform the development of targeted interventions.