E-cigarettes and Cessation: Asking Different Questions Requires Different Methods

We appreciate the comments we received in response to our analysis of the relationship between e-cigarette use and cigarette smoking cessation using the Population Assessment of Tobacco and Health (PATH) Study.^1,2 Pierce et al. noted that our findings diverged from those in a prior study on e-cigarette use, which used a different subsample of PATH participants.³ Indeed, differences in inclusion and exclusion criteria can result in dissimilar findings, but cohort selection in observational studies necessarily shifts depending on the questions being asked. Importantly, the study by Pierce et al. was framed as an intervention study; thus, it examined smoking cessation and use of e-cigarettes in one’s last quit attempt only among those who reported having made a quit attempt in the past year. In that defined subsample, the authors found that e-cigarettes were equally as effective as pharmacotherapy and no product assistance.

Our study attempted to answer a broader question: the impact of e-cigarettes on cigarette smoking cessation among the full sample of smokers in the PATH data set. Unlike nicotine replacement therapy and other pharmacologic cessation aids, e-cigarettes are consumer products that are widely available and used for a variety of reasons. As such, their effects on patterns of cigarette use are not limited to their deliberate use as smoking cessation aids. For example, some people who smoke may take up e-cigarettes as an occasional alternative to smoking—not as a quitting tool—and gradually migrate to cigarette abstinence over time. E-cigarette and smoking cessation studies that only include those who make a serious quit attempt also introduce bias by not considering this real-world context. Although mimicking randomized controlled clinical trials offers an important perspective,⁴ observational studies conducted on broader populations are also imperative to determine the impact of lower-risk nicotine substitutes—such as e-cigarettes—on priority populations of combusted tobacco users (e.g., those not interested in quitting cigarettes).

Pierce et al. also indicated that we did not adhere to the recommended guidelines from the National Academies of Sciences, Engineering and Medicine (NASEM).⁵ These guidelines focus more narrowly on analyses of e-cigarettes as an intervention for smoking cessation, and they were cited out of context given the purpose of our study. The NASEM report also asserts that the public health impact of e-cigarettes is both a function of their efficacy for smoking cessation, and more broadly, their reach as an appealing consumer product. Our study’s design incorporated these dimensions, thereby offering a more complete picture of e-cigarette use and its relationship to smoking patterns outside of a treatment paradigm.

The primary concerns raised by Pierce et al.—namely, that the results from our study diverge from and are more biased than the study by Pierce et al.—are misguided. Each study answers different questions, and as with any observational study, both measured and unmeasured biases are unavoidable and limit the findings. The two studies differ in a number of ways. First, we did not limit our inclusion criteria to a narrow subsample of smokers to mimic an intervention trial. Although propensity score matching techniques can be useful to construct artificial comparison groups, they do not rule out inherent biases from utilizing strict inclusion criteria. Second, we included daily and nondaily smokers to examine variation in smoking cessation based on the frequency of smoking. Third, we included additional covariates in our models and defined our cessation outcome differently to assess short- and long-term quitting. The issue regarding temporality of the exposure and outcome is a limitation we acknowledged in our article. Although some individuals may have started using e-cigarettes after quitting smoking, documenting even a temporally imprecise relationship between e-cigarette use and smoking cessation is critical, as it captures realistic trajectories related to harm reduction behaviors that influence population health. For example, some individuals may begin using e-cigarettes to avoid relapse to smoking.

A key strength of our study, which was not present in the study by Pierce et al., was our attention to key criteria established by Villanti et al. to improve the methodological rigor of observational studies examining the impact of e-cigarettes—specifically, assessment of the dose, duration of exposure, type, and quality of the e-cigarette products used.⁶ When compared with similarly designed observational studies using PATH and other national data sets, as well as the recently updated Cochrane Systematic Review,⁷ our findings are consistent: frequent and continuous use of e-cigarettes is associated with increased smoking cessation.^8,9 Multiple study designs are needed to triangulate findings and determine how the data converge to inform effective treatments and policies. Data that reflect real-world use of e-cigarettes can also be used to inform model assumptions when projecting the population impact of e-cigarettes in simulation models.

Although we agree with Pierce et al. need for rigorously defined studies, we believe our study only adds to the range of studies available on this topic and offers a meaningful and more comprehensive perspective than the studies that he has cited. To inform policy and interventions, we need evidence from studies of real-world use in addition to ideal, cessation-focused conditions. Notably, our article does not necessarily contradict Pierce et al.’s finding that e-cigarettes were equally as effective as nicotine replacement therapy and quitting without product assistance. Rather, by incorporating a broader sample of individuals who smoke, our study adds new information to these results that can more fully inform the Food and Drug Administration’s public health impact criterion.

Supplementary Material

A Contributorship Form detailing each author’s specific involvement with this content, as well as any supplementary data, are available online at https://academic.oup.com/ntr.

Declaration of Interests

None declared.