Electronic Nicotine Delivery Systems Marketing and Initiation Among Youth and Young Adults

The Tobacco Center of Regulatory Science on Youth and Young Adults has 2 rapid-response longitudinal surveillance systems that are used to assess recall of tobacco marketing and tobacco use among school-going youth (ages 12–17 years at baseline) and young adults (ages 18–29 years at baseline). The 2 studies are being conducted in the 5 counties surrounding the 4 largest cities in Texas: Houston, Dallas-Fort Worth, San Antonio, and Austin. The Texas Adolescent Tobacco and Marketing Surveillance System (TATAMS) includes students in sixth, eighth, and 10th grades at baseline (n = 3907; N = 461 069) from 79 middle and high schools who were recruited by using a complex multistage probability design.³³ The Marketing and Promotions Across Colleges in Texas (M-PACT) project includes students (N = 5482) enrolled in 24 2- and 4-year colleges.³⁴ After providing active informed consent, participants voluntarily completed 5 Web-based surveys every 6 months through the spring of 2017, and retention rates for both studies were >70% at the 2.5-year follow-up. For the purposes of this study, only participants who reported never using ENDS at baseline and who had complete data were included. Thus, data from 2288 TATAMS participants and 2423 M-PACT participants were included in the current study. Additional details about study methods are described elsewhere for TATAMS³³ and M-PACT.³⁴ 

This study was approved by the Center for the Protection of Human Subjects at the University of Texas Health Science Center at Houston (HSC-SPH-13-0377) and Institutional Review Board at the University of Texas at Austin (2013-06-0034) and meets the ethical standard outlines in the Helsinki Declaration of 1975 as revised in 2000. Informed consent was obtained from all participants and from parents for those age <18 years.

Baseline survey items were used to assess recalled exposure to ENDS marketing through 5 channels: TV, radio, billboards, retail stores, and Internet (see Supplemental Table 3 for items). Dichotomized Internet exposure (1 = exposed; 0 = unexposed) was derived from an item in which students were asked to check where they remember seeing marketing for ENDS. For all other channels, questions used to assess frequency of recalled exposure were recoded to dichotomous variables (1 = exposed; 0 = unexposed) for consistency across channels, similar to research conducted by Nicksic et al.²⁵ For TV, radio or Internet radio, and billboards, students who recalled observing ENDS marketing “occasionally,” “frequently,” and “very frequently” were considered exposed; those who “never” or “rarely” recalled ENDS marketing or those who did not use the channel (M-PACT only) were considered unexposed. For retail stores, students who recalled observing ENDS marketing “some of the time,” “most of the time,” or “every time” were considered exposed. Those who did not visit any stores in the past month or those who did not recall ENDS marketing (“never/not that I remember”) were considered unexposed.

The primary outcome variable was ENDS initiation, which was assessed at all 5 follow-up waves with the same item asking participants if they had ever used ENDS (see Supplemental Table 3 for item). ENDS never users who reported ever using ENDS in any of the subsequent 5 follow-up waves were considered ENDS initiators and assigned a score of 1. Those who confirmed never using ENDS at the 2.5-year follow-up (ie, wave 6) were considered noninitiators and were assigned a score of 0.

Covariates assessed at baseline were sex (1 = male, 0 = female), race and/or ethnicity (1 = Hispanic, 2 = Non-Hispanic white or other race, and 3 = Non-Hispanic African American), current other tobacco use (0 = did not use cigarettes, cigars, hookahs, or smokeless tobacco in the past 30 days; 1 = used at least 1 other product on at least 1 day in the past 30 days), and peer ENDS use (number of close friends who use ENDS; scored 1 [none] to 5 [all]). Three additional baseline covariates were assessed that varied across the 2 samples. Sensation seeking was measured by using the 4-item Brief Sensation Seeking Scale 4³⁵ for both groups, but responses ranged on a 5-point scale for the TATAMS sample and on a 4-point scale for the M-PACT sample. Moreover, TATAMS included grade level (sixth, eighth, and 10th) as a covariate, whereas M-PACT included age and type of college (2- vs 4-year college) as covariates.

Because of the complex survey design of TATAMS, we applied sampling weights for all statistical analyses and adjusted for clustering within schools and for nonresponse bias, which allowed results to be generalized to the population of sixth, eighth, and 10th grade students in the sampled counties (see Pérez et al³³). For M-PACT, multilevel analyses were used to accommodate the clustering of participants in college by including college in each model as a random effect. All analyses were performed by using Stata 15.³⁶ 

Logistic regression analyses were used to determine if baseline recall of ENDS marketing predicted subsequent ENDS initiation up to 2.5 years later. Separate models were examined for the youth and young adult samples, and only participants reporting never using ENDS at baseline were included. Models included all 5 types of recalled marketing exposure (TV, radio or Internet radio, billboards, retail stores, and Internet) as independent variables to control for the other types of exposure and were additionally adjusted for baseline covariates of sex, race and/or ethnicity, grade level (TATAMS only), age and college type (M-PACT only), current other tobacco use, sensation seeking, and peer ENDS use.

Data from 2288 youth who reported never using ENDS at baseline were included in analyses; an additional 926 youth were excluded because of missing data. Data from 2423 young adults who reported never using ENDS at baseline were also included in analyses; an additional 423 young adults were excluded because of missing data. χ² and t tests were conducted to determine differences on all baseline variables between those included and those excluded. For youth, results indicated there were differences between the 2 groups on 4 of 11 study variables: excluded youth were more likely to be in sixth or eighth grade compared with 10th grade, were more likely to be Hispanic or African American, and were less likely to recall ENDS marketing exposure in retail stores or on the Internet. For young adults, there were group differences for 3 of 12 variables: excluded young adults were more likely to be Hispanic or African American, were more likely to be enrolled in 2-year compared with 4-year colleges, and were more likely to recall exposure to ENDS marketing on billboards.

Of the 2288 youth who reported never using ENDS at baseline, ∼14% (weighted percentage; n = 359) initiated ENDS use during the 2.5-year follow-up period. As shown in Table 1, ∼50% of the sample were girls, and more than half identified as Hispanic. Sixth- and eighth-graders each composed more than one-third of ENDS never users, with 28% in 10th grade at baseline. Only baseline ENDS marketing at retail stores was recalled by a majority of the sample (58%), whereas all other forms of marketing exposure were recalled by less than half. Only 1.2% of the sample reported current use of other tobacco products at baseline.

Of the 2423 young adults who reported never using ENDS at baseline, 25% (n = 611) initiated ENDS use during the 2.5-year follow-up period. As shown in Table 1, 67% of the sample were women; 64% identified as non-Hispanic white or other, and just more than one-quarter identified as Hispanic. At baseline, participants were, on average, 20.3 years old, and most were enrolled in a 4- vs 2-year college. A majority of the sample recalled ENDS marketing at retail stores and on the Internet at baseline (∼56%), whereas <30% recalled all other forms of exposure. Approximately 11% of the sample reported current use of other tobacco products at baseline.

Baseline recall of ENDS marketing exposure in retail stores was the only significant marketing predictor of subsequent ENDS initiation up to 2.5 years later. Recall of marketing in retail stores was significant even after controlling for the significant effects of the following covariates: race and/or ethnicity, grade level, and peer ENDS use. Youth who recalled ENDS marketing at retail stores had 1.99 times higher odds of ENDS initiation up to 2.5 years later compared with those who did not recall store-based marketing (see Table 2).

Baseline recall of ENDS marketing exposure in retail stores significantly increased the odds of subsequent ENDS initiation up to 2.5 years later, with odds of initiation 1.30 times higher for young adults who recalled store-based ENDS marketing compare with those who did not. Additionally, recall of ENDS marketing on TV significantly increased odds of subsequent ENDS initiation, with odds of initiation 1.29 times higher for those who recalled marketing on TV compared with those who did not. These effects were significant even after controlling for the significant effects of current other tobacco use, sensation seeking, and peer ENDS use (see Table 2).

Findings reveal that recalled exposure to ENDS marketing at retail stores makes a unique contribution to ENDS initiation up to 2.5 years later for both youth and young adults. For young adults, ENDS marketing on TV also uniquely contributes to ENDS initiation. It is notable that recall of these marketing exposures was significant even after we controlled for important covariates, including other tobacco use, sensation seeking, and peer ENDS use. These findings build on a small body of evidence that reveals a temporal link between self-reported ENDS marketing and use^25,–27,37 similar to that observed for other tobacco products.⁶ 

Results reveal the importance of examining marketing channels individually rather than cumulatively because different marketing channels have varying influences on behavior. Understanding how and why these differences exist remains a critical question because characteristics, such as content, opportunities for user engagement, and placement, may affect how marketing channels influence behavior.¹ Drawing on the elaboration likelihood model, 1 mechanism of influence may be the degree to which viewers actively engage with marketing (influence via the “central route”) versus passively absorbing (“peripheral route”).^38,39 Although marketing on TV, in magazines, on radio, and on billboards tends to be passively observed, retail stores and the Internet provide both cues and opportunities for immediate response. For example, retail stores often display marketing near checkout counters, simultaneously marketing ENDS products and providing direct access to them.⁴⁰ Notably, the use of self-reported measures of marketing exposure in this study likely means that findings most closely reflect the influence of central-route exposure on behavior because to actively recall seeing ENDS marketing reveals conscious recognition of its presence. Different odds ratios by channel therefore reflect not only marketing dosage and content but also the likelihood that young people who have never used ENDS consciously register and engage with those channels.

With this study, we extend existing research by examining the impact of ENDS marketing on both youth and young adults. Both are important developmental periods, but research on young adults is generally lacking. Although the results presented here reflect 2 distinct study samples and cannot be directly compared, observable differences suggest the importance of considering these 2 groups separately and of investigating ENDS-use trajectories across the transition to adulthood. Recall of ENDS marketing at retail stores predicted ENDS initiation among both groups, but ENDS marketing on TV predicted only young adult initiation, although more youth appeared to recall ENDS marketing on TV (43.8%) compared with young adults (28.9%). Although it is not clear why recall of TV marketing predicted only young adult initiation, 1 possibility lies in the content of TV advertisements and their appeal (or lack thereof) to youth. Previous research used to examine ENDS advertisements appearing on TV and/or social media between 2009 and 2014 revealed that a large proportion contained messages that are not typically appealing to youth, such as messages that are critical of cigarettes, that compare ENDS use with cigarette use, and/or that describe health effects.⁷ However, given the lack of research examining the impact of advertisements through various channels on ENDS-use behaviors across developmental levels, more research that tracks youth and young adults across time is needed. Continued surveillance is important because industry marketing and participant exposure and engagement may change over time. Results from a prospective national study of 12- to 17-year-old adolescents reveal that engagement with online tobacco and ENDS marketing increased from 8.7% in 2013–2014 to 20.9% in 2014–2015.⁴¹ It is likely that engagement will continue to rise because the prevalence of using both ENDS and social media has continued to rise.^3,42 

The current study fills an important gap in the literature regarding ENDS marketing and subsequent ENDS initiation. However, there are limitations. Data are specific to Texas and may not be generalizable to the larger population. Reliance on self-reported data is also a limitation. Although some research reveals a moderate correlation between recalled exposure and objective proxy measures,⁴³ self-reports introduce potential for reporting biases. For example, recall of ENDS marketing on TV and the Internet may capture informational content (eg, news stories or blogs)⁴⁴ and not just advertisements and promotions for ENDS. However, self-report likely reflects primarily conscious exposure (ie, central-route exposure in the elaboration likelihood model), and findings have unique implications in this context. An important direction for future research is using objective exposure measures to study the influence of marketing on behavior. Finally, although initiation is 1 of the first stages in a progression of tobacco- and/or nicotine-product use behaviors, it is not possible to determine who will become a regular user. Additional longitudinal research is needed to determine the role of ENDS marketing exposure on subsequent stages of youth and young adult ENDS use as they unfold.

Recall of ENDS marketing exposure in retail stores uniquely predicts youth and young adult ENDS initiation, and recall of marketing on TV predicts young adult initiation. These findings underscore the need for regulation of ENDS marketing, particularly in retail stores. An important regulatory consideration for the US Food and Drug Administration includes reducing the size, visibility, and availability of ENDS advertisements and products in retail stores. A consideration for state and local governments is restricting price promotions, discounts, and coupons that are prevalent in the retail environment⁴⁵ given that young people are more likely to purchase tobacco and/or nicotine products as the prices decrease.⁴⁶ Finally, state governments can consider entering into assurances of voluntary compliance with retailers who sell ENDS products to further limit the type and placement of advertisements in their retail stores.⁴⁷ Given the increasing popularity of ENDS products among young people, limiting ENDS marketing in retail stores is of paramount importance.

ENDS

electronic nicotine delivery system

M-PACT

Marketing and Promotions Across Colleges in Texas

TATAMS

Texas Adolescent Tobacco and Marketing Surveillance System