Co-occurring Change in Electronic Nicotine Delivery System (ENDS) and Cigarette Harm Perceptions and Use in Young Adults

Abstract

Introduction

The use of electronic nicotine delivery systems (ENDS) may contribute to cigarette use and nicotine addiction by shifting perceptions and norms around tobacco, but little is known about whether or how ENDS use and norms are related to cigarette use and norms, particularly among young adults. This study tested two potential mechanisms by which END use may facilitate cigarette use: decreasing tobacco harm perceptions (desensitization) and increasing favorability of tobacco use (renormalization).

Aims and Method

Analyses included data from 2187 young adults in a longitudinal panel who reported any ENDS or combustible cigarette use at ages 21, 23, or 26. The sample was 53.6% male, 19.3% Hispanic, and 68.0% White. Parallel process latent growth models were fitted to test how change in ENDS use and harm perceptions about ENDS use co-occurred with change in cigarette use and harm perceptions about cigarette use across a 6-year period from ages 21, 23, and 26 (years 2014–2019).

Results

When ENDS harm perceptions decreased and ENDS use increased, cigarette harm perceptions decreased and the favorability of cigarettes increased. Decreases in ENDS harm perception were differentially associated with the likelihood of transition to cigarette use (an increase) and frequency of use when it occurred (a decrease).

Conclusions

Changing tobacco harm perceptions and favorability are two processes by which ENDS use may underlie transitions to cigarette use. Tobacco prevention messaging should emphasize the potential harms of ENDS use that could occur through shifting tobacco perceptions, especially among young adults who are not already frequent cigarette smokers.

Implications

Increasing favorability of ENDS (and increased ENDS use) may generalize to combustible cigarettes. Continued use of ENDS can increase risk of cigarette use when this exposure desensitizes nonsmoking young adults from the dangers of smoking and renormalizes pro-tobacco attitudes. Findings suggest that prevention messaging around tobacco products should emphasize the potential harms of ENDS use (including the increased likelihood of cigarette use after initiating ENDS), especially among young adults who are not already frequent cigarette smokers.

The increasing popularity of electronic nicotine delivery systems (ENDS) could pose a significant threat to the downward trend in combustible cigarette use seen over the past 50 years in young adults.¹ While cigarette use rates have declined due in part to successful public health campaigns promoting awareness of harm and reducing pro-smoking norms,^2,3 ENDS use has been increasing in young adults.⁴ There is significant concern that the novelty of ENDS devices, the suggestion that they are healthier than cigarettes, and their variety of flavor options will lead nonsmoking or formerly smoking adults to use ENDS, which may contribute to nicotine addiction and cigarette use.^5–8 Ample data have shown an increased risk of tobacco use for nonsmoking adults who initiate ENDS.^5,6,8–10 There is additional concern that marketing of ENDS as a cessation aid may stimulate cigarette use among persistently smoking adults and interfere with cessation.^11,12

ENDS use may facilitate cigarette use by shifting perceptions about the harms of tobacco use and lead to more pro-cigarette norms, but little is known about how ENDS use and norms are related to combustible cigarette use and norms. Two complementary mechanisms may be at play: desensitization to the harms of cigarette use and renormalization of cigarette use by boosting positive beliefs about cigarettes. Both mechanisms are rooted in the health belief model, which places beliefs and attitudes—which are shaped by past experiences—as precursors to health decision-making.¹³ For the desensitization mechanism, positive experiences with ENDS may decrease the harm perception of ENDS and other nicotine products, including cigarettes.¹⁴ For example, lack of immediate harm from ENDS may lead users to believe nicotine use in general is not overly harmful and to escalate ENDS and cigarette use. Furthermore, a renormalization process could occur, whereby increasing prevalence and more favorable attitudes about ENDS may contribute to more favorable attitudes about cigarette smoking.^15,16 Many individuals who use ENDS more frequently may be doing so in settings where the use of ENDS and other nicotine and tobacco products is more normative and stigma around any kind of use is lower, leading to more positive individual attitudes about smoking and, subsequently, to smoking behaviors. Like desensitization, renormalization is related to the perception of safety, but is important to examine separately because cigarettes have been increasingly stigmatized not only for their health risks but also for smelling bad, yellowing of teeth and fingers, and second-hand smoke consequences.^15,16 A change toward more favorable attitudes suggests a removal of such stigma. Although young adults have tended to view ENDS more favorably¹⁷ and as less harmful¹⁸ than cigarettes, desensitization and renormalization processes could upend these trends and increase the risk of cigarette use and nicotine addiction.

One way to explore mechanisms consistent with desensitization and renormalization is to measure co-occurring change in perceptions and attitudes about nicotine use in a “cross-product” fashion. For example, change in harm perception for ENDS products would co-occur with or lead to change in harm perception for cigarette products. Most published work has focused on establishing relationships between ENDS use and cigarette perceptions, reporting that increases in ENDS use are associated with increases in favorable perceptions and attitudes toward cigarettes.^14,19–24 However, there is a paucity of studies examining co-occurring change in perceptions, both for ENDS and cigarettes. Assessing these relationships is critical for identifying cross-product shifts in nicotine perceptions that may underlie transitions from ENDS to cigarette use. In addition, most work on cross-product associations has been done in adolescent samples,^{14,19–21,23,24} or cross-sectionally in adults.²² Prospective, longitudinal work assessing these trends is needed in young adults, given the legality and ease of accessing nicotine products in this population, and because motivations for ENDS use may relate more to exploration in adolescents, and more to addiction or to smoking cessation in adults.⁶ Adults in their 20s represent an important age group for studying nicotine perceptions, given the reduced risk for mortality among young adults who quit smoking before age 35.²⁵

Finally, two distinct cigarette use outcomes may be influenced by ENDS perceptions in different ways: onset of cigarette use from nonuse, and frequency of cigarette use when it occurs. Information about ENDS-related smoking onset is carried primarily by people who initially are nonsmoking in young adulthood. Information about ENDS-related cigarette use frequency (when it occurs) is carried primarily by individuals with histories of at least occasional or greater cigarette use, and it is qualitatively distinct in carrying information about increasing toward daily or heavy daily cigarette use. Assessing both cigarette onset and frequency of use provides insight into how young adults with different smoking histories interface with ENDS in different ways. Capturing this heterogeneity is critical for the effective tailoring of prevention messaging around ENDS.

Current Study

Employing a sample of young adults who used nicotine—including ENDS and/or cigarettes—the current study assessed co-occurring change in ENDS and cigarette perceptions and use. Doing so enabled the identification of how ENDS perceptions and use may influence cigarette perceptions and use over time, as illustrated in a conceptual diagram in Figure 1. We measured ENDS-related outcomes at three time points between ages 21 and 26, which corresponds to a critical historical period (2014–2019) when ENDS use proliferated on the US market and may have contributed to changing perceptions of ENDS.⁴ We tested three hypotheses within a desensitization framework: when ENDS harm perception decreases, cigarette harm perception also decreases (hypothesis 1), and cigarette use increases (hypothesis 2); and when ENDS use increases, cigarette harm perception decreases (hypothesis 3). We tested two additional hypotheses within a renormalization framework: we expected that when ENDS harm perception decreases (hypothesis 4) and ENDS use increases (hypothesis 5), increases in cigarette favorability would co-occur. Two additional hypotheses were included because they provide informative context. Focusing exclusively on nicotine use outcomes, we expected that as ENDS use increases, cigarette use also increases (hypothesis 6). In addition, focusing only on ENDS, we predicted that as ENDS harm perception decreases, ENDS use increases (hypothesis 7). Finally, when testing hypotheses about change in cigarette use, we accounted for potentially distinct influences of ENDS on the likelihood of cigarette use onset from nonuse versus frequency of cigarette use when it occurs.

Figure 1.

Conceptual parallel-process latent growth models. Note: Diagrams follow the notation for traditional structural equation modeling. Hypothesis testing for co-occurring change is carried out via slope-slope covariance coefficient terms, indicated by thick double arrows on the right side within the larger growth models. Indicator variables (eg, ENDS 21, Cig 21) are shorthand for the type of tobacco product being referenced (ENDS, cigarettes) and time point for the young adult cohort (ages 21, 23, and 26) in the growth model. Model estimates include cross-product covariance terms (ie, ENDS outcome co-varied with a cigarette outcome). ENDS = electronic nicotine delivery system; Cig = Cigarette.

Method

Participants

The current study used data from the Community Youth Development Study (CYDS), a community-randomized trial of the Communities That Care (CTC) prevention system that aims to reduce substance use and delinquency outcomes in young people. Detailed information about the CYDS longitudinal panel is reported elsewhere.^26–29 The current study did not have any intervention-related aims. The CYDS longitudinal panel includes participants who grew up in 24 small- to moderate-sized towns in seven states. Communities were matched in pairs within state on population, racial diversity, economic indicators, and crime rates. Communities were randomly assigned to the CTC condition and received coalition support for implementing evidence-based prevention interventions, or to a control condition that did not receive such support. Beginning in fall 2004, all grade 5 youth attending public school in CYDS communities were invited to participate in the trial. A total of 76.4% of parents provided consent for their children’s participation, and 4407 of these youth (2205 CTC, 2002 from control) completed baseline surveys in grade 5 or 6. Participants were surveyed annually from grades 5 to 10, and at grade 12 and ages 21, 23, and 26. The study uses data from ages 21, 23, and 26 collected in 2014, 2016, and 2019, respectively. Retention rates were 91.4% at age 21, 88.0% at age 23, and 87.1% at age 26.

Analyses included the subset of 2187 young adults who reported ENDS or cigarette use on at least three past-year occasions at ages 21, 23, or 26 (50.4% of active living sample at age 26). Participants were from both control (n = 1013) and CTC (n = 1174) arms, with pooling of the sample guided by confirmation that control and CTC arms had similar model results in sensitivity analyses. There were no sustained effects of CTC programming on tobacco use by age 21.³⁰ The analysis sample was 53.6% male. Participants selected multiple categories to identify their race: 55.8% chose White, 12.2% American Indian or Alaska Native, 5.1% Black or African American, 2.9% Asian, 2.5% Native Hawaiian or Pacific Islander, and 35.1% other (of which 70.1% identified as Latino or Hispanic). The on-time high school graduation rate was 75.9%. Over half (59.2%) of participants reported lifetime use of combustible cigarettes by age 19. The study protocol was approved by the University of Washington Human Subjects Review Committee.

Measures

See Table 1 for descriptive statistics on key study variables at ages 21, 23, and 26.

Table 1.

Descriptive statistics for ENDS and cigarette outcomes

		Age 21	Age 23	Age 26
		(2014)	(2016)	(2019)
Perception outcomes	Scale range	Mean (SD)
ENDS harm perception	1 = no risk to 4 = great risk	2.52 (0.94)	2.86 (0.93)	3.01 (0.87)
Cigarette harm perception	1 = no risk to 4 = great risk	3.54 (0.76)	3.61 (0.71)	3.62 (0.70)
Cigarette favorable attitudes	1 = very wrong to 4 = not at all wrong	2.31 (1.09)	2.21 (1.10)	2.17 (1.08)
Use outcomes (past year)	Response options	%
ENDS use	Nonusing	48.9	59.3	56.4
	Occasionally	20.9	15.1	12.5
	Monthly	13.4	11.0	8.9
	Weekly	9.5	6.2	9.0
	Daily	7.4	8.3	13.3
Cigarette use	Nonusing	39.7	43.2	53.0
	Occasionally	11.0	10.6	9.0
	Monthly	5.5	5.7	4.2
	Weekly	8.2	7.1	5.6
	Daily	26.4	24.5	20.1
	Heavy daily	9.2	8.9	8.0
Dual ENDS + cigarette use (occasional or greater use for both substances)		39.6	29.5	17.3

Based on Community Youth Development Study sub-sample of 2187 adults reporting any use of ENDS or cigarettes at any time point between ages 21 and 26. Values in italics are standard deviations.

ENDS Outcomes

ENDS harm perception was measured with one item: “How much do people risk harming themselves (physically or in other ways) if they use electronic cigarettes or e-cigarettes (‘Vapes’), such as Ruyan, NJOY, or Juul?” (1 = no risk to 4 = great risk).

ENDS use was measured with items asking “On how many occasions (if any) have you used electronic cigarettes or e-cigarettes (‘Vapes’), such as Ruyan, NJOY, or Juul…” “…in the past 30 days?” (for past-month use) and “…in the past 12 months?” (for past-year use). Response options both for past-month and past-year use were 1 = 0 occasions, 2 = 1–2 occasions, 3 = 35 occasions, 4 = 6–9 occasions, 5 = 10–19 occasions, 6 = 20–39 occasions, 7 = 40 or more occasions. For analyses, a five-level composite variable was created from past-year and past-month use to capture meaningful differences in behavior among regular (ie, monthly, weekly, and daily) and occasional users.^5,31,32 Participant use was categorized as: 0 = nonusing (0 occasions past month and < 3 occasions past year), 1 = occasional use (1–2 occasions past month, or 3–9 occasions past year and 0 occasions past month), 2 = monthly use (3–9 occasions past month, or 10–39 occasions past year and 0–2 occasions past month), 3 = weekly use (10–39 occasions past month, or 40 + occasions past year and 0–9 occasions past month), 4 = daily use (40 + occasions past month).

Cigarette Outcomes

Cigarette harm perception was measured with one item: “How much do people risk harming themselves (physically or in other ways) if they smoke one or more packs of cigarettes a day?” (1 = no risk to 4 = great risk).

Favorable attitudes toward cigarette use were measured with one item: “How wrong do you think it is for someone your age to smoke one or more packs of cigarettes per day?” (1 = very wrong to 4 = not wrong at all).

Cigarette use was measured with items querying past-month use (“How frequently have you smoked cigarettes in the past 30 days?”; 1 = not at all, 2 = less than 1 per day, 3 = 1 to 5 per day, 4 = ½ pack per day, 5 = 1 pack per day, 6 = 1 ½ packs per day, 7 = 2 + packs per day) and past-year use (“On how many occasions (if any) have you smoked cigarettes in the past 12 months?”; 1 = 0 occasions, 2 = 1–2 occasions, 3 = 3–5 occasions, 4 = 6–9 occasions, 5 = 10–19 occasions, 6 = 20–39 occasions, 7 = 40 or more occasions). A composite variable was created with six levels, including a clinically important distinction between daily and heavy daily use.³³ Use was categorized as: 0 = nonusing (no use past month and < 3 occasions past year), 1 = occasional use (< 1 daily cigarette past month and 3–9 occasions past year), 2 = monthly use (< 1 daily cigarette past month and 10–39 occasions past year), 3 = weekly use (< 1 daily cigarette past month and 40 + occasions past year), 4 = daily use (1 cigarette to ½ pack of cigarettes daily past month and 40 + occasions past year), 5 = heavy daily use (1–2 packs of cigarettes daily past month and 40 + occasions past year).

Analysis Plan

A series of latent growth models were fitted in Mplus version 8 to measure change in ENDS and cigarette processes across ages 21, 23, and 26. Full information maximum likelihood (FIML) estimation was used to handle missing data. Missingness for all variables ranged from 5.9% to 6.1% at age 21, 9.6% to 9.7% at age 23, and 10.4% to 11.2% at age 26. Models were adjusted for intervention status (control = 0, CTC = 1) to account for potential differences in substance use risk factors that may have been attributable to CTC exposure. We also added dummy-coded demographic covariates (male, White, Hispanic, on-time high school graduation), based on associations with ENDS and cigarette use in extant literature.^5,34 Lifetime smoking by age 19 was also included as a covariate to control for effects of smoking history but was not included in models estimating a growth term for cigarette past-year use due to collinearity issues. Although CYDS is a community-randomized trial, analyses did not control for community clustering as, by young adulthood, at least half of the sample were not living in the original study community anymore and were dispersed across almost all states in the United States. Prior to hypothesis testing, single-process growth models were fitted for each outcome variable to assure fit (Table S1).

Parallel-Process Latent Growth Modeling: Hypothesis Testing

Parallel-process latent growth models were fitted to test co-occurring change in ENDS and cigarette outcomes (Figure 1).³⁵ A parallel linear growth model for each process was fitted (ie, model 1 for hypothesis 1), whereby the relationship between the two processes was identified simultaneously. Hypothesis testing focused on slope-slope covariances, which indicate the degree to which change in one outcome co-occurred with change in the other outcome, controlling for intercept-intercept covariances and slope-on-intercept regression coefficients.

To test co-occurring change in perception across products, five parallel-process models were specified. Model 1 estimated dual change in ENDS harm perception and cigarette harm perception. Model 2 estimated dual change in ENDS harm perception and cigarette past-year use. Model 2 was a two-part (semi-continuous) growth model³⁶ that estimated the association between ENDS harm perception and cigarette past-year use via the TWOPART function in Mplus. The TWOPART function decomposed the distribution of cigarette use into two parts: a binary part to indicate whether use occurred (“cigarette any use”; 0 = nonusing, 1 = any use; growth here signifies going from nonusing to using), and a continuous part to indicate the frequency of use when it occurred (“cigarette use frequency”; 1 = occasional use to 5 = heavy daily use; growth here signifies escalating use among adults who smoke). The TWOPART function modeled growth in the binary and continuous variables simultaneously, with the binary and continuous parts analyzed as logistic and continuous growth terms, respectively. Model 3 estimated co-occurring change in ENDS past-year use (for example, movement from occasional to monthly use) and cigarette harm perception. We also estimated dual change in ENDS harm perception and favorable attitudes toward cigarettes (model 4), and ENDS past-year use and favorable attitudes toward cigarettes (model 5).

Models Testing Past-Year Cigarette and ENDS Use

Model 6 estimated dual change in ENDS past-year use and two-part cigarette past-year use, again using the TWOPART function. Model 7 estimated dual change in ENDS harm perception and ENDS past-year use. See Table S2 for details about models 1–7 in the sample of nicotine-using young adults, including fit indices and parameter estimates for key covariance terms.

Sensitivity Analyses

Heterogeneous Trajectories.

The main analyses aggregate across a heterogenous group of nicotine users, with some decreasing (n = 516) and others increasing (n = 844) in ENDS harm perception (groupings were based on change scores from age 21 to 26). Models 1–7 were repeated in these subgroups separately to confirm whether co-occurring change patterns would generalize across different trajectories of ENDS harm perception.

Nonusing Adults.

Since the main models were conducted with nicotine-using subsamples, it was important to test whether similar processes are also occurring among non-nicotine users. In this lower-risk group, higher perceptions that cigarette use is harmful could contribute to later onset of ENDS, seen as a less harmful option. Using a sub-sample of adults from CYDS reporting no nicotine use across ages 21, 23, and 26 (n = 1987), Models 1 and 4 were repeated to test whether changes in harm perception for ENDS and cigarettes co-occur.

Results

Parallel-Process Latent Growth Modeling: Hypothesis Testing

Model 1

ENDS harm perception and cigarette harm perception. The slope-slope covariance coefficient term showed that when decreasing ENDS harm perception occurred, decreasing cigarette harm perception also occurred, with the co-movement of these processes in the same direction indicated by a positive coefficient (B = 0.35, CI [0.31, 0.38]).

Model 2

ENDS harm perception and cigarette use. In a two-part cigarette use model, when ENDS harm perception decreased, the likelihood of cigarette past-year any use increased (B = −0.29, CI [−0.52, −0.06]), indicating co-movement in opposing directions. This association was driven primarily by the transition to use onset among historically nonsmoking adults. In contrast, when ENDS harm perception decreased, there was a corresponding decrease in cigarette past-year use frequency when cigarette use occurred (B = 0.05, CI [0.01, 0.09]), with this association driven primarily by adults who are smoking consistently.

Model 3

ENDS use and cigarette harm perception. When ENDS past-year use increased, cigarette harm perception decreased, with these opposing trajectories resulting in a negative coefficient (B = −0.08, CI [−0.15, −0.01]).

Model 4

ENDS harm perception and favorable attitudes toward cigarettes. When a decrease in ENDS harm perception occurred, there was an opposite direction co-occurrence of increasing favorability of cigarettes (B = −0.09, CI [−0.15, −0.04]).

Model 5

ENDS use and favorable attitudes toward cigarettes. When ENDS past-year use increased, the favorability of cigarettes also increased (B = 0.23, CI [0.14, 0.33]).

Other Models Tested

Model 6

ENDS use and cigarette use. In a two-part cigarette use model and extending the pattern of results observed in model 2, an increase in ENDS past-year use was associated with an increase in the likelihood of cigarette past-year any use (B = 1.46, CI [1.10, 1.82]), and a decrease in cigarette past-year use frequency when smoking occurred (B = −0.08, CI [−0.14, −0.02]). This indicated that ENDS use was associated with an increase in uptake in cigarette use, and a substitution effect in decreasing frequency among cigarette users.

Model 7

ENDS harm perception and ENDS use. A decrease in ENDS harm perception was associated with an increase in ENDS past-year use (B = −0.42, CI [−0.52, −0.33]).

Sensitivity Analyses

Heterogenous Trajectories

Slope-slope correlation estimates from models 1–7 in the primary sample were replicated in the sub-sample of adults who increased in ENDS harm perception. These adults drove the mean-level trend of nicotine harm perceptions increasing and use decreasing in the larger sample (see Table S2). In addition, significant and strong slope-slope correlations in the decreasing harm perception group were observed, not inconsistent with a desensitization mechanism in this subsample.

Nonusing Adults

In the subsample of nicotine-nonusing adults, models 1 and 4 replicated findings from the primary sample of nicotine-using adults.

Discussion

This study found support for co-occurring changes in ENDS and cigarette perceptions and use that suggest that positive perceptions of ENDS may be associated with more positive perceptions of cigarettes over time. Co-occurring change patterns involving cigarette attitudes and ENDS were consistent with a desensitization framework, suggesting lowered harm perceptions for cigarettes in the case when young adults began to view ENDS as less harmful (model 1) and increased ENDS use (model 3). Although harm perceptions were increasing and use decreasing both for ENDS and cigarettes over time on average, there was considerable heterogeneity in these change patterns, with many adults decreasing in harm perceptions over time. As confirmed in sensitivity analyses, the fact that cigarette harm perceptions changed alongside ENDS harm perceptions—regardless of whether ENDS harm perceptions were increasing or decreasing—makes it less likely the current results merely reflect a maturing-out process for all forms of nicotine use.³⁷

Co-occurring change patterns were also consistent with a renormalization framework, suggesting a lessening of non-health-related stigmas associated with cigarettes in the case when young adults began to view ENDS as less harmful (model 4) and increased ENDS use (model 5). Although our results reflect individual-level change, these shifts occurred in a larger societal context of increasing prevalence and acceptability of ENDS at the population level during a period when ENDS were emerging on the market.^11,21,34 Our results suggest that population-level shifts around ENDS could facilitate more pro-tobacco social norms, whereby smoking behaviors across similar-looking products would become more visible and would be viewed more favorably by young adults.¹⁵ Future studies testing renormalization should include measures of social norms regarding ENDS and cigarette use as predictors of attitude change for smoking.

A notable caveat to the pattern of co-occurring change was that ENDS perceptions (model 2) and ENDS use (model 6) were associated with cigarette onset (from nonuse) and cigarette use frequency (when it occurs) in opposing ways, even though both onset and use frequency were decreasing at the mean level. This trend may be explained by different pathways into ENDS between adults who do or do not smoke cigarettes. For example, for cigarette use onset among nonsmoking adults, a purported pathway begins with using ENDS for recreational purposes (eg, enjoyment, curiosity^4,38), which leads to the uptake of cigarettes via desensitization and renormalization. The trend for cigarette onset raises concern, given increasing ENDS use in the US adult population among never smokers.^4,16 In contrast, among frequently smoking adults, viewing ENDS as increasingly less harmful (and increasing their use) may coincide with smoking less, as would be the case when adults in this group begin to substitute ENDS for cigarettes as a cessation tool.^11,12

Limitations and Strengths

Although the sample is representative of the communities that were included in the Communities That Care intervention, most youth grew up in rural environments where smoking is more prevalent than in urban or suburban areas.³⁹ Analytic procedures need to be replicated with nationally representative samples. Second, in addition to population-level changes in ENDS use⁴ and perceptions,18 the ENDS marketplace evolved considerably during the data collection window (2014–2019), including the availability of higher potency e-liquid and enhanced marketing of ENDS products. Associations between ENDS and cigarette perceptions need to be replicated with more current samples, and an examination needs to be conducted of whether different ENDS products (ie, vapes, e-cigars) with different substances (ie, nicotine, marijuana) have similar associations with cigarette use and attitudes. Strengths include a large and diverse longitudinal sample with prospectively collected data on ENDS and cigarette use and attitudes, and capturing important heterogeneity in ENDS perceptions depending on cigarette smoking frequency. This is one of the first studies to explore ENDS-related mechanisms through which population-level increases in tobacco use may be seen in the future. Although the study design was not set up for claims about the causal direction of co-occurring changes in use and perception, results give clear direction for future inquiries and have implications for prevention.

Conclusions

The current study is the first to capture longitudinal trajectories of co-occurring change in ENDS and cigarette perceptions and use among nicotine-using young adults, extending published work on ENDS perceptions in adolescent samples.^14,20,21 Results demonstrate that when young adults developed more favorable perceptions about and increased their ENDS use, they also did so with cigarettes. Continued use of ENDS can increase the risk of cigarette use when this behavior desensitizes nonsmoking young adults from the dangers of smoking and renormalizes pro-tobacco attitudes. Our findings suggest that prevention messaging around tobacco products should emphasize the potential harms of ENDS use, especially among young adults who are not already frequent cigarette smokers.

Supplementary Material

Supplementary material is available at Nicotine and Tobacco Research online.

Funding

This work was supported by the National Cancer Institute at the National Institutes of Health (grant number NCI R37CA225690 to MAE), with co-funding for the Community Youth Development Study from the National Institute on Drug Abuse at the National Institutes of Health (grant numbers R01DA015183 to SO, R56DA044522 to SO, and R01DA044522 to SO & MRK). The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the funding agencies.

Declaration of Interests

The authors declare that they have no conflicts of interest.

Data Availability

The data, analytic code, and materials necessary to reproduce the analyses or attempt to replicate the findings in this article are not publicly accessible but are available from the first author upon reasonable request.

Human Subjects Ethical Treatment

The study was conducted in accordance with the ethical standards of the American Medical Association. All research protocols involving human participants were approved by the University of Washington Institutional Review Board and were in accordance with the ethical standards of the 1964 Helsinki declaration and its later amendments. Parent/guardian consent was obtained for all research protocols involving individuals under the age of consent. An earlier version of this manuscript was presented at the annual meeting of the Society for Research on Nicotine and Tobacco held in March 2023 in San Antonio, TX.

Author Contributions

Justin Caouette (Conceptualization, Formal analysis, Investigation, Methodology, Visualization, Writing—original draft, Writing—review & editing [lead]), Marina Epstein (Conceptualization, Formal analysis [supporting], Funding acquisition [lead], Investigation, Methodology, Project administration [supporting], Supervision [lead], Writing—original draft, Writing—review & editing [supporting]), Max Halvorson (Conceptualization, Investigation, Methodology, Writing—review & editing [supporting]), Sarah Danzo (Conceptualization, Investigation, Methodology, Writing—review & editing [supporting]), Margaret Kuklinski (Conceptualization, Funding acquisition, Resources, Writing—review & editing [supporting]), and Sabrina Oesterle (Conceptualization, Funding acquisition, Investigation, Writing—review & editing [supporting])