Exploring Polysubstance Use with a Data Mining Approach in Individuals with and without Mental Health Conditions

Polysubstance use is widespread among individuals who use drugs in the U.S. and can result in serious consequences. This study aims to understand frequent polysubstance use patterns among U.S. adults and compare these patterns between individuals with and without mental health (MH) conditions. Using data from the National Survey on Drug Use and Health for 2019-2022, we identified individuals who used one of ten substances in the past year. We then applied association analysis to extract co-use patterns of substances and estimate the strength of these associations. More than 42% of individuals reported using more than one substance in the past year. The top 10 polysubstance use patterns were largely consistent between individuals with and without MH conditions, although the prevalence and association of these patterns were higher among the MH group. In both groups, the most frequently co-used substances were marijuana and prescription opioids. Among the five illicit substances examined, heroin showed the highest conditional probability of co-use with cocaine or methamphetamine. This study demonstrated that there are similarities and differences in complex polysubstance use patterns between individuals with and without MH conditions. These findings can inform the development of targeted interventions tailored to particular substances and at-risk populations.