Although total cessation of nicotine and tobacco products would be most beneficial to improve public health, exclusive e-cigarette use has potential health benefits for smokers compared to cigarette smoking. This study investigated differences between dual users and exclusive e-cigarette users provide information to optimize health communication about smoking and vaping. A cross-sectional survey (n = 116) among 80 current, adult dual users and 36 current, adult-exclusive e-cigarette users was conducted in the Netherlands. The questionnaire assessed four clusters of factors: (1) Past and current smoking and vaping behavior, (2) product characteristics used, (3) attractiveness and reasons related to cigarettes and e-cigarettes, and (4) socio-cognitive factors regarding smoking, vaping, and not smoking or vaping. We used random forest-a machine learning algorithm-to identify distinguishing features between dual users and e-cigarette users. We are able to discern a dual user from an exclusive e-cigarette user with 86.2% accuracy based on seven factors: Social ties with other smokers, quantity of tobacco cigarettes smoked in the past (e-cigarette users) or currently (dual users), self-efficacy to not vape and smoke, unattractiveness of cigarettes, attitude towards e-cigarettes, barriers: accessibility of e-cigarettes, and intention to quit vaping (A). This combination of features provides information on how to improve health communication about smoking and vaping.
Keywords: attractiveness; dual use; e-cigarettes; machine learning; public health; random forest; smoking behavior; socio-cognitive factors.