A random forest analysis identifying important predictors and risk factors among those who reported use of...

According to the Centers for Disease Control and Prevention (CDC), in 2023 alone, there were approximately 35,000 deaths that involved methamphetamine. Because methamphetamine is commonly used with alcohol, our objective was to investigate risk factors for concurrent use, which are underexplored.

This was a secondary analysis of deidentified publicly available data (2021-2023) collected by the National Survey of Drug Use and Health (NSDUH). Our primary aim was to identify risk factors for concurrent use of methamphetamine and alcohol. We used the random forest method and conducted variable importance analyses to identify the most important predictors for concurrent use, and subsequently binomial logistic regression for identifying risk factors.

Between 2021 and 2023, there were a total of 219 (0.13% of all respondents) respondents (449,590 respondents weighted) that reported concurrent use of methamphetamine and alcohol. The top three predictors per the random forest model were (reflective of past year behaviors): driving under the influence of a selected illicit drug (Gini: 0.08), emergency department visit for substance use (Gini: 0.05), and fentanyl misuse (Gini: 0.046). Of the top 10 most important predictors, only 3 were statistically significant: income (p < 0.001), driving under the influence of any drug (aOR: 18.6, p < 0.0001); and perceived health status (p = 0.02).

Our study identified several important predictors and risk factors for concurrent use of methamphetamine and alcohol. Due to our study's low rate for the outcome of interest, future studies should further investigate risk factors and targeted interventions should be put into place to mitigate risky behaviors.