In a recent article, Yong et al.1 focus on whether the regulatory environment for e-cigarettes influences their effectiveness for smoking cessation. This topic is relevant given the rapid growth in e-cigarette use for cessation in some countries 2 and the polarized view within the public health community on their likely effectiveness. 3 However, we believe the data used in the present study to address whether regulatory environments influence the impact of e-cigarettes on cessation severely limit firm conclusions for at least three reasons.First, the measurement of e-cigarette use was only valid in one of the ten waves of the data used. Respondents were only queried about the quantity and frequency of their e-cigarette use in the 2013-2014 surveys. For the previous nine waves of data, e-cigarette use was not specifically measured, but respondents had the opportunity to add it as part of an open response to an "other product" question. Measuring e-cigarette use in this way could lead to underreporting or misclassification. Restriction to the portion of the sample with valid measures (2013-2014 waves) could address this issue; however, if the authors were to restrict their sample they would only have two replicates for their "regulatory environment" categories, as neither the United States nor Canada had any respondents who were in the 2013 cohort.Second, the analyses suffer from inadequate sample size, drawing into question the generalizability of the sample to the population they are purported to represent. For instance, there are only 50 respondents from either Canada or Australia who reported using an e-cigarette over the entire 11-year period.Third, the authors consider how the association of e-cigarette use with 30-day cigarette abstinence varies across countries categorized according to their regulatory environment (two countries for each option), but the validity of this proposed singular distinction has not been demonstrated. The authors do suggest one possibility for how the regulatory environment might modify the relationship between e-cigarette use and cessation is a greater ability for respondents to experiment with e-cigarette products in an environment that is less restrictive of e-cigarette use. Yet, the authors have not directly assessed this hypothesis empirically in their analysis. Any observed differences across the four countries could depend on many other factors. In other settings, such as the impact of medical marijuana laws on marijuana usage, it has been shown that accounting for measured and unmeasured confounding factors (eg, social norms about drug use) constant over time through the use of fixed effect models can affect a study's conclusions. 4 In conclusion, there are at least three limitations in this article that severely temper the conclusions reached by the authors and, in our view, cannot be addressed by the supporting data. Given the importance of the research question, it is equally important that firm conclusions be generated from appropriate data.
FundingNone declared. ...