Machine Learning of Smoking Relapse: the Role of Racial Differences and E-Cigarette Vaping Characteristics on Former Smokers.

BACKGROUND: Machine learning models can help identify multifaceted factors influencing tobacco use transitions. A random forest model is developed to predict smoking relapse, focusing on racial differences and vaping characteristics. METHODS: Data are drawn from the Population Assessment of Tobacco and Health (PATH) study adult interview files. Former combustible cigarette smokers at baseline (Wave 5) were followed up 1 year later (Wave 6). Predictors (n = 100) include a wide range of social demographics, psychosocial factors, health status, tobacco and substance use behaviors, and vaping characteristics. RESULTS: Among 4693 former smokers at baseline, 4.4% relapsed to smoking within 4 years. Random forest models achieved high prediction accuracies across racial groups, with area under the curve (AUCs) of 0.77 for Whites, 0.88 for Blacks, and 0.70 for Hispanics. Quit history (i.e., recent vs. long-term quitters) was one of the top predictors across all racial and ethnic groups. Tobacco addiction was one of the top predictors among White and Hispanic former smokers but not among their Black and other race counterparts. Marijuana use was one of the top predictors for Blacks but not for other racial and ethnic individuals. Vaping status predicted relapse across all groups, but the importance of vaping characteristics differed. E-cigarette nicotine concentration levels and e-cigarette devices ranked higher for Whites and Hispanics than for Blacks and Others. CONCLUSIONS: The findings reveal notable racial differences in smoking relapse predictors, along with distinct roles of vaping characteristics across racial groups. Unique social, behavioral, and health factors are crucial for improving smoking cessation outcomes.