Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Oct 1.
Published in final edited form as: Addict Behav. 2023 Jun 19;145:107784. doi: 10.1016/j.addbeh.2023.107784

An experimental evaluation of the effects of banning the sale of flavored tobacco products on adolescents’ and young adults’ future nicotine vaping intentions

Michael Dunbar a,*, Claude M Setodji a, Steven C Martino a, Desmond Jensen b, Rosemary Li a, Armenda Bialas a, William G Shadel a
PMCID: PMC10478339  NIHMSID: NIHMS1911855  PMID: 37364525

Abstract

Background:

Some U.S. states and municipalities have banned the sale of flavored tobacco products to help curb youth vaping. However, evidence supporting such bans is limited. This experiment tested whether removing flavored tobacco products from a retail setting diminished adolescents’ (ages 11–20) future intentions to use vaping products.

Methods:

The study was implemented in the RAND StoreLab, a life-sized model convenience store. The display of flavored tobacco products in the store was manipulated with these conditions: 1) tobacco, sweet, and menthol/mint flavors displayed; 2) only tobacco and menthol/mint displayed; and 3) only tobacco flavors displayed. Participants were randomly assigned to shop in one of these conditions and completed measures of future vaping intentions post-shopping. Separate logistic regression models assessed effect of condition on future intentions to use different flavors (tobacco-, menthol/mint-, and sweet-flavored) and any flavor (composite score across flavor categories) of vaping products.

Results:

Study condition was not associated with intentions to use menthol/mint-, sweet-flavored, or any flavor. Compared to the condition in which all flavored products were displayed, removing menthol/mint- and sweet-flavored products significantly increased future intentions to use tobacco-flavored vaping products (OR = 3.97, 95 % CI [1.01, 15.58], p < .05). This effect was only observed among adolescents with history of vaping (OR = 11.30, 95 % CI [1.42, 89.96], p = .02).

Conclusions:

Flavor bans may not affect adolescents’ intentions to use menthol/mint, sweet, or “any” flavor of vaping products but may increase intentions to use tobacco-flavored products for teens who have already started vaping.

Keywords: E-cigarettes, Vaping, Adolescent, Young adult, Tobacco flavor ban, Tobacco policy

1. Introduction

Prevalence of e-cigarette use (vaping) among U.S. adolescents rose dramatically in the past decade and remains high (Gentzke et al., 2022; Miech et al., 2019; Park-Lee et al., 2022; Wang et al., 2020). Although less harmful to health than combustible cigarettes, e-cigarettes contain nicotine and other toxicants that pose risks to young people such as nicotine dependence and exacerbation of respiratory symptoms (The National Academies of Sciences Engineering and Medicine, 2018). As such, there is a strong need for actions that curb adolescents’ use of vaping products.

Restricting or banning the sale of non-tobacco flavored e-cigarettes (hereafter, “flavor restrictions” or “flavor bans”) has been proposed and/or implemented as a policy action to help curb vaping among young people. Although they also appeal to adult tobacco users (Gades et al., 2022; Harrell et al., 2016), characterizing flavors likely played a central role in the rise in nicotine vaping among adolescents in the past decade (Gaiha et al., 2022; Leventhal et al., 2019; Wang et al., 2020; U.S. Department of Health and Human Services, 2016). Prior work indicates that nearly all adolescents (98 %) and most young adults (71 %) who vape nicotine initiated use with a flavored product (Harrell et al., 2016), and approximately 85 % of teens who currently vape nicotine report using e-cigarettes with characterizing flavors other than tobacco (Cooper et al., 2022). For these reasons and others, restricting the sale of characterizing non-tobacco flavors is theorized to reduce adolescents’ initiation and continued use of vaping products.

Primarily in response to the recent rise in adolescent e-cigarette use (Ali et al., 2022; Gades et al., 2022; Goniewicz & Stanton, 2022), hundreds of local municipalities and several states have enacted policies restricting sales of flavored e-cigarettes and other tobacco products (Campaign for Tobacco-Free Kids). Policy approaches vary considerably (e.g., products, flavors affected) (Campaign for Tobacco-Free Kids), with comprehensive bans on sales of all non-tobacco flavors for all tobacco products (e.g., as in San Francisco) representing the most robust model. Yet few studies have rigorously examined effects of flavor restrictions on youth vaping. Available data from non-experimental local and state-level evaluations suggest that these policies may reduce prevalence of vaping and other tobacco use among adolescents. For example, Hawkins et al. (2021) found that the adoption of flavor restrictions in Massachusetts was associated with a significant reduction in county-level prevalence of nicotine vaping for adolescents under age 18. Other studies have also shown a significant reduction in youths’ use of both flavored and unflavored tobacco products (including e-cigarettes) following adoption of flavor restrictions in Massachusetts municipalities (Kingsley et al., 2019, 2020). Another recent study examining state-wide e-cigarette flavor restrictions in New York found that past-month prevalence of e-cigarette use among teens (ages 16–19 years old) decreased significantly following restrictions (Schneller et al., 2022). Additionally, some data indicate that comprehensive tobacco product flavor restrictions in San Francisco were associated with decreased prevalence of young adult e-cigarette use (Kingsley et al., 2019, 2020), although a more recent analysis of California Healthy Kids Survey data found no significant change in e-cigarette use prevalence among high school students pre-to-post San Francisco’s restrictions (Dove et al., 2022). Collectively, these findings suggest that flavor restrictions may reduce use of any vaping products among adolescents at the population level, perhaps by decreasing initiation.

However, the impact of flavor bans on teens who have already initiated vaping is less clear. Some studies suggest that, among young people who use e-cigarettes, flavor restrictions may have limited impact on flavored product use. For example, the New York state evaluation referenced above found that over 95 % of teens who reported vaping following implementation of statewide flavor restrictions reported using flavored e-cigarettes (Schneller et al., 2022), which suggests that, among youths who already vape, banning sales of flavored products may not significantly deter use of flavored products. Other studies suggest that flavor restrictions may prompt some individuals to substitute or switch to available flavors (e.g., menthol/mint or tobacco-flavored) rather than discontinuing vaping entirely. For example, in a mixed methods study of young adult e-cigarette users’ behaviors and attitudes surrounding flavor restrictions, Romm et al. (2022) found that most participants indicated that federal e-cigarette flavor restrictions had no impact on their e-cigarette use, with 36 % indicating that they continued to use available flavors, and some respondents reported switching to available menthol or tobacco flavors following restrictions (Romm et al., 2022). Further, in a sample of young adults with history of tobacco use in San Francisco recruited through an online crowdsourcing platform, one study found that most survey respondents indicated that they were still able to obtain flavored tobacco products through multiple channels following the city’s comprehensive ban (Yang et al., 2020). Another recent study by Ali et al. (2022) found that statewide restrictions on flavored e-cigarette sales in Massachusetts, Rhode Island, New York, and Washington were associated with reductions in total e-cigarette sales; however, reductions were “partially compensated by an increase in tobacco-flavored e-cigarette sales in all 4 states,” suggesting that some individuals transitioned to vaping tobacco-flavored products (Ali et al., 2022). Although evidence is limited, particularly with respect to youth, these findings suggest that flavor restrictions may have different effects on future vaping behavior among individuals who have already begun vaping compared to those who have not. This has implications for gauging the potential for restrictions to reduce harm for individuals who have already started vaping and underscores a need to better understand how these policies may differently affect youths who have and have not initiated vaping.

As more municipalities and states consider flavor restrictions as a lever for reducing adolescent vaping, experimental evidence of the potential impact of these actions is urgently needed. The present study provides such evidence by experimentally testing the impact of enacting tobacco flavor restrictions at retail point-of-sale on adolescents’ and young adults’ future intentions to use vaping products. Examining this issue at retail point-of-sale is critical, given that most young people are exposed to vaping products at retail outlets. Specifically, we use a convenience store shopping experiment to examine the impact of removing the display of menthol/mint- and sweet-flavored tobacco products in a tobacco powerwall on young peoples’ intentions to use sweet-flavored, menthol/mint flavored, and tobacco-flavored vaping products. We hypothesized that, compared to a condition in which all flavors were displayed, exposure to flavor ban conditions would lessen future intentions to use vaping products. Furthermore, we hypothesized that effects on intentions to use vaping products would vary by flavor category, such that flavor restrictions would lead to greater reductions for intentions to use sweet and menthol/mint flavors compared to tobacco-flavored products. Because some data indicate that effects of flavor restrictions may vary by individuals’ vaping status, we also explored whether effects of condition differed for young people with and without lifetime vaping history.

2. Methods

2.1. Participant Recruitment and consent

Adolescents and young adults were recruited from the Pittsburgh, PA area via print, online, and radio advertisements between September 2020 and May 2022. Recruitment materials advertised an opportunity to participate in a study about convenience store shopping patterns but did not specify that the study pertained to tobacco to reduce sampling bias. Interested parents (or young adults) were screened to determine eligibility: aged 11–20 years, no medical or psychological condition(s) that would preclude participation in the study, and no previous participation in a RAND StoreLab study. Eligible participants attended an in-person session during which they were provided a brief description of the study (e.g., that it involved a shopping task and questionnaires to assess convenience store purchase patterns and attitudes). Participants and parents (of minor participants) were informed that some details of the study could not be disclosed upfront because doing so could affect the study results and that they would be provided with complete information at the end of the session. Individuals then provided written informed consent or assent (plus parental consent for minor participants). A total of 276 individuals (n = 149 young adults; n = 118 adolescents under 18) consented/assented to participate and completed the study. The study was approved by the RAND Human Subjects Protection Committee (IRB #2018–0822).

2.2. Procedures

The study was conducted in the RAND StoreLab, a replica of a mid-sized convenience store that was designed to experimentally assess tobacco regulatory actions in a model retail point-of-sale setting (Shadel et al., 2016). Briefly, the RAND StoreLab is located within an office building in Pittsburgh, PA, and is only accessible to research participants. It stocks over 650 products (e.g., food, beverages, magazines/newspapers, toiletries, tobacco). Advertisements for tobacco and other products are displayed on the walls, shelves, and windows and a tobacco powerwall is located directly behind a check-out counter. For the present study, the powerwall (display size 58.5 square feet) included 181 tobacco products: 120 cigarette packages, 27 e-cigarette packages, 12 little cigar/cigarillo packages, and 22 smokeless tobacco products packages.

2.3. Experimental conditions

Experimental conditions included a control condition in which the powerwall displayed products with all existing characterizing flavor profiles (All Flavors; consistent with the status quo in Pittsburgh, Pennsylvania), a condition in which sweet characterizing flavors other than tobacco or menthol/mint were banned (Tobacco/Menthol), and a condition in which all non-tobacco flavors were banned (Tobacco Only; i.e., menthol/mint and sweet flavors removed). In the present study, we differentiated between tobacco flavors, menthol/mint flavors, and other flavors to align with the definition of characterizing flavors used by the US FDA (21 U.S.C. 387g - Tobacco product standards). For conditions in which flavored products were removed from the powerwall, sweet- and menthol/mint- flavored products were replaced with tobacco-flavored products of the same brand and brand subtype. The power wall size, location, number of products displayed (total and within each product class), price information, and branded posters were identical in all study conditions. Fig. 1 shows images of the powerwall used in each of the three conditions.

Fig. 1.

Fig. 1.

Open in a new tab

Study conditions.

2.4. Shopping task and pre- and post-shopping questionnaires

Prior to the experiment, participants completed a baseline questionnaire that assessed demographics, tobacco use attitudes and beliefs, and convenience store shopping experiences, along with ‘filler’ items that addressed behaviors unrelated to tobacco to help mask the purpose of the study (e.g., use of and beliefs about soft drinks, potato chips). After completing the baseline questionnaire at the in-person session, participants were randomized to one of the experimental conditions (see Fig. 1) and participated in a simulated shopping task (Shadel et al., 2016). Briefly, a study research assistant instructed individuals to ‘shop’ in the store for whatever item(s) they wanted, to spend at least $5, and to check-out and pay as they would in a typical convenience store using a pre-paid debit card, which they were told was loaded with $10. Another research assistant, who was not involved in the consent/assess or questionnaire administration procedures, served as the cashier and facilitated the check-out process. Although we assessed attempted tobacco purchases as part of our study protocol, no participant attempted to purchase tobacco from the RSL. If a participant had attempted to purchase tobacco in the RSL, the “cashier” would have requested age verification and the purchase would have been declined because all participants were under age 21. After completing the shopping task, participants were instructed to complete a post-shopping questionnaire, which assessed the same constructs included in the baseline questionnaire.

Participants then underwent a debriefing procedure in which they were told about the study purpose, told that any items they purchased would need to be returned, addressed any questions they had about the study, and viewed a 20-minute tobacco media literacy video followed by a guided discussion. They then received printed tobacco prevention materials, a $40 gift card, and reimbursement for parking/transportation costs.

3. Measures

3.1. Baseline demographic characteristics

Participants provided information about their age, education (high school degree vs. not), gender (male; female; nonbinary or another gender), race (Black/African American, White, Asian, another race), and Hispanic/Latino ethnic identity in response to the baseline questionnaire.

3.2. Baseline nicotine vaping and tobacco use history

Lifetime nicotine vaping was assessed with the question, “Have you ever used electronic nicotine products in your life?” (yes/no) and an accompanying definition of electronic vaping products from the Population Assessment of Tobacco and Health (PATH) study (Hyland et al., 2017). Individuals who reported lifetime use were also asked about past-month use of any nicotine vaping products (yes/no) and whether they had tried sweet- or menthol-flavored vaping products. Participants also completed parallel items assessing any lifetime use of other tobacco products (cigarettes; cigars/cigarillos; smokeless tobacco); we created a derived variable from these items and lifetime vaping to indicate any lifetime tobacco use (yes/no).

3.3. Baseline measure of future nicotine vaping intentions

Future vaping intentions were assessed using separate items adapted from a measure shown to predict future adolescent tobacco use (Choi et al., 2001; Cole et al., 2019). Items used to assess tobacco-flavored vaping intentions were: “Do you think you will try an unflavored vaping product (unflavored = tobacco-flavored vaping product) anytime soon?”, “Do you think you will use an unflavored vaping product (unflavored = tobacco-flavored vaping product) anytime in the next year?”; and “If one of your best friends offered you an unflavored vaping product (unflavored = tobacco-flavored vaping product), would you use it?”. For menthol- or mint-flavored vaping products, three parallel questions were asked, specifying “…a vaping product flavored with menthol or mint.” For sweet-flavored vaping products, the items asked about “a vaping product flavored with fruit, candy, alcohol or non-alcohol drink, or some other sweet flavor.”

Responses to all items were made on a 1 (Definitely Not) to 10 (Definitely Yes) scale and summed within each flavor profile to produce a future vaping intentions score, for which higher scores indicate greater future intentions (all α’s > 0.91). Due to skewness in these measures for all flavor profiles (i.e., over 69 % of participants had the lowest score possible on each scale, i.e., a ‘3’), scores for each scale were dichotomized: those who scored a ‘3’ were recoded as ‘0’ (no intention) and any scores greater than ‘3’ were coded as ‘1’ (any intention) (Choi, Gilpin, Farkas, & Pierce, 2001; Cole, Kennedy, Chaurasia, & Leatherdale, 2019; Shadel et al., 2016) Additionally, to assess the impact of condition on intentions to use any flavor of vaping product, we created a composite variable from our three dichotomized intention variables for Tobacco-, Menthol/mint-, and Sweet-flavored vaping products. Those who reported any intention to use a vaping product regardless of flavor category were coded as ‘1’ (any intention) on this composite measure; those with intention scores of zero for all three flavor categories were coded as ‘0’ (no intention) on this composite measure.

3.4. Dependent measures (post-shopping): Future nicotine vaping intentions

Identical vaping intentions items included in the baseline questionnaire were administered immediately after the shopping task. As with baseline measures, scores were dichotomized due to skew, such that those with scores greater than three were coded as 1 (any intention) and those with scores of 3 were coded as 0 (no intention).

3.5. Analyses

We first conducted univariate tests (means, frequencies) on sample characteristics of interest and bivariate chi-square and t-tests to assess demographic differences by lifetime vaping history. To assess the impact of study condition on future vaping intentions, we conducted separate multiple logistic regression models to assess the effect of study condition on each dependent variable: future intentions to use tobacco-flavored vaping products, sweet-flavored vaping products, menthol/mint-flavored vaping products, and any flavor of vaping products. Models controlled for baseline intentions (i.e., corresponding to the dependent variable for that model), lifetime vaping history (yes/no), and demographic characteristics known to correlate with vaping (age, race/ethnicity, gender, education) (Dai et al., 2021; Leventhal et al., 2019; Wang et al., 2020). To explore whether effects of condition varied by vaping history, we added an interaction term for study condition by vaping history to the logistic models described above. Finally, to examine effects of study conditions among those who had and had not previously vaped nicotine, we conducted separate logistic models within each vaping history subgroup (past vaping; no past vaping).

4. Results

4.1. Sample characteristics

Table 1 summarizes demographic characteristics of the sample. Briefly, participants averaged 16.56 (SD = 3.11) years old and 55 % (n = 147) had graduated high school. A majority of participants were female (65 %), 31 % male, and 4 % identified as nonbinary or another gender. Approximately two thirds (62 %) were White, 12 % were Black, 17 % Asian, 7 % identified as two or more races, 6 % identified as Hispanic/Latino, and 2 % identified as another race/ethnicity. Approximately 29 % (n = 78) of the sample reported any lifetime nicotine vaping and 9 % (n = 25) reported any past-month vaping. Of those reporting lifetime vaping, 86 % (n = 67) reported having used a sweet-flavored vaping product and 69 % (n =54) reported having used a menthol/mint vaping product.

Table 1.

Sample characteristics at baseline.

Characteristic Full sample (N = 267) Lifetime Vaping History Subgroups
Ever vaped (n = 78) Never vaped (n = 189) Subgroup differencea p-value
Participant age (mean (SD)) 16.56 (3.11) 18.40 (1.90) 15.80 (3.21) <0.001
Graduated high school (% yes) 54.85 % 82.05 % 43.68 % <0.001
Gender identity 0.10
 Female 65.17 % 73.08 % 61.9 %
 Male 31.09 % 21.79 % 34.92 %
 Nonbinary or other gender 3.75 % 5.13 % 3.17 %
Race 0.44
 Asian 16.54 % 14.29 % 17.46 %
 Black 12.03 % 14.29 % 11.11 %
 White 62.03 % 63.64 % 61.38 %
 Other 2.26 % 3.90 % 1.59 %
 Two or more races 7.14 % 3.90 % 8.47 %
Hispanic/Latino 6.02 % 9.09 % 4.76 % 0.18
Lifetime use of any tobacco product (% yes) 32.58 % 100 % 4.76 % <0.001
Baseline future vaping intentions (% any intention)
 Tobacco-flavored 22.32 % 50.00 % 10.26 % <0.001
 Menthol/mint-flavored 27.11 % 65.28 % 9.15 % <0.001
 Sweet-flavored 30.40 % 68.49 12.34 % <0.001
 Any flavorb 31.00 % 69.86 % 12.82 % <0.001
Open in a new tab

Note. SD = standard deviation.

a

Vaping history subgroup differences were assessed using bivariate chi-square (categorical variables) and t-tests (continuous variables). Bolded values indicate significant group difference at p < .05.

b

Intention to use “Any flavor” was a composite of Tobacco-, Menthol/mint-, and Sweet-flavored intention variables. Individuals who reported intention to use Tobacco-, Menthol/mint-, and/or Sweet-flavored vaping products were categorized as having intention to use “Any flavor”; those who reported no intention (score = 0) for all three flavor categories were categorized as having no intention to use “Any flavor” of vaping products.

4.2. Main effects of flavor ban condition on future vaping intentions

Table 2 shows effects of study condition on future vaping intentions. Adjusting for demographic characteristics, baseline vaping intentions and vaping history were significantly, positively associated with future use intentions in all models (see Table 2). Additionally, there was a significant main effect of study condition on future intentions to use tobacco-flavored vaping products such that, compared to the All Flavors condition, the Tobacco Only condition (i.e., removing all flavors, including menthol/mint), was associated with greater future intentions to use tobacco-flavored vaping products (aOR = 3.97 95 % CI 1.01–15.58, p < 0.05). Study condition was not associated with future intentions to use menthol/mint-, sweet-flavored, or any flavor vaping products (see Table 2).

Table 2.

Main effects of flavor ban condition on future vaping intentions.

Independent variable Dependent variable: Future vaping intentions (yes/no)
Tobacco-flavored Menthol/mint-flavored Sweet-flavored Any flavor a
AOR LCL UCL p-value AOR LCL UCL p-value AOR LCL UCL p-value AOR LCL UCL p-value
Participant age 1.11 0.74 1.66 0.626 1.08 0.71 1.65 0.702 1.30 0.86 1.97 0.218 1.34 0.89 2.01 0.158
Graduated high school (reference = no) 0.73 0.07 8.17 0.798 0.62 0.06 6.89 0.701 0.79 0.07 8.42 0.847 0.50 0.05 4.94 0.550
Race (reference = non-white) 1.17 0.39 3.48 0.779 0.70 0.24 2.05 0.517 1.28 0.44 3.75 0.647 1.04 0.36 3.02 0.945
Gender (reference = not female) 0.33 0.10 1.06 0.063 0.40 0.12 1.33 0.134 0.83 0.26 2.67 0.749 0.64 0.21 1.99 0.443
Baseline future vaping intentionsb 66.63 20.50 216.50 <0.001 64.90 20.28 207.65 <0.001 124.55 32.36 479.41 <0.001 132.34 36.33 482.11 <0.001
Lifetime vaping (reference = ever vaped) 0.12 0.03 0.46 0.002 0.10 0.03 0.39 0.001 0.24 0.07 0.80 0.020 0.34 0.10 1.17 0.087
Flavor ban condition (reference group = All Flavors)
 Tobacco and Menthol/mint Only 1.42 0.37 5.47 0.614 1.24 0.34 4.47 0.745 1.21 0.35 4.14 0.766 1.01 0.29 3.53 0.985
 Tobacco only 3.97 1.01 15.58 0.048 1.26 0.32 4.92 0.741 1.23 0.34 4.52 0.751 1.08 0.30 3.84 0.911
Open in a new tab

Note. Values are estimates from multiple logistic regression models. Separate models were conducted for each dependent variable. AOR = adjusted odds ratio. LCL = lower 95 % confidence limit. UCL = upper 95 % confidence limit. Bolded values indicate statistically significant effect at p < .05.

a

Any flavor intentions variable was as a composite of the three flavor category intention variables. Individuals who reported intention to use Tobacco-, Menthol/mint-, and/or Sweet-flavored vaping products were categorized as having intention to use “Any flavor”; those who reported no intention (score = 0) for all three flavor categories were categorized as having no intention to use “Any flavor” of vaping products.

b

Each model controlled for baseline intention score corresponding to the dependent variable in that model (e.g., the model examining future tobacco-flavored vaping intentions after the shopping task controlled for future tobacco-flavored vaping intentions at baseline). As expected, baseline intention was highly correlated with post-shopping intention and as such led to very high odds and wide confidence intervals.

4.3. Differences by lifetime vaping history

To better understand the effect observed in the Tobacco Only condition, the next set of models examined whether condition effects (and non-effects) were moderated by vaping history. No significant interaction between the Tobacco Only condition and vaping history was observed for models examining effects on menthol/mint- (interaction p = .22), sweet-flavored (interaction p = .48), or any flavor (interaction p = .22) vaping intentions. The study condition-by-vaping history interaction was marginally significant (b = −2.61, 95 % CI −5.37–0.15, p = .06) in the model examining future intentions to use tobacco-flavored vaping products. To further understand this trend, we modeled the effects of condition on future vaping intentions separately for individuals with and without lifetime vaping (see Table 3). Among individuals with lifetime vaping, there was a significant main effect of condition such that the Tobacco Only condition was associated with a significantly higher likelihood of future tobacco-flavored vaping intentions compared to the All Flavors condition (aOR = 11.30, 95 % CI 1.42–89.96, p = .02). No effects of condition on future vaping intentions were observed among those who have not previously vaped (all p’s for effect of condition ≥ 0.42). These sub-analyses suggest that the observed positive main effect of condition on future tobacco-flavored vaping intentions was specific to individuals who had previously vaped.

Table 3.

Effects of flavor ban condition on future vaping intentions within ever- and never-vaped subgroups.

Ever Vaped Subgroup (n = 78)
Predictor Dependent variable: Future vaping intentions (yes/no)
Tobacco-flavored Menthol/mint-flavored Sweet-flavored Any flavor a
AOR LCL UCL p-value AOR LCL UCL p-value AOR LCL UCL p-value AOR LCL UCL p-value
Participant age 1.02 0.57 1.83 0.941 0.85 0.40 1.80 0.675 0.94 0.35 2.53 0.909 1.00 0.42 2.38 0.994
Graduated high school (reference = no) 1.07 0.06 19.82 0.964 2.38 0.07 76.42 0.625 2.30 0.04 138.76 0.691 0.89 0.02 39.02 0.953
Race (reference = non-white) 2.82 0.67 11.84 0.158 1.14 0.24 5.44 0.873 1.79 0.25 12.76 0.563 1.99 0.30 13.35 0.478
Gender (reference = not female) 0.37 0.07 2.12 0.266 0.96 0.13 6.88 0.966 2.38 0.17 32.79 0.518 1.13 0.12 10.83 0.918
Baseline future vaping intentions 23.45 5.06 108.64 <0.001 32.75 6.85 156.66 <0.001 168.14 12.32 2295.50 <0.001 80.93 11.65 562.37 <0.001
Flavor ban condition (reference group = All Flavors)
 Tobacco and Menthol/mint Only 1.73 0.36 8.26 0.491 1.93 0.38 9.70 0.426 1.09 0.16 7.34 0.930 1.13 0.18 7.21 0.898
 Tobacco only 11.30 1.42 89.96 0.022 2.99 0.31 29.27 0.346 2.09 0.12 37.69 0.617 3.92 0.24 64.52 0.340
Never Vaped Subgroup (n = 189)
Dependent variable: Future vaping intentions (yes/no)
Predictor Tobacco-flavored Menthol/mint-flavored Sweet-flavored Any flavora
AOR LCL UCL p-value AOR LCL UCL p-value AOR LCL UCL p-value AOR LCL UCL p-value
Participant age 1.22 0.61 2.42 0.577 1.31 0.71 2.42 0.382 1.52 0.91 2.55 0.113 1.56 0.91 2.67 0.104
Graduated high school (reference = no) 0.38 0.01 25.27 0.653 0.20 0.01 7.71 0.386 0.38 0.02 7.24 0.520 0.28 0.01 5.79 0.409
Race (reference = non-white) 0.36 0.05 2.74 0.327 0.52 0.10 2.68 0.431 1.17 0.29 4.73 0.825 0.88 0.21 3.63 0.862
Gender (reference = not female) 0.30 0.04 2.11 0.225 0.20 0.03 1.15 0.072 0.55 0.14 2.09 0.379 0.45 0.12 1.75 0.249
Baseline future vaping intentions 279.09 34.41 2263.43 <0.001 160.24 19.99 1284.12 <0.001 100.02 18.47 541.49 <0.001 175.88 27.25 1135.25 <0.001
Flavor ban condition (reference group = All Flavors)
 Tobacco and Menthol/mint Only 0.51 0.03 8.97 0.643 0.55 0.07 4.55 0.577 1.10 0.20 5.98 0.912 0.77 0.13 4.37 0.766
 Tobacco only 1.33 0.13 13.31 0.807 0.44 0.06 3.28 0.424 0.90 0.20 4.11 0.892 0.61 0.13 2.93 0.536
Open in a new tab

Note. Values are estimates from multiple logistic regression models stratified by lifetime vaping history. Separate models were conducted for each dependent variable. AOR = adjusted odds ratio. LCL = lower 95 % confidence limit. UCL = upper 95 % confidence limit. Bolded values indicate statistically significant effect at p < .05. Each model controlled for baseline intention score corresponding to the dependent variable in that model (e.g., the model examining future tobacco-flavored vaping intentions after the shopping task controlled for future tobacco-flavored vaping intentions at baseline). As expected, baseline intention was highly correlated with post-shopping intention and as such led to very high odds and wide confidence intervals.

a

Any flavor intentions variable was as a composite of the three flavor category intention variables. Individuals who reported intention to use Tobacco-, Menthol/mint-, and/or Sweet-flavored vaping products were categorized as having intention to use “Any flavor”; those who reported no intention (score = 0) for all three flavor categories were categorized as having no intention to use “Any flavor” of vaping products.

5. Discussion

To our knowledge, this is the first study to experimentally assess the impact of restricting flavored tobacco products on young people’s future intentions to use tobacco-, menthol/mint-, and sweet-flavored vaping products. We found that, compared to a condition in which all flavors were displayed, removing menthol/mint- and sweet-flavored products from a convenience store setting increased future intentions to use tobacco-flavored vaping products among adolescents who had previously vaped. Study conditions had no effect on adolescents’ intentions to use menthol/mint- or sweet-flavored vaping products.

These results suggest that policy actions like flavor restrictions could lead some young people who vape nicotine (i.e., those who use flavored products – the vast majority of adolescents and young adults who vape nicotine (Harrell et al., 2016; Gaiha et al., 2022; Leventhal et al., 2019; Romm et al., 2022) to substitute flavored vaping products with tobacco-flavored products in response to a ban – especially those individuals who are nicotine dependent and are unable or unwilling to quit. This is consistent with recent studies that have observed evidence for substitution following implementation of restrictions (Ali et al., 2022; Li et al., 2022). Municipalities that choose to enact flavor restrictions to reduce teen vaping should consider adopting additional prevention and intervention strategies (Adams et al., 2021; American Lung Association, 2021; Truth Initiative, 2021) to help support vaping cessation among young people and ensure that restrictions do not simply shift teens who have already started vaping toward tobacco-flavored products.

These findings may also be driven in part by increased prominence of tobacco-flavored products in the Tobacco-only (vs. status quo) condition. Our decision to retain a fixed powerwall display size and replace flavored products with tobacco-flavored products was based on a desire to minimize confounding factors across conditions as well as anticipated retailer behavior following restrictions and absent other regulations. Because powerwall product displays are dictated by contracts between retailers and tobacco companies that determine the display space devoted to the company’s products and promotional materials (Reimold et al., 2022), flavor restrictions alone are unlikely to prompt retailers to reduce powerwall display size or number of products displayed. The tobacco industry has an extensive history of shifting their marketing and product promotion strategy at point-of-sale in response to regulations (Bar-Zeev et al., 2022; Evans-Reeves et al., 2019). As such, it is likely that retailers would replace displays of ‘banned’ products with those for which sales remain permitted, which is the scenario we examined in the present study. Replacing menthol-/mint- and sweet-flavored products with tobacco-flavored products in our Tobacco-only condition could have increased participants’ attention to tobacco-flavored products, which may in turn have driven increases future use intentions. Although we did not assess attention to the powerwall or to specific products in the present study, our previous work has shown that greater attention to the powerwall is associated with increased tobacco use intentions (Martino, Setodji, Dunbar, & Shadel, 2019). More research is needed to understand potential mechanisms through which flavor restrictions may exert effects on future product use intentions.

Study conditions had no effect on intentions to use menthol/mint- or sweet-flavored products in the full sample or in any subgroup analyses. This is consistent with some previous work suggesting that, following implementation of flavor restrictions, teens and young adults who vape continue to overwhelmingly use flavored products (Romm et al., 2022; Schneller et al., 2022). This underscores a need for strict enforcement of restrictions and comprehensive regulations to prevent young people from accessing flavored products following restrictions (e.g., from informal sources or non-compliant retailers) or switching to devices that may facilitate continued use of preferred flavors (Posner et al., 2022; Romm et al., 2022; Schneller et al., 2022; Yang et al., 2020). For example, following FDA’s 2020 guidance prioritizing enforcement of flavored cartridge-based e-cigarettes (e.g., Juul, Vuse), data indicated increases in adolescents’ use of other device types (e.g., disposable e-cigarettes like PuffBar) that continued to be sold in a range of characterizing flavors (Gaiha, Lempert, McKelvey, & Halpern-Felsher, 2022; Park-Lee et al., 2022). This may indicate a need for comprehensive regulations coupled with strong enforcement in order to close regulatory gaps and optimize potential benefits of flavor restrictions for young people.

Additionally, we observed no significant condition effects on intentions to use any flavor of vaping products. Firstly, this suggests that flavor restrictions at point-of-sale may not dampen overall vaping intentions among young people. Secondly, this underscores the potential importance of distinguishing between flavor categories to provide a more comprehensive understanding of the impact of flavor restrictions on youth vaping intentions. That is, focusing only on intentions to use any vaping products may fail to capture meaningful variability in use intentions across product subcategories such as characterizing flavor profiles. Depending on the research questions and resources, future studies examining the impact of regulatory actions on product use intentions or behavior may benefit from asking more detailed questions about product subcategories (e.g., flavor profiles, device types) in order to characterize potential heterogeneity. Such data may be useful in identifying potential targets for prevention/intervention and informing regulatory decisions about product standards.

Flavor restriction conditions had no impact on future vaping intentions in the subset of teens who had not previously vaped. This may be attributable to a preponderance of committed never-tobacco users, for whom future use intentions may be essentially fixed, in this group: most reported no intention to use vaping products in the future at baseline and very few ever used any tobacco. This is consistent with prior work, showing that tobacco advertising and regulatory actions may have limited effects on individuals who are tobacco naïve and committed to not using tobacco in the future (Dunbar, Setodji, Martino, & Shadel, 2019; Shadel et al., 2019; West, Romero, & Trinidad, 2007). Although perhaps unsurprising, our findings suggest that flavor restrictions are unlikely to increase risk of future use among those already at low risk.

Findings from the study should be considered in the context of limitations. The RAND StoreLab is a model convenience store and thus may not generalize to real-world settings. Additionally, effects were observed following a single, brief exposure. It is unclear whether or how repeated exposures to real-world retail settings in which flavor bans are enacted may affect perceived product availability, use intentions, or other factors over time and whether/how effects may persist or diminish. Although the intention measures we used have been shown to predict future tobacco use behavior, we did not assess the effect of study conditions on participants’ vaping behavior. Furthermore, low cell size precluded examination of differences based on current vaping status and by use of flavored products. Finally, the present study focused on adolescents and young adults under age 21. Because characterizing flavors also play an important role in the appeal of vaping products for adults, restrictions may also affect vaping intentions and behavior among adults (Gades et al., 2022). Decisions surrounding adoption of flavor restrictions should consider potential consequences of these policies on all segments of the population, including adolescents and adults who already use tobacco.

In conclusion, this study adds to the small literature examining the impact of flavor restrictions on adolescent vaping. Our findings suggest that restricting characterizing flavors for tobacco products may not affect intentions to vape flavored products in the future and may increase interest in tobacco-flavored vaping products among teens who have already initiated vaping. Flavor restrictions may need to be accompanied by additional strategies to support vaping cessation among the sizable proportion of young people who have already used vaping products.

Acknowledgments

The authors would like to thank Amanda Meyer, Samantha Ryan, Grace Falgoust, and Abigail Torbatian for their work on this project. Funding for this study was provided by a grant from the National Cancer Institute (NCI) (R01CA236608).

Role of funding sources

Funding for this study was provided by a grant from the National Cancer Institute (NCI) (R01CA236608). NCI had no role in the study design, collection, analysis or interpretation of the data, writing the manuscript, or the decision to submit the paper for publication.

Footnotes

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Disclaimer

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

CRediT authorship contribution statement

Michael Dunbar: Conceptualization, Methodology, Writing – original draft. Claude M. Setodji: Data curation, Software, Formal analysis, Writing – original draft. Steven C. Martino: Validation, Writing – review & editing. Desmond Jensen: Writing – review & editing. Rosemary Li: Writing – review & editing. Armenda Bialas: Writing – review & editing. William G. Shadel: Funding acquisition, Data curation, Supervision, Writing – review & editing.

Data availability

Data will be made available on reasonable request.

References

  1. Adams ZW, Kwon E, Aalsma MC, Zapolski T, Dir A, & Hulvershorn LA (2021). Treatment of adolescent e-cigarette use: limitations of existing nicotine use disorder treatment and future directions for e-cigarette use cessation. Journal of the American Academy of Child and Adolescent Psychiatry, 60(1), 14–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Ali FRM, Vallone D, Seaman EL, et al. (2022). Evaluation of statewide restrictions on flavored e-cigarette sales in the US from 2014 to 2020. JAMA Network Open, 5(2), Article e2147813. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. American Lung Association (2021). N-O-T: Not On Tobacco—Proven Teen Smoking and Vaping Cessation Program. https://www.lung.org/quit-smoking/helping-teens-quit/not-on-tobacco. Accessed May 1, 2023.
  4. Bar-Zeev Y, Berg CJ, Khayat A, et al. (2022). IQOS marketing strategies at point-of-sales: a cross-sectional survey with retailers [published online ahead of print, 2022 Feb 9]. Tobacco Control. tobaccocontrol-2021–057083. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Campaign for Tobacco-Free Kids. States and localities that have restricted the sale of flavored tobacco products. https://www.tobaccofreekids.org/assets/factsheets/0398.pdf.
  6. Choi WS, Gilpin EA, Farkas AJ, & Pierce JP (2001). Determining the probability of future smoking among adolescents. Addiction, 96(2), 313–323. [DOI] [PubMed] [Google Scholar]
  7. Cole AG, Kennedy RD, Chaurasia A, & Leatherdale ST (2019). Exploring the predictive validity of the susceptibility to smoking construct for tobacco cigarettes, alternative tobacco products, and E-cigarettes. Nicotine & Tobacco Research, 21(3), 323–330. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cooper M, Park-Lee E, Ren C, Cornelius M, Jamal A, & Cullen KA (2022). Notes from the field: E-cigarette use among middle and high school students - United States, 2022. MMWR Morbidity and Mortality Weekly Report, 71(40), 1283–1285. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Dai H, Ramos AK, Faseru B, Hill JL, & Sussman SY (2021). Racial disparities of E-cigarette use among US youths: 2014–2019. American Journal of Public Health, 111 (11), 2050–2058. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Dove MS, Gee K, & Tong EK (2022). Flavored tobacco sales restrictions and teen E-cigarette use: Quasi-experimental evidence from California. Nicotine & Tobacco Research, 25(1), 127–134. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dunbar MS, Setodji CM, Martino SC, & Shadel WG (2019). Graphic health warning posters increase some adolescents’ future cigarette use susceptibility by changing normative perceptions of smoking: A case of mediated moderation. Psychology of Addictive Behaviors, 33(7), 649–658. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Evans-Reeves KA, Hiscock R, Lauber K, & Gilmore AB (2019). Prospective longitudinal study of tobacco company adaptation to standardised packaging in the UK: Identifying circumventions and closing loopholes. BMJ Open, 9(9), Article e028506. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Gades MS, Alcheva A, Riegelman AL, & Hatsukami DK (2022). The role of nicotine and flavor in the abuse potential and appeal of electronic cigarettes for adult current and former cigarette and electronic cigarette users: A systematic review. Nicotine & Tobacco Research, 24(9), 1332–1343. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Gaiha SM, Lempert LK, McKelvey K, & Halpern-Felsher B (2022). E-cigarette devices, brands, and flavors attract youth: Informing FDA’s policies and priorities to close critical gaps. Addictive Behaviors, 126, Article 107179. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Gentzke AS, Wang TW, Cornelius M, et al. (2022). Tobacco product use and associated factors among middle and high school students — National Youth Tobacco Survey, United States, 2021. MMWR Surveillance Summaries, 71, 1–29. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Goniewicz ML, & Stanton CA (2022). Impact of flavour restricting policies on non-cigarette tobacco products. Tobacco Control, 31(S3), S159–S160. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Harrell MB, Weaver SR, Loukas A, Creamer M, Marti CN, Jackson CD, et al. (2016). Flavored e-cigarette use: Characterizing youth, young adult, and adult users. Preventive Medicine Reports, 5, 33–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Hawkins SS, Kruzik C, O’Brien M, & Coley RL (2021). Flavoured tobacco product restrictions in Massachusetts associated with reductions in adolescent cigarette and e-cigarette use. Tobacco Control, 31(4), 576–579. [DOI] [PubMed] [Google Scholar]
  19. Hyland A, Ambrose BK, Conway KP, et al. (2017). Design and methods of the Population Assessment of Tobacco and Health (PATH) Study. Tobacco Control, 26(4), 371–378. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kingsley M, Setodji CM, Pane JD, et al. (2019). Short-term impact of a flavored tobacco restriction: Changes in youth tobacco use in a Massachusetts community. American Journal of Preventive Medicine, 57(6), 741–748. [DOI] [PubMed] [Google Scholar]
  21. Kingsley M, Song G, Robertson J, Henley P, & Ursprung WWS (2020). Impact of flavoured tobacco restriction policies on flavoured product availability in Massachusetts. Tobacco Control, 29(2), 175–182. [DOI] [PubMed] [Google Scholar]
  22. Leventhal AM, Miech R, Barrington-Trimis J, Johnston LD, O’Malley PM, & Patrick ME (2019). Flavors of e-cigarettes used by youths in the United States. Journal of the American Medical Association, 322(21), 2132–2134. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Li D, Ossip DJ, Bansal-Travers M, & Xie Z (2022). Impact of the FDA flavour enforcement policy on flavoured electronic cigarette use behaviour changes. Tobacco Control, 31, s176–s183. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Martino SC, Setodji CM, Dunbar MS, & Shadel WG (2019). Increased attention to the tobacco power wall predicts increased smoking risk among adolescents. Addictive Behaviors, 88, 1–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Miech RA, Johnston LD, O’Malley PM, Bachman JG, & Patrick ME (2019). Trends in adolescent vaping, 2017–2019. New England Journal of Medicine, 381(15), 1490–1491. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Park-Lee E, Ren C, Cooper M, Cornelius M, Jamal A, & Cullen KA (2022). Tobacco product use among middle and high school students - United States, 2022. MMWR Morbidity and Mortality Weekly Report, 71(45), 1429–1435. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Posner H, Romm KF, Henriksen L, Bernat D, & Berg CJ (2022). Reactions to sales restrictions on flavored vape products or all vape products among young adults in the United States. Nicotine & Tobacco Research, 24(3), 333–341. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Initiative Truth. (2021). This is Quitting. https://truthinitiative.org/thisisquitting. Accessed May 1, 2023.
  29. Reimold AE, Lee JGL, & Ribisl KM (2022). Tobacco company agreements with tobacco retailers for price discounts and prime placement of products and advertising: a scoping review [published online ahead of print, 2022 Jan 24]. Tobacco Control;tobaccocontrol-2021–057026. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Romm KF, Henriksen L, Huang J, et al. (2022). Impact of existing and potential e-cigarette flavor restrictions on e-cigarette use among young adult e-cigarette users in 6 US metropolitan areas. Preventive Medicine Reports, 28, Article 101901. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Schneller LM, Kasza KA, Hammond D, Bansal-Travers M, O’Connor R, & Hyland A (2022). E-cigarette and tobacco product use among NYS youth before and after a state-wide vaping flavour restriction policy, 2020–2021. Tobacco Control, 31 (S3), S161–S166. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Shadel WG, Martino SC, Setodji C, Dunbar M, Kusuke D, Lanna S, & Meyer A (2019). Placing Antismoking Graphic Warning Posters at Retail Point-of-Sale Locations Increases Some Adolescents’ Susceptibility to Future Smoking. Nicotine & Tobacco Research, 21(2), 220–226. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Shadel WG, Martino SC, Setodji CM, Scharf DM, Kusuke D, Sicker A, & Gong M (2016). Hiding the tobacco power wall reduces cigarette smoking risk in adolescents: using an experimental convenience store to assess tobacco regulatory options at retail point-of-sale. Tobacco Control, 25, 679–684. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. The National Academies of Sciences Engineering and Medicine. (2018). Public health consequences of e-cigarettes. Washington, D.C.: The National Academies of Sciences Engineering and Medicine. [Google Scholar]
  35. U.S. Department of Health and Human Services (2016). E-cigarette use among youth and young adults. A report of the surgeon general. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. [Google Scholar]
  36. Wang TW, Neff LJ, Park-Lee E, Ren C, Cullen KA, & King BA (2020). E-cigarette use among middle and high school students - United States, 2020. MMWR Morbidity and Mortality Weekly Report, 69(37), 1310–1312. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. West JH, Romero RA, & Trinidad DR (2007). Adolescent receptivity to tobacco marketing by racial/ethnic groups in California. American Journal of Preventive Medicine, 33(2), 121–123. [DOI] [PubMed] [Google Scholar]
  38. Yang Y, Lindblom EN, Salloum RG, & Ward KD (2020). The impact of a comprehensive tobacco product flavor ban in San Francisco among young adults. Addictive Behaviors Reports, 11, Article 100273. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

Data will be made available on reasonable request.

RESOURCES