Appeal and Use of Customizable E-cigarette Product Features in Adolescents

Abstract

Objectives:

We evaluated the appeal and use of e-cigarette product features in a sample of high school students in Connecticut.

Methods:

In 2015, we conducted school-wide surveys in 8 high schools and analyzed data from 1765 students (877 susceptible tobacco-naïve never-e-cigarette users, 315 susceptible tobacco-exposed never-e-cigarette users, and 656 current e-cigarette users). Students were asked: “What is cool about e-cigarettes?” Response options included e-liquid ingredients (eg, flavors, nicotine content, propylene glycol/vegetable glycerin ratio [ie, PG/VG), the heating source (eg, battery), and physical design (eg, color, shape, appearance). Students were asked: “How do you customize your e-cigarette?” (Change flavors, voltage, PG/VG levels, temperature, skin/tank color, other) and use of nicotine in the e-liquid (yes/no). Multivariable regression assessed associations between current e-cigarette users’ customization and e-cigarette use frequency (#days/month).

Results:

Flavors were the most appealing product feature among all groups. Among current users, using nicotine e-liquid was more prevalent than changing flavors (44.2% vs 29.6%). Using nicotine e-liquids and changing PG/vG ratios were associated with more frequent e-cigarette use among current users.

Conclusions:

These findings can inform tobacco regulatory actions that put forth evidence-based recommendations to manufacture e-liquid flavors, nicotine concentrations and PG/VG ratios that prevent e-cigarette initiation and escalation of use.

From 2011 to 2016, the prevalence of lifetime e-cigarette use rose by over 10-fold in adolescents; now, e-cigarette use is more prevalent than traditional cigarette smoking in this age group.¹ Although e-cigarette use rates declined slightly from 2015–2016, there is continued concern that e-cigarettes increase the risk for initiating cigarette smoking and may have adverse long-term health effects.^2–4 The Food and Drug Administration (FDA) began to regulate the manufacturing, sales, and distribution of e-cigarettes as tobacco products in August 2016.⁵ In service of its mission to reduce e-cigarette-related harm, the FDA can regulate e-cigarette product components and parts (henceforth referred to as “product features”) such as e-liquid ingredients (eg, flavors, nicotine content, propylene glycol/vegetable glycerin ratio [ie, PG/ VG]), the heating source (eg, battery), and physical product design features (eg, color, shape, appearance). Thus, there is a need to generate empirical evidence that will help determine how the availability of e-cigarette product features may affect youth, a population inherently vulnerable to tobacco initiation and the development of life-long tobacco use.⁶

When e-cigarettes were introduced to the US market in 2007, e-cigarettes resembled traditional cigarettes, in terms of size and shape (cig-a-like models). However, e-cigarette design has evolved rapidly over time, and newer products generally do not resemble tobacco cigarettes and encourage customization of a wide range of product features that were not available on early cig-a-like models. Advanced-generation e-cigarettes often include an empty, refillable tank that allows users to easily customize e-liquids with their choice of flavors, nicotine concentrations, PG, and VG (ie, the medium that delivers nicotine and flavorings).^7,8 In some cases, advanced generation products also allow for the use of pre-filled cartridges that contain e-liquids with a variety of flavors.⁸ Thus, advanced-generation products allow users to customize nicotine concentrations and PG/VG ratios to facilitate a stronger “throat hit”⁹ and/or better taste.^8,10 Users also can customize the voltage of the battery, resulting in either increased or decreased e-liquid temperature, nicotine yield, or puff volume.^8,11,12 Lastly, users can customize the external appearance of their devices through the use of different colors, designs, and battery sizes. Several studies demonstrate that adult cigarette smokers find the customizable product features of e-cigarettes appealing. However, the scientific evidence regarding the appeal and use of these customizable product features among adolescents is limited despite evidence that advanced generation e-cigarettes are popular among e-cigarette users in this population.¹³

Emerging evidence suggests that adolescents are particularly attracted to the customizable features of e-liquid (eg, flavors, nicotine concentrations, PG/VG); interest in diverse e-liquid flavors is associated with e-cigarette curiosity and experimentation in adolescents,^14,15 and national survey data indicate that 81.5% of adolescent e-cigarette users report that flavors are the leading reason for their continued e-cigarette use.¹⁵ Additionally, a sizable proportion of youth use nicotine e-liquids (eg, 3 mg/ml to > 30 mg/ml). However, one study found that 34% of adolescents did not know the nicotine content of the e-liquid they regularly use.¹⁵ Youth also are interested in “smoke tricks,” which, among other customizations, are facilitated by customizing PG/VG ratios to produce thicker vapor clouds.^9,14 Although these data strongly suggest that youth may be customizing specific aspects of e-liquids, it currently is unknown how commonly youth customize each of these features and how customization translates to use behavior.

Additionally, little is known about the appeal and use of the e-cigarette battery and its temperature, voltage, and physical design features. A 2013 survey of Connecticut high school students found that 60.1% of youth usually used e-cigarettes with rechargeable batteries, whereas only 11.3% used a disposable type.¹⁶ Data also indicate that adolescent e-cigarette users, like adults, “drip” e-liquids directly onto the heating source,¹⁷ which anecdotally increases the intensity of the flavor, throat hit, and thickness of the vapor.¹⁸ Whereas evidence suggests that adolescents customize their e-cigarette batteries, voltage, and/or temperature, there is an overall lack of empirical evidence on the subject.

Of central importance, the customization of these e-cigarette product features has been shown to play an important role in adults’ e-cigarette use and cessation behaviors. Specifically, experienced adult users report that the ability to customize flavors and nicotine concentrations in e-cigarettes helps to promote traditional cigarette reduction or abstinence.^8,19 Although much of the evidence regarding the appeal and use of different product features is relevant to experienced adult e-cigarette users, many of whom are current or former cigarette smokers,^8,19 there are several important reasons to examine the appeal and use of customizable e-cigarette product features in adolescents. First, it is plausible that the availability of certain product features may increase interest in e-cigarette experimentation in youth. For instance, adolescents cite appealing flavors as a primary reason for initiation.¹⁴ In addition, adolescents who are relatively naïve to tobacco products may be attracted to e-cigarettes without nicotine but transition over to e-cigarettes with nicotine. Indeed, our previous data showed that adolescents report initiating e-cigarette use with nicotine-free e-liquids but later switching to e-liquids that contain nicotine.¹⁶ Therefore, it is essential to examine how product features appeal to adolescents who have never tried e-cigarettes but are susceptible to initiation. Second, the use of certain product features, such as nicotine e-liquid, may promote or hasten regular e-cigarette use, and therefore, it is essential to examine how current (ie, past-month) e-cigarette users customize their products and whether customization relates to e-cigarette use frequency.²⁰

Thus, the current study evaluates the appeal and use of customizable e-cigarette product features in a sample of high school students in Connecticut. We specifically examine product features that can be regulated including e-liquid ingredients (flavors, nicotine concentrations, PG/VG), the heating source (voltage/temperature of the battery), and the appearance (lights, color/shape). We examine whether the appeal of various e-cigarette product features differs among youth who have already initiated e-cigarette use when compared with those who have not by examining differences among susceptible never-e-cigarette users who are naïve to all tobacco products, susceptible never-e-cigarette users who have tried at least one other type of tobacco product, and current e-cigarette users. Second, this study aims to determine to what degree current e-cigarette users customize product features, and if the use of customizable product features is associated with increased e-cigarette use frequency. We included use of traditional cigarettes and use of alternative tobacco products and owning an e-cigarette device as covariates because prior research has shown that youth who use e-cigarettes are more likely to engage in these behaviors.^21,22 We also included device ownership as a covariate when examining the association between customization and e-cigarette use frequency, as youth who own their own device (rather than shared someone else’s device) would presumably have more opportunity to customize e-cigarettes. Overall, the findings from this study may begin to inform tobacco product regulations that aim to reduce the appeal of e-cigarettes to youth.

METHODS

Survey Procedures

We conducted a school-wide survey ofadolescents from 8 Southeastern Connecticut high schools (N = 7045) in June 2015. Full methods have been published elsewhere.^17,23 In brief, we surveyed a convenience sample of high schools in southeastern CT. In total, 36 schools were invited to take part in the study, and we enrolled interested schools with the goal of having at least one school from each district reference group (DRG). DRGs are groupings of schools in CT based on socioeconomic features (ie, family income levels, parental education levels, parental occupation status, and use of non-English home language).²⁴ The final sample consisted of 8 schools representing 7 of 9 district reference groups (DRG’s) in southeastern CT.

Information sheets were mailed to parents along with the option of refusing participation for their child. Two parents refused their child’s participation in the surveys. Prior to completing the survey, students were informed that participation was voluntary and that their data were anonymous. Students completed the survey during their homeroom periods. The response rate was 83.8% based on student attendance on the day of survey administration.

Measures

Demographics.

Students reported their sex, race, and age.

E-cigarette use status.

This was measured using responses to: (1) “Have you ever tried e-cigarettes?” and (2) “How many days out of the past 30 days did you use an e-cigarette?” Never e-cigarette users were defined as responding “No” to question 1. Current e-cigarette users were defined as respondents who reported using e-cigarettes ≥1 day in the past month.

E-cigarette susceptibility.

Never e-cigarette users were classified as susceptible to e-cigarette use if they answered “Probably not,” “Probably yes,” or “Definitely yes” (rather than “Definitely not”) to 2 questions: “If one of your best friends offered you an e-cigarette, would you smoke it?” and “Do you think that in the future you might experiment with e-cigarettes?” These questions were adapted from the Pierce et al measure of cigarette susceptibility.^4,16 Our previous analysis of a longitudinal cohort of middle and high school students in CT showed that never e-cigarette users who indicated susceptibility on both items were more than 5 times as likely to initiate e-cigarettes within 6 months and were more than 7 times as likely to report using e-cigarettes in the past month, compared to never e-cigarette users who were not susceptible. Furthermore, this measure of e-cigarette susceptibility predicted subsequent initiation and use beyond other baseline predictors, such as age, other tobacco use, and other substance use.²⁵

Appealing e-cigarette product features.

These were measured with “What is cool about e-cigarettes?” [select all that apply]. We examined answer choices related to e-liquid features (“Different flavors,” “Can customize flavors,” “Can customize nicotine level,” “Can customize vapor level”). We also examined answer choices related to the heating source (“Can change temperature,” “Can change voltage”) and the appearance (“Its shape,” “Lights up”). Additional response choices included: “I don’t think e-cigarettes are cool and “Other” with a write-in response option. Write-in responses in the “Other” category were re-coded to an appropriate category where possible.

Use of customizable e-cigarette product features.

Students were asked: “How do you customize your e-cigaretter” [select all that apply]. We developed the items that were used to assess e-cigarette product customization from product descriptions found on popular e-cigarette websites. Answer choices included those related to the e-liquid (“Change the flavors,” “Change the PG/VG levels”), the heating source (“Change the voltage, temperature”), the color/shape appearance (“Change the skin/tank color”), and “other.” Due to small sample sizes, the voltage and temperature categories were combined into a voltage/temperature variable. Responses to the customization questions were summed to determine the total number of product features that users reported customizing (scale of 0 to 5).

We also assessed the use of e-liquids with varying nicotine concentrations. Students were asked: “What concentration of nicotine do you typically use?” Response options included: “I do not use e-cigarettes,” “0 mg,” “about 6 mg,” “about 12 mg,” “about 18 mg,” “about 24 mg,” and “I don’t know.” Like previous studies, participants were classified as using nicotine-free e-liquids (0 mg), nicotine e-liquids (about 6 mg or more), or not knowing the concentration of the e-liquid used (“I don’t know”).²⁰

Lastly, we assessed e-cigarette device ownership with the question: “When you use an e-cigarette do you...” [select all that apply]. Response options included: “use a device you personally own,” “use a device you share with someone else,” “take one or more puffs from someone else’s device.” Participants were classified as owning their own device if they selected “Use a device you personally own.”

Additional Measures

Lifetime cigarette smoking.

This was assessed with the question: “How old were you when you tried a cigarette, even just 1 or 2 puffs?” Response options included: (1) the option to fill in an age in years; and (2) the option to choose “I never smoked, even just 1 or 2 puffs.” Participants who provided an age in response option 1 were classified as having a history of cigarette smoking.”

Alternative (ie, non-cigarette) tobacco product use.

Students were shown pictures and provided descriptions of each tobacco product and were asked: “Have you tried (blunts, cigars, cigarillos, hookah, smokeless tobacco)?” Response options included “yes” and “no.” “Yes” responses were summed to create a composite alternative tobacco use variable (scale of 0 to 5).¹⁷

Identification of the Sample

In total, 7045 students completed the survey. We identified: (1) susceptible never-e-cigarette users who were naïve to all other tobacco products (N = 883); (2) susceptible never-e-cigarette users who tried at least one other type of tobacco product (hereafter referred to as tobacco-exposed never-e-cigarette users; N = 337); and (3) current e-cigarette users (N = 775). To help ensure that the sample’s e-cigarette use status was valid, we excluded 119 current e-cigarette users (15.4%) who later provided an incongruent response (“I do not use e-cigarettes”) on the question assessing nicotine concentration. We also excluded an additional 28 never-e-cigarette users (6 tobacco-naïve and 22 tobacco-exposed) who indicated they used a nicotine concentration in their e-liquid. The 147 excluded adolescents were more likely than those with congruent data to be in 11^th grade and less likely to be in 10 ^th grade (p < .02), but did not differ on other demographics. The final analytic sample included 1765 students (877 susceptible tobacco-naïve never-e-cigarette users, 315 susceptible tobacco-exposed never-e-cigarette users, and 656 current e-cigarette users).

Data Analytic Plan

Descriptive frequencies were used to describe the prevalence of reporting product features as appealing and to describe the use of customizable product features. To address the first aim, we used chi-square tests to evaluate whether the 3 groups differed in the appeal of e-cigarette product features. To address the second aim (ie, to assess whether use of customizable product features was associated with e-cigarette use frequency among current e-cigarette users (N = 656), multivariable linear regression evaluated the effects of use of product features (flavors, nicotine content, PG/VG, voltage/temperature, and the appearance of the product) on frequency of e-cigarette use (number of days used in past 30 days) when all independent variables were entered simultaneously. For the multivariable analysis, we modeled the dependent variable (days of e-cigarette use) as a count variable and used Proc FMM in SAS v.9.4 (SAS Institute, Cary, NC) with a negative binomial distribution to account for over dispersion of the data among current e-cigarette users.

Given the existing literature demonstrating the association between cigarette smoking, alternative tobacco product use, and e-cigarette use, we included these measures as covariates.^21,22 We also controlled for sex, age, race/ethnicity, school, and device ownership. Multivariable results were considered significant at p < .05.

RESULTS

The final sample consisted of 50.5 % girls and 66.3% white, 18.7% Hispanic, 16.5% black, and 5% Asian students with an average age of 15.9 ± 1.2 years old. Among the susceptible never-e-cigarette users, most participants reported “probably not” to at least one of the susceptibility measures (85.4%), 23.1% reported “probably yes” to one or both measures, and only 1.6% reported “definitely yes.” About one-fourth (25.5%) of the current e-cigarette users reported using only one day in the past month, and 19.8% all 30 days in the past month, with the remaining current e-cigarette users dispersed between 2 and 29 days [median = 5 days; interquartile range (IQR) 1–20].

Figure 1 depicts the prevalence of reporting e-cigarette product features as appealing among all groups. Overall, the groups reported the availability of different flavors as the most appealing product feature and current users were more likely than both groups of susceptible users to endorse each product feature as appealing (p < .001). A majority of tobacco-naïve never-e-cigarette users (68.1%) did not report any product feature as appealing.

Figure 1.

Appeal of Customizable E-cigarette Product Features among Susceptible Never Users and Current Users^a

Note.

a: A larger proportion of current users than susceptible never-users reported that each product feature was appealing (p-values < .001).

Overall, 41% of current e-cigarette users reported that they customized at least one e-cigarette product feature (Figure 2). The most prevalent behaviors were using e-liquids containing nicotine (44.2%), changing flavors (29.6%), and changing PG/VG ratios (18.6%) (Figure 2). Figure 3 demonstrates that current users reported using nicotine e-liquid of varying concentrations (ie, 6 mg [14.8%], 12 mg [8.7%], 18 mg [5.8%], 24 mg [6.1%], and ≥30 mg [8.8.2%]. Many current users used nicotine-free e-liquid (23.0%) or did not know the nicotine concentration of the e-liquid (34.1%). Multivariable models demonstrated that, among current e-cigarette users, using nicotine e-liquid (β = 0.52, SE = 0.12; p < .001), and changing PG/VG ratios (β = 0.41, SE = 0.13; p = .002) was associated with more days of e-cigarette use in the past month (Table 1).

Figure 2.

Use of Customizable E-cigarette Product Features among Current Users

Figure 3.

Use of Nicotine Concentrations among Current E-cigarette Users

Note.

a: Students who responded “I don’t know” to the nicotine concentration question.

Table 1.

Association between Customization of E-cigarette Product Features and E-cigarette Use Frequency

	Estimate	Standard Error	z Value	Pr > |t|
Customization of Product Characteristics
E-liquid Characteristics
Change flavors (Yes vs No)	0.11	0.11	1.02	.31
Use Nicotine in E-liquid
Nicotine vs Nicotine-Free	0.52	0.11	4.57	< .0001
Unknown vs Nicotine-Free	−0.12	0.12	−1.00	.32
Change PG/VG ratios (Yes vs No)	0.32	0.13	2.45	.01
Heating Source Characteristics
Change voltage/temperature (Yes vs No)	−0.01	0.15	−0.10	.92
Product Appearance
Change color (Yes vs No)	0.10	0.12	0.80	.42
Other	0.29	0.18	1.58	.11
Covariates
Demographics
Female vs Male Participant	−0.11	0.09	−1.16	.25
Age	−0.02	0.04	−0.55	.58
School	0.00	0.02	0.14	.89
Race/Ethnicity
White (Yes vs No)	0.02	0.14	0.14	.89
Black (Yes vs No)	0.38	0.16	2.44	.01
Asian (Yes vs No)	0.13	0.22	0.57	.57
Hispanic (Yes vs No)	0.00	0.14	0.01	.99
Tobacco Use Behaviors
Number of alternative tobacco products tried	0.11	0.03	3.32	.001
Ever cigarette smoker (Yes vs No)	0.21	0.10	2.09	.04
Own an e-cigarette (Yes vs No)	0.48	0.10	5.00	< .0001

DISCUSSION

Our study describes the appeal and use of customizable e-cigarette product features among adolescents. The availability of flavors was the most appealing product feature among both susceptible never users and current e-cigarette users. Overall, 41% of current e-cigarette users reported that they customized at least one e-cigarette product feature. Using nicotine e-liquid was more prevalent than was changing flavors among current users (44.2% vs. 29.6%). Furthermore, using nicotine e-liquids and changing PG/VG ratios were associated with greater frequency of e-cigarette use. Use of these product features was associated with e-cigarette use frequency beyond important covariates, such as other tobacco product use and owning an e-cigarette.

Among susceptible never e-cigarette users, the availability of a variety of flavors and the ability to customize flavors were the most appealing product features. The current findings are consistent with previous research indicating that appealing flavors are the most commonly endorsed reason for e-cigarette initiation.¹³ Among current e-cigarette users, we did not detect an association between changing flavors and e-cigarette frequency. This finding aligns with a previous longitudinal study of a cohort of adolescents that found that endorsing flavors as a reason to try e-cigarettes at baseline was not associated with more frequent e-cigarette use one year later.²⁶ A potential hypothesis for the lack of association between changing flavors and e-cigarette frequency is that more frequent users have chosen a preferred flavor and are less likely to change flavors. However, given that this is a cross-sectional association, future research with longitudinal cohorts of youth is needed to determine how flavor experimentation aligns with the natural history of e-cigarette use.

Among current users, using nicotine e-liquid (44.2%) with concentrations of 6 to ≥30 mg was more common than changing flavors (29.6%) or PG/VG ratios (18.4%). Similar to previous studies,²⁰ a sizeable proportion of current users reported that they did not know the nicotine concentrations in the e-liquid they used, which is concerning, because users could inadvertently expose themselves to nicotine (or higher levels of nicotine than intended). Our study also indicated that use of nicotine e-liquids and changing PG/VG ratios were associated with more frequent e-cigarette use. Although this cross-sectional study precludes inferences about causality, this association may indicate that more frequent users simply have more opportunities to customize their products. Alternatively, frequent e-cigarette users may be more likely to be nicotine dependent and use nicotine in the e-liquid to prevent withdrawal. Unique to e-cigarettes, customizing PG/VG ratios may be associated with more frequent use because it provides a more satisfying user experience through improved taste, stronger throat hit, and/or the production of a larger and denser vapor cloud.^8,10,27 Although, the findings from the adult literature have led some experts to support the availability of customizable e-liquids in the service of promoting smoking cessation,⁷ future research is needed to determine whether the ability to customize features like flavors, nicotine concentrations, and PG/VG ratios is similarly related to reduction in traditional cigarette use among adolescent cigarette smokers.

In general, tobacco-naïve never e-cigarette users and tobacco-exposed never e-cigarette users endorsed similar appealing product features, although a larger proportion of susceptible tobacco-naive never users reported that nothing was “cool” about e-cigarettes. The tobacco-exposed group may have more propensity to try new things, as evidenced by their previous experimentation with other tobacco products, and therefore, exhibit more interest in e-cigarette product features. The high prevalence of reporting nothing was “cool” among the tobacco naïve never users also may indicate that, in this group, susceptibility to initiation is better explained by factors other than product appeal, such as social norms or peer use. Another explanation for this finding is that our measure of susceptibility may have been too broad. However, our measure of susceptibility uses a similar definition as the well-validated Pierce measure of cigarette susceptibility²⁸ and has been shown to measure increased likelihood of e-cigarette initiation among adolescents.²⁵ Lastly, never users may not know which product features they find appealing because they do not have first-hand experience with the product.

Our findings should be considered considering several limitations. The cross-sectional design prohibits inferences about causality. The generalizability of our findings also may be limited by our convenience sample of high school students in Connecticut. Examinations of nationally representative samples are needed to confirm these findings. We are unable to ascertain whether specific nicotine concentrations or PG/VG levels promote more frequent e-cigarette use. Further, we are unable to determine whether the students understood specialized terms such as “PG/VG.” We developed the items that were used to assess e-cigarette product customization from product descriptions found on popular e-cigarette websites. However, future research, including cognitive testing, is needed to ensure the psychometric properties of these items and to ensure full coverage of the construct (ie, content validity). Lastly, as e-cigarette products rapidly evolve, the patterns and behaviors observed may not accurately reflect current trends. Future studies are needed to improve understanding of which device types (eg, disposable, cartridge, tank, mods) appeal to adolescents and whether use of certain device types relates to an increased likelihood of experimentation or long-term e-cigarette use in youth.

IMPLICATIONS FOR TOBACCO REGULATION

This study provides important new information about the appeal and use of customizable e-cigarette product features among adolescents who are susceptible never users and current e-cigarette users. Our findings must be considered in parallel to the emerging literature in adults, which indicates that experienced e-cigarette users perceive using a variety of flavors helps with switching from cigarettes to e-cigarettes and cigarette cessation.^8,19 Our data additionally suggest that nicotine use and changing PG/VG ratios are associated with the frequency of e-cigarette use in youth. To inform tobacco regulation better, future research is needed to determine whether changing nicotine concentrations and PG/ VG ratios escalates user satisfaction or the emergence of nicotine dependence. Future studies are also needed to determine whether more frequent e-cigarette use among adolescent cigarette smokers is associated with cigarette smoking cessation. Nevertheless, our study may inform tobacco regulatory actions that put forth evidence-based recommendations to manufacture e-cigarettes with e-liquid nicotine concentrations and PG/VG ratios that simultaneously promote population harm reduction and limit the escalation of nicotine dependence in youth. Additionally, these data add to the emerging evidence that suggests that prevention strategies and tobacco regulatory efforts should consider the potential impact of youth access to e-liquid flavors, nicotine e-liquids, and customizable PG/VG levels.

Human Subjects Statement

The Institutional Review Board of Yale University and the local school administrators approved this study.