The popularity of electronic cigarettes (e-cigarettes) in the United States has increased dramatically in recent years. Less than 1% of U.S. adults reported ever trying e-cigarettes in 2009 [1
]; in 2013, 15% reported ever trying them [2
]. Use of e-cigarettes by smokers is particularly high. In a 2013 national U.S. survey, half of smokers reported ever trying e-cigarettes, and 21% reported currently using them some days or every day [3
]. Use by smokers is similarly high in other countries [4
]. For instance, more than one-third (37%) of current smokers and recent ex-smokers in Great Britain reported ever use of e-cigarettes in 2012, and 21% currently used them [6
Among the most common reasons for using e-cigarettes are believing they are healthier than regular cigarettes and can help smokers quit or reduce smoking [7
]. Because e-cigarettes do not rely on combustion or contain cured tobacco, they produce far fewer harmful constituents than regular cigarettes [8
]. However, some studies of e-cigarette aerosol and e-liquid have detected formaldehyde, nitrosamines, and other harmful constituents in amounts that varied considerably across brands, flavors, and models [9
]. The extent to which e-cigarettes help smokers quit is still unclear [12
]. One randomized-controlled trial found similar cessation success rates for e-cigarettes and the nicotine patch [15
]. Some population-based studies indicate that e-cigarettes can help smokers quit [16
], while others do not [17
]. Other reported reasons for trying or using e-cigarettes are avoiding smoking restrictions [19
], alleviating cravings for nicotine [20
], enjoying the flavors [23
], and preventing others’ exposure to cigarette smoke [21
There is little research on stopping use of e-cigarettes outside of surveys with purposeful sampling of e-cigarette users from online forums [20
] or smokers exiting tobacco shops [24
]. It is important to understand why people stop using e-cigarettes for multiple reasons. First, e-cigarettes may provide an opportunity for harm reduction should nicotine-addicted smokers be willing to switch from regular cigarettes to non-combustible electronic ones. Only if switching is both successful and widespread would their use offer a net public health benefit. Second, understanding who stops using e-cigarettes and why allows researchers to distinguish among user types. Just as there are subtypes of cigarette smokers, like “social smokers” [25
], subtypes of e-cigarette users may also exist. Reaching the right users with the right messages is critical for the success of any future public health campaigns focused on e-cigarettes. The aim of the present study was to investigate reasons for starting and stopping use of e-cigarettes and to examine differences in discontinuation by reason for trying among a large, population-based sample of U.S. adults.
This survey relied on data collected as part of the Tobacco Control in a Rapidly Changing Media Environment (TCME) study. The TCME project examined the relationship between recall of receiving tobacco-related information through multiple media channels and tobacco-related attitudes, beliefs, and behavior. In March 2013, 17,522 U.S. adult TCME participants completed an online survey. Most respondents (75%) were members of KnowledgePanel, a nationally representative online survey panel constructed using random-digit dialing, supplemented by address-based sampling to capture cell phone-only households. The remainder of the sample constituted members of a separate consumer survey panel recruited through online ads for joining the panel. The survey company screened individuals who were part of this consumer sample based on demographics and tobacco use behaviors and issued invitations to this survey in order to quota match with the KnowledgePanel sample. KnowledgePanel members received monthly compensation for being part of the panel (usually as points redeemable for $4–6 of goods) as well as entry into a sweepstakes with an additional cash or prize reward for completing this specific survey. Respondents from the separate consumer panel received points worth up to $2 for taking this survey. Before taking the survey, all participants provided consent online. Of the 34,097 KnowledgePanel members surveyed, 61% completed the screening; among those who were eligible (n = 13,531), 97% completed the survey. Response rates for the convenience sample cannot be calculated because there is no known sampling frame. For this study, we report data only from the 3878 participants who reported ever having tried an e-cigarette. Institutional review boards at the University of Illinois and the National Cancer Institute approved the study.
While viewing generic images of e-cigarettes, participants read an introductory statement: “The next questions are about electronic cigarettes, often called e-cigarettes. An e-cigarette looks like a regular cigarette, but it runs on a battery and produces vapor instead of smoke. There are many types of e-cigarettes. Some common brands are Smoking Everywhere, NJOY, Blu, and Vapor King. Below are some pictures of e-cigarettes.” Participants who were aware of e-cigarettes (assessed with the item, “Before today, had you ever heard of e-cigarettes?”) answered items about ever use (“Have you ever used an e-cigarette, even one puff?”) and current use (“Do you now use e-cigarettes every day, some days, or not at all?”). We defined “trying” or “starting” as ever use and “current use” as using e-cigarettes either every day or some days. We defined “stopping” or “discontinuing” as ever use without current use.
One survey item assessed reasons for trying e-cigarettes: “What are the reasons you first tried e-cigarettes? (check all that apply).” The response options were: they are affordable; I can use them in places where smoking cigarettes isn’t allowed; they might be less harmful to me than regular cigarettes; they might be less harmful to people around me than regular cigarettes; e-cigarettes come in flavors I like; e-cigarettes can help me quit or cut back on smoking regular cigarettes; e-cigarettes don’t smell bad; using an e-cigarette feels like smoking a regular cigarette; e-cigarettes don’t bother people who don’t use tobacco; the advertising for e-cigarettes appeals to me; they help me deal with cravings to smoke; I have a friend or family member who uses e-cigarettes; I was curious about e-cigarettes; I received an e-cigarette as a holiday gift; I made a New Year’s resolution to quit smoking regular cigarettes; and other (please specify). We recoded open-ended responses that clearly fit into one of the available response options. For example, we recoded a response of “So I can quit smoking” into the category “e-cigarettes can help me quit or cut back on smoking regular cigarettes”.
Based on the responses to the open-ended item, we created an omnibus category for the reason “a friend or family member used, gave, or offered e-cigarettes”; this category included positive responses for “I have a friend or family member who uses e-cigarettes”, “I received an e-cigarette as a holiday gift”, and any open-ended responses mentioning a gift or offer from a friend or family member. We also asked respondents to select the main reason they tried e-cigarettes out of all of the reasons they selected on the previous item. We categorized the main reason as “non-goal-oriented” if respondents said they tried e-cigarettes because of curiosity, a friend or family member, liking the ads, other (non-recoded responses), or said they did not know. We classified the remaining main reasons (e.g., stopping or reducing smoking, using where smoking is not allowed) as “goal-oriented”.
Another survey item assessed reasons for stopping e-cigarette use, among those who reported no longer using them: “What are the reasons you stopped using e-cigarettes? (check all that apply).” Response options were: e-cigarettes cost too much money; I didn’t like how they tasted; using e-cigarettes didn’t help me deal with cravings to smoke; using e-cigarettes didn’t help me quit or cut back on smoking regular cigarettes; using e-cigarettes didn’t feel like smoking regular cigarettes; e-cigarettes are poor quality, defective, or break easily; I was concerned about the health risks caused by using e-cigarettes; I didn’t like the side effects of using them; I was just experimenting with e-cigarettes; and other (please specify). Again, we recoded open-ended responses that clearly fit into one of the available response options.
The survey also assessed use of regular cigarettes, gender, age, education, race and ethnicity, marital status, employment, region of residence, and household income. We classified “non-smokers” as those who had smoked fewer than 100 cigarettes in their lifetimes and “former smokers” as those who smoked 100 or more cigarettes but did not currently smoke. Current daily and non-daily smokers received a question about their intentions to quit (“Do you plan to quit smoking for good…?” with six response options: in the next 7 days, in the next 30 days, in the next 6 months, in the next year, more than one year from now, or I do not plan to quit smoking for good). To assess understanding and appropriateness of item wording and ease of responding to survey items, we conducted cognitive interviews and refined the list of reasons for starting and stopping use of e-cigarettes based on this feedback. We pre-tested the revised survey with 160 respondents. For all variables, we recoded missing scores (<0.5% for each item) to the mean or mode of that item.
2.3. Data Analysis
We examined bivariate associations between respondent characteristics and the three most common reasons for trying e-cigarettes using logistic regression. We included all statistically significant bivariate correlates (p < 0.05) in a multivariate model. We used the same procedure to examine associations between respondent characteristics and stopping use of e-cigarettes. To compare rates of discontinuation among those who tried e-cigarettes for goal-oriented vs. non-goal-oriented reasons, we conducted an additional logistic regression with discontinuation as the dependent variable and reason type (goal-oriented vs. non-goal-oriented) as the independent variable. Analyses were run in Stata Version 12. Frequencies are unweighted. Percentages and all other analyses used the “svy” command and post-stratification weights to adjust for the representativeness of the sample compared to the U.S. population and the sampling design, including the combination of probability and non-probability samples. Statistical tests were two-tailed with a critical alpha of 0.05.
In this large national study of U.S. adults who had ever used e-cigarettes, the most common reasons for trying e-cigarettes were curiosity, because a friend or family member used, gave, or offered them, and quitting or reducing smoking. Two in three people who tried e-cigarettes later stopped using them, but this pattern differed by reason for trial. Those whose main reason for trying was goal-oriented had much higher rates of continued use than those who tried for non-goal-oriented reasons like curiosity or the influence of a friend or family member. This pattern suggests there may be distinct subgroups of e-cigarette users with differing experiences of use.
Curiosity, the most common reason for e-cigarette trial, is a natural reason for experimentation. E-cigarettes are a novel product, only introduced in the U.S. market in 2007 [26
]. They are widely advertised [27
] and covered in the popular media [28
], both of which help to spread information about and promote experimentation with any innovation [31
]. Individuals with higher education tried e-cigarettes because of curiosity more often than those with lower education. Higher education individuals in general are more likely to adopt innovations [31
]. In addition, experimentation might be easier in this group because they have greater access to e-cigarettes, which are more available in high socioeconomic status neighborhoods than low socioeconomic status neighborhoods [32
]. E-cigarette users also frequently cited the influence of friends or family as a reason for trying the product. Social connections motivate the diffusion of most innovations throughout a population [31
]. Smoking behaviors in particular spread through social contacts, partly because similar people gather together in social or family groups [33
]. Women, who tend to have larger and denser family networks [35
] and more intimate friendships [36
] than men, were more likely to report the influence of friends or family as a reason for trying e-cigarettes.
In the e-cigarette literature, quitting or reducing smoking is a frequently documented reason for use [7
], as we found in this study. Not surprisingly, smokers were more likely than non-smokers to report this as a motivator for trial. Three participants who reported trying e-cigarettes in order to quit smoking met our definition of non-smokers (i.e.
, smoked less than 100 cigarettes in their lifetime). It is possible that these participants sought to use e-cigarettes in order to curtail their smoking before it became a regular habit. The most common reasons for trial in our study, curiosity and social influence, are less commonly reported in the literature, perhaps because researchers did not allow for these answer choices in closed-ended survey items or because these factors motivated initial trying e-cigarettes but not continued use of e-cigarettes, so they would not be mentioned when asked about reasons for use.
Most respondents who tried e-cigarettes for goal-oriented reasons like smoking cessation or the ability to use e-cigarettes in places where they could not smoke regular cigarettes continued to use them, while those who started for non-goal-oriented reasons like curiosity later stopped using them. We suspect that those who try e-cigarette use for non-goal-oriented reasons have no intention to use them on a regular basis, perhaps because they do not personally identify with the image in their mind of the typical e-cigarette user. Individuals who have positive prototypes of cigarette smokers and identify with those prototypes are more susceptible to smoking [37
] and more likely to relapse after quitting [38
], and holding more negative beliefs about a typical cigarette smoker is associated with lower interest in trying e-cigarettes [39
]. In contrast, only about 20% of those who tried e-cigarettes for goal-directed reasons, most commonly smoking cessation and harm reduction, discontinued use. Value expectancy models of health behavior, such as the Health Belief Model [40
], focus on rational, goal-directed behavior. These models describe willingness to engage in health-protective behavior as related to the desire to avoid illness and the belief that this behavior would prevent that illness. Individuals who try e-cigarettes for smoking cessation or harm reduction (the two most common goal-oriented reasons in this study) may be explicitly or implicitly trying to reduce their chances of developing a smoking related-illness. Among this group, those who then discontinue might do so because the product did not help them achieve their goals (e.g.
, smoking cessation) or because negative experiences (e.g.
, bad taste) overrode their goal. Those who continue to use might feel that e-cigarettes are helping them to smoke less or they enjoy the experience of using e-cigarettes once having tried them for cessation reasons.
In general, discontinuation was less common among smokers and more common among those with higher education. This pattern likely reflects the differences in reasons for trial. Individuals with higher education are more likely to try e-cigarettes because of curiosity and then discontinue because they were merely experimenting. Smokers were more likely to try e-cigarettes as a means to quit smoking and then continue to use them for the goal-related reasons described above. After experimentation, the next most common reasons for stopping use were that e-cigarettes did not feel like regular tobacco cigarettes and users did not like their taste. Most e-cigarette users are smokers [7
]. If they are interested in electronic cigarettes as a direct substitute for regular ones, they may be disappointed. Some smokers say that e-cigarettes do not feel like or taste like the “real thing” [22
]. Indeed, one frequent response in the open-ended field was that e-cigarettes were “too heavy,” presumably in comparison to the weight of regular cigarettes that smokers are accustomed to. In the meantime, e-cigarette technology and design are improving (e.g.
, stronger batteries, options for user customization), and this trend will likely continue as multinational tobacco corporations invest in the e-cigarette market. Newer models might be more appealing to smokers, and dissatisfaction with feel, taste, or cost may decline. If newer models deliver nicotine more efficiently, fewer smokers might discontinue use because of failure to help manage their cravings. Better products might also be better cessation tools, reducing the number of smokers who stop using because the product fails to help them quit. Alternatively, there may be more smokers who stop using the product because they used it to successfully help them quit nicotine altogether. In the present study, only a small number of respondents reported discontinuing use for that reason.
The present results suggest two distinct groups of e-cigarette users: casual experimenters who almost always stop using and motivated users whose discontinuation decisions relate to the utility or experience of using e-cigarettes. This distinction has implications for research, specifically measurement. Researchers typically report and examine correlates of “ever use” of e-cigarettes. However, “ever users” appear to comprise two different groups. Finer-grained measurement is needed. To distinguish these groups, researchers might assess the number of instances of previous use and rely on a cut-off point for defining ever use as is currently done for established smoking, which is typically defined as having smoked at least 100 cigarettes in a lifetime.
The distinction between casual experimenters and motivated users also has implications for public health practice. If research can identify why certain vulnerable groups are attracted to and start using e-cigarettes (e.g., youth find the candy and fruit flavors appealing), public health campaigns can deliver more effective counter-messages when appropriate. In this sample, the most common reason for trying e-cigarettes was the same among smokers and non-smokers: curiosity. Reducing the appeal of e-cigarettes (e.g., restricting advertising that makes e-cigarettes seem glamorous) or making the product more difficult to obtain might deter non-smokers from trying the product out of mere curiosity.
Ultimately, the extent to which practitioners treat these two groups of e-cigarette users differently will likely depend on whether studies find that e-cigarettes can help smokers quit. If e-cigarettes do not help smokers quit or have other detrimental effects on public health (e.g., lead to smoking initiation among non-smokers), distinguishing casual experimenters from motivated users could be helpful, as these groups might need different strategies for discouraging use. Alternatively, if future studies show that e-cigarettes are useful cessation tools, practitioners will also need to distinguish between the two types of users. They will need to establish which reasons for trying e-cigarettes are linked to successful smoking cessation. If some reasons for trying (e.g., curiosity, social influence) are not associated with cessation, practitioners should discourage use among individuals who are motivated by those reasons. Should e-cigarettes prove to be useful cessation tools, regulators will also need to ensure that any future policies, such as advertising restrictions, are balanced to ensure that ads can still reach smokers who would use the product as a smoking substitute. In any case, the degree to which e-cigarettes could serve as a harm reduction tool at the population level will depend on the trial and discontinuation choices of different user types.
Limitations to our study include that the TCME survey described e-cigarettes as looking like regular cigarettes, as was the case with first generation models. As e-cigarettes now comprise a class of electronic nicotine delivery devices that do not always look like cigarettes or go by the name “e-cigarettes,” some participants might have misidentified themselves as “never users” because the product they had tried was not called an “e-cigarette” or did not look like a cigarette. Measuring the degree of endorsement of different reasons using more than two responses would have allowed finer-grained analysis. In addition, this study was cross-sectional and assessed only reasons for first trial. Thus, participants’ recollections of their reasons for first trying e-cigarettes could be subject to recall bias. Further, we could not identify participants who intended to stop using e-cigarettes but had not yet done so and could not examine the reasons for the most recent use among participants who had used e-cigarettes repeatedly. We also could not determine whether stated reasons for trying e-cigarettes prompted later changes in behavior, e.g., whether those who tried e-cigarettes in order to quit smoking actually did so. The study data reflect probability and non-probability samples, but the use of quota sampling and survey weights enabled us to align our estimates with the U.S. population. During survey design, we cognitively tested the survey instrument, but we may have missed some response options for reasons to start and stop using e-cigarettes, such as completely quitting nicotine. The partial overlap between some of the established categories (e.g., a friend or family member uses e-cigarettes) and some of the open-ended responses (e.g., a friend or family member offered a puff of an e-cigarette) suggests that response options could have been more precise. Strengths of the study include the large national sample, extensive formative work, and novel research questions.