Causal inference in tobacco research: a public health challenge.

Abstract

Causal inference represents a rapidly expanding interdisciplinary subfield that involves various assumptions, study designs, and estimation strategies, allowing researchers to establish causal relationships from both clinical trials and observational data. In tobacco research, numerous studies address causal questions, including the contentious issue of whether vaping in nonsmoking youth leads to smoking initiation, known as the "gateway effect." Determining the effectiveness and safety of many health interventions will continue to rely on observational [mainly longitudinal] data because randomized trials are not always feasible, ethical, or timely. Therefore, review articles that are synthesizing evidence on the gateway effects of electronic nicotine delivery systems [ENDS] on subsequent cigarette smoking must also consider observational studies as first-rate evidence that can help bring together the polarized tobacco research community and help better understand the "gateway effect." In addition, this will help ongoing efforts to rigorously prevent ENDS use by youth while expanding the cessation potential of ENDS among adult established smokers who are unwilling to quit otherwise. In this commentary, we discuss causal inference tobacco research as one of the public health challenges and provide some recommendations/implications.