Article Text

Correlates of tobacco product cessation among youth and adults in the USA: findings from the PATH Study Waves 1–3 (2013–2016)
Free
1. Karin A Kasza1,
2. Kathryn C Edwards2,
3. Zhiqun Tang2,
4. Cassandra A Stanton2,3,
5. Eva Sharma2,
6. Michael J Halenar2,
7. Kristie A Taylor2,
8. Elisabeth A Donaldson4,
9. Lynn C Hull4,
10. Maansi Bansal-Travers1,
11. Jean Limpert4,
12. Izabella Zandberg4,
13. Lisa D Gardner4,
15. Nicolette Borek4,
16. Heather L Kimmel5,
17. Wilson M Compton5,
18. Andrew Hyland1
1. 1 Department of Health Behavior, Roswell Park Comprehensive Cancer Center, Buffalo, NY, USA
2. 2 Behavioral Health and Health Policy Practice, Westat, Rockville, MD, USA
3. 3 Department of Oncology, Georgetown University Medical Center, Washington, DC, USA
4. 4 Office of Science, Center for Tobacco Products, Food and Drug Administration, Silver Spring, MD, USA
5. 5 National Institute on Drug Abuse, National Institutes of Health, Bethesda, MD, USA
1. Correspondence to Dr Karin A Kasza, Department of Health Behavior, Roswell Park Comprehensive Cancer Center, Buffalo, NY 14263, USA; karin.kasza{at}roswellpark.org

## Abstract

Objective To report on demographic and tobacco use correlates of cessation behaviours across tobacco products (cigarettes, electronic nicotine delivery systems (ENDS), cigars, hookah and smokeless tobacco) among the US population.

Design Data were drawn from the first three waves (2013–2016) of the Population Assessment of Tobacco and Health Study, a nationally representative, longitudinal cohort study of US youth (ages 12–17) and adults (ages 18+) . Past 30-day (P30D) tobacco users at Wave 1 (W1) or Wave 2 (W2) were included (n=1374 youth; n=14 389 adults). Generalised estimating equations were used to evaluate the association between demographic and tobacco use characteristics at baseline, with cessation behaviours at follow-up (discontinuing use, attempting to quit, quitting), over two 1-year periods (W1–W2, W2–Wave 3).

Results Among adult users of each type of tobacco product, frequency of use was negatively associated with discontinuing use. Among adult cigarette smokers, non-Hispanic white smokers, those with lower educational attainment and those with lower household income were less likely to discontinue cigarette use; ENDS use was positively associated with making quit attempts but was not associated with cigarette quitting among attempters; smokeless tobacco use was positively associated with quitting among attempters; tobacco dependence was negatively associated with quitting among attempters. Among youth cigarette smokers, tobacco dependence was negatively associated with making quit attempts.

Discussion Demographic correlates of tobacco cessation behaviours underscore tobacco use disparities in the USA. Use of ENDS and use of smokeless tobacco products are positively associated with some adult cigarette cessation behaviours.

• cessation
• disparities
• non-cigarette tobacco products

## Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

## Introduction

In 2013–2016, tobacco product cessation rates among youth and adults were lowest for cigarettes,1 followed by smokeless tobacco,2 electronic nicotine delivery systems (ENDS) among youth3 and cigars among adults4 and were highest for hookah among youth and adults.5 Understanding correlates of cessation behaviours across tobacco products can help in targeted intervention efforts and can enable researchers to better predict the potential impacts of regulatory actions and other public health efforts.

Population level tobacco cessation rates depend on the fraction of tobacco users who make a quit attempt and the fraction of those who quit when they try.6 For cigarette smoking, there are similarities and differences in factors associated with quit attempt and quit success rates.7 Two nationally representative longitudinal studies prospectively evaluated cigarette cessation behaviours among adults: the International Tobacco Control 4-Country survey8 9 and the Tobacco Use Supplement to the Current Population Survey (TUS-CPS)10 evaluated discontinuing cigarette smoking (ie, from use to non-use),8 10 attempting to quit cigarette smoking8 9 and quitting among attempters.8 9 These studies found that adults who are younger,8 have higher educational attainment9 and have lower tobacco dependence8 are more likely to attempt to quit cigarette smoking than their counterparts. Further, lower tobacco dependence8 and higher income9 are positively associated with quitting among attempters.

As the scope of tobacco products has expanded and concurrent use of multiple tobacco products has become common among tobacco users in the USA,11–15 large-scale prospective studies are beginning to evaluate how ENDS may impact cigarette cessation. Using the Population Assessment of Tobacco and Health (PATH) Study, Coleman et al 16 found that, among adult cigarette smokers, daily ENDS use and lower tobacco dependence were positively associated with cigarette smoking ‘not at all’ 1 year later. Also using PATH Study data, Buu et al 17 found that exclusive cigarette smokers who started using ENDS and used ENDS more frequently had lower cigarette consumption and tobacco dependence 1 year later than those who used ENDS less frequently. Benmarhnia et al 18 found that cigarette smokers who used ENDS to quit cigarette smoking had higher rates of not smoking for at least the past 30 days at 1 year follow-up than those who did not use ENDS. Using data from the National Health Interview Survey (NHIS) and the TUS-CPS, Johnson et al 19 found that, among cigarette smokers, ENDS use was associated with greater likelihood of making a quit attempt and with higher rates of not smoking for at least the past 90 days at 1-year follow-up. On the other hand, a meta-analysis of 38 studies concluded that e-cigarette use was associated with 28% lower odds of quitting cigarettes, though these studies varied in design and in how e-cigarette use was defined.20

Using the TUS-CPS, Messer et al 21 found that cigarette smokers who also used smokeless tobacco were more likely to make a cigarette quit attempt but were less likely to quit when they tried compared with those who did not use smokeless tobacco. Nationally representative data on cessation of non-cigarette tobacco products are limited; Coleman et al 16 found that non-daily ENDS use, non-use of combustible tobacco products and use of non-customisable ENDS devices were positively associated with using ENDS ‘not at all’ 1 year later.

Understanding factors associated with the transition from tobacco product use to non-use is important to efforts aimed at decreasing tobacco use prevalence in the population.22 The purpose of this study is to report on demographic and tobacco use correlates of the following cessation behaviours among the US population of youth (ages 12–17) and adult (ages 18+) tobacco users, for different types of tobacco products, using the PATH Study: (1) discontinuing tobacco product use overall, (2) attempting to quit tobacco product use and (3) quitting tobacco product use among those who attempted to quit.

## Methods

### Data source and participants

The PATH Study is an ongoing, nationally representative, longitudinal cohort study of youth and adults in the USA. Data were collected from 12 September 2013 through 14 December 2014 (Wave 1 (W1)); from 23 October 2014 through 30 October 2015 (Wave 2 (W2)) and from 19 October 2015 through 23 October 2016 (Wave 3 (W3)). The study uses audio computer-assisted self-interviews (ACASI) administered in English or Spanish to collect self-reported information on tobacco-use patterns and associated health behaviours. At W1, the overall weighted response rate was 78.4% for the youth interview and 74.0% for the adult interview; at W2, it was 87.3% for the youth interview and 83.2% for the adult interview and at W3, it was 83.3% for the youth interview and 78.4% for the adult interview. Further details regarding the PATH Study design and W1 methods are published elsewhere.23 Details on interview procedures, questionnaires, sampling, weighting, response rates and accessing the data are described in the PATH Study Restricted Use Files (RUF) User Guide, and Non-response Bias Analysis Reports for W1–W3 are available at https://doi.org/10.3886/Series606.24 The study was conducted by Westat and approved by the Westat Institutional Review Board. All respondents ages 18 and older provided informed consent, with youth respondents ages 12 to 17 providing assent and each youth’s parent/legal guardian providing consent. Data in this paper were drawn from respondents present in W1, W2 and W3 of the PATH Study, which includes 25 384 adults at W1 or W2 and 12 993 youth at W1 or W2.

This paper describes tobacco product cessation behaviours over two 1-year periods. The analytic sample was restricted to respondents who were past 30-day (P30D) users of at least one type of tobacco product at W1 or W2, which includes 14 389 adults and 1374 youth. W1 and W2 are each considered the ‘baseline’ wave to the subsequent wave, such that W1 is the baseline wave to W2, and W2 is the baseline wave to W3. Inclusion in the youth analyses versus the adult analyses was determined based on age at the respective baseline wave. (That is, youth who aged into the adult cohort at W2 were included in the youth analyses between W1 and W2 (N=352 current users) and were included in the adult analyses between W2 and W3 (N=334 current users). ‘Shadow youth’ who aged into the youth cohort at W2 (N=14 current users) were included in the youth analyses only between W2 and W3.) See online supplemental table 1 for additional details. The weighted estimates presented in this paper represent the civilian, non-institutionalised population of the USA at the time of W3 who were age 9 or older at W1, through application of population and replicate weights that adjust for the complex study design characteristics (eg, oversampling at W1) and non-response at W1, W2 and W3.

### Measures

#### Tobacco product use

Tobacco products were grouped into five types: cigarettes, ENDS (e-cigarettes at W1 and e-cigarettes, e-cigars, e-pipes and e-hookah at W2 and W3), cigars (traditional cigars, cigarillos and filtered cigars), hookah and smokeless tobacco (loose snus, moist snuff, dip, spit, chewing tobacco and snus pouches). For each of these five types of tobacco products, P30D use, current established use (adults only) and P30D non-light cigarette smoking (youth only) were defined, as shown in table 1. Current established use and P30D non-light cigarette smoking were additionally categorised according to frequency of use (table 1).

Table 1

Definitions

#### Cessation behaviours

The following cessation behaviours were assessed at follow-up (table 1): (1) discontinuing use (P30D use to no P30D use), (2) making a quit attempt and (3) quitting among quit attempters (current established use to no current established use among quit attempters (adults) and P30D non-light cigarette smoking to no P30D cigarette smoking among quit attempters (youth)).

#### Tobacco dependence

Sixteen items assessing symptoms of tobacco dependence were included in the PATH Study; responses to these items were combined and scaled to produce an overall tobacco dependence score as described and validated by Strong and colleagues.25

#### Demographic characteristics

Demographic characteristics were assessed at baseline wave and categorised as shown in the tables. Missing data on age, sex, race and Hispanic ethnicity at W1 were imputed as described in the PATH Study RUF User Guide (Imputed sex and race/ethnicity were carried forward to also represent these characteristics at W2; however, age at W2 was used since the time between interviews may not have yielded one additional year in all instances.).24

### Statistical analyses

For each type of tobacco product, generalised estimating equations (GEEs) were used to evaluate the association between correlates assessed at baseline and cessation behaviours assessed at follow-up, over two 1-year periods (W1–W2 and W2–W3). This statistical method allows for the inclusion of transitions from both periods in a single analysis while statistically controlling for interdependence among observations contributed by the same individuals.26 27 Specifically, GEE logistic regression models specified unstructured covariance and within-person correlation matrices and the binomial distribution of the dependent variable using the logit link function. Analyses were weighted using the W3 ‘all-wave’ weights (including full-sample and 100 replicate weights) to produce nationally representative estimates. Variances were computed using the balanced repeated replication method28 with Fay’s adjustment set to 0.3 to increase estimate stability.29 All analyses were conducted using SAS 9.4 software (SAS Institute, Cary, North Carolina, USA). See online supplementary material for the SAS macro code created to run weighted GEE analyses and calculate adjusted ORs (aORs) and CIs. Analyses were run on the W1–W3 RUF (https://doi.org/10.3886/ICPSR36231.v18).

For each type of tobacco product, cessation behaviours were evaluated with respect to the given tobacco product regardless of other products used (However, depending on whether respondents were exclusive or multiple product users, they may have been asked about their cessation behaviour regarding the specific product, or tobacco products generally (see table 1 for more details).). All analyses were conducted among adults and youth at baseline separately and were adjusted for all correlates plus wave (some significant wave effects were observed). Estimates with a relative standard error >30 or with a denominator <50 are suppressed since these estimates may provide unreliable precision and to protect respondent confidentiality.

## Results

### Discontinuing use

Among adult P30D users of each type of tobacco product at baseline, rates of discontinuing use of each product at approximately 1-year follow-up were 13.4% (95% CI: 12.8 to 14.1) for cigarettes, 52.6% (95% CI: 50.9 to 54.4) for ENDS, 48.6% (95% CI: 47.0 to 50.2) for cigars, 63.8% (95% CI: 61.6 to 66.0) for hookah and 25.9% (95% CI: 24.2 to 27.7) for smokeless tobacco (table 2).

Table 2

Correlates of discontinuing P30D use among P30D users (adults 18+) at baseline

##### Cigarettes

Those ages 25–39 or 40–54 had lower odds of discontinuing cigarette use than those ages 18–24 (aOR=0.7, 95% CI: 0.6 to 0.9; aOR=0.6, 95% CI: 0.5 to 0.7, respectively); Hispanic smokers had higher odds than non-Hispanic white smokers (aOR=1.3, 95% CI: 1.1 to 1.6), those with some college or associate degree (aOR=1.4, 95% CI: 1.1 to 1.6) or bachelor’s degree or more (aOR=1.5, 95% CI: 1.2 to 1.8) had higher odds than those with less than a high-school diploma, those with income US$25 000–US$74 999 (aOR=1.3, 95% CI: 1.1 to 1.5) or US$75 000+ (aOR=1.5, 95% CI: 1.3 to 1.9) had higher odds than those with income <US$25 000, and frequent cigarette smokers had much lower odds than infrequent smokers (aOR=0.2, 95% CI: 0.2 to 0.2), of discontinuing cigarette use (table 2).

##### ENDS

Non-Hispanic black ENDS users had higher odds than non-Hispanic white users (aOR=1.3, 95% CI: 1.0 to 1.8), those with income US$75 000+ had lower odds than those with income <US$25 000 (aOR=0.8, 95% CI: 0.6 to 1.0), those who used cigarettes had higher odds than those who did not use cigarettes (aOR=1.5, 95% CI: 1.2 to 1.8), those who used hookah had lower odds than those who did not use hookah (aOR=0.8, 95% CI: 0.6 to 1.0) and those who were frequent ENDS users had much lower odds than infrequent users (aOR=0.3, 95% CI: 0.3 to 0.4), of discontinuing ENDS use (table 2).

##### Cigars

Those ages 40–54 or 55+ had lower odds than those ages 18–24 (aOR=0.7, 95% CI: 0.6 to 1.0; aOR=0.4, 95% CI: 0.3 to 0.6, respectively), male users had lower odds than female users (aOR=0.8, 95% CI: 0.7 to 1.0), non-Hispanic black users had lower odds than non-Hispanic white users (aOR=0.7, 95% CI: 0.5 to 0.8), with those who identified as gay or lesbian had higher odds than those who identified as straight/heterosexual (aOR=1.6, 95% CI: 1.0 to 2.6), those who used hookah had lower odds than those who did not use hookah (aOR=0.8, 95% CI: 0.6 to 0.9) and those who were frequent cigar users had much lower odds than infrequent users (aOR=0.4, 95% CI: 0.3 to 0.5), of discontinuing cigar use (table 2).

##### Hookah

Non-Hispanic other users had lower odds than non-Hispanic white users (aOR=0.5, 95% CI: 0.3 to 1.0), and those who were frequent hookah users had lower odds than infrequent users (aOR=0.7, 95% CI: 0.4 to 1.0), of discontinuing hookah use (table 2).

##### Smokeless tobacco

Male users had lower odds than female users (aOR=0.7, 95% CI: 0.4 to 1.0), those who smoked cigarettes had higher odds than those who did not smoke cigarettes (aOR=1.9, 95% CI: 1.5 to 2.5) and those who were frequent smokeless tobacco users had much lower odds than infrequent users (aOR=0.2, 95% CI: 0.2 to 0.3), of discontinuing smokeless tobacco use (table 2).

#### Youth

Among youth P30D users of each type of tobacco product at baseline, rates of discontinuing use of each product at approximately 1 year follow-up were 32.5% (95% CI: 29.0 to 36.1) for cigarettes, 55.6% (95% CI: 51.5 to 59.6) for ENDS, 54.8% (95% CI: 49.1 to 60.3) for cigars, 70.8% (95% CI: 65.4 to 75.6) for hookah and 36.3% (95% CI: 29.2 to 44.1) for smokeless tobacco (online supplementary table 2). For youth, GEE models failed to converge for discontinuing use of cigars, hookah and smokeless tobacco; thus, correlates for discontinuing use of these products are not reported, though rates of discontinuing use for these products, stratified by correlate, are shown in online supplementary table 2. Significant correlates were age, with older youth less likely than younger youth, sexual orientation, with those who did not identify as straight/heterosexual less likely than those who did identify as straight/heterosexual, and frequency of cigarette use, with those who used 20–30 days less likely than those who used 1–19 days, to discontinue cigarette use (online supplementary table 2).

### Making a quit attempt

Among adult current established users of each type of tobacco product at baseline, rates of making a quit attempt at follow-up were 35.6% (95% CI: 34.6 to 36.7) for cigarettes, 45.8% (95% CI: 43.8 to 47.8) for ENDS, 56.2% (95% CI: 53.6 to 58.7) for cigars, 61.5% (95% CI: 58.6 to 64.3) for hookah and 39.2% (95% CI: 36.5 to 42.0) for smokeless tobacco (table 3).

Table 3

Correlates of making a quit attempt among current established users (adults 18+) at baseline

##### Cigarettes

Similar to correlates of discontinuing cigarette use, young adults (18–24 years), Hispanic smokers, those with more education and those who were not daily smokers at baseline all had higher odds of making a quit attempt than their counterparts (table 3). Additionally, those who used ENDS had higher odds of making a quit attempt than those who did not use ENDS (aOR=1.2, 95% CI: 1.1 to 1.3; table 3).

##### ENDS

Similar to correlates of discontinuing ENDS use, non-Hispanic black users, those who used cigarettes and those who were not daily ENDS users all had higher odds of making a quit attempt than their counterparts (table 3). Additionally, those with higher tobacco dependence scores had higher odds of making a quit attempt than those with lower scores (aOR=1.1, 95% CI: 1.0 to 1.3, table 3; not assessed for discontinuing use outcome).

##### Cigars

Somewhat similar to correlates of discontinuing cigar use, young adults (ages 18–24 years, compared with those ages 40–54 years or 55+ years) and those who were not daily cigar users (compared with daily users) had higher odds of making a quit attempt (table 3). Additionally, Hispanic users had higher odds than non-Hispanic white users (aOR=1.6, 95% CI: 1.1 to 2.3), those with a bachelor’s degree had lower odds than those with less than a high-school diploma (no diploma) or GED (aOR=0.6, 95% CI: 0.4 to 0.9), those with household income $75 000+ had lower odds than those with household income <US$25 000 (aOR=0.7, 95% CI: 0.5 to 1.0), those who used cigarettes had higher odds than those who did not use cigarettes (aOR=1.4, 95% CI: 1.0 to 1.9) and those with higher tobacco dependence scores had higher odds than those with lower scores (aOR=1.3, 95% CI: 1.1 to 1.5, not assessed for discontinuing use outcome) of making a quit attempt (table 3).

##### Hookah

Somewhat similar to correlates of discontinuing hookah use, non-Hispanic other users had lower odds than non-Hispanic white users of making a quit attempt (table 3). Additionally, those who used cigarettes had higher odds of making a quit attempt than those who did not use cigarettes (aOR=1.6, 95% CI: 1.1 to 2.4; table 3).

## Youth

Among youth P30D non-light cigarette smokers at baseline who made a quit attempt at follow-up, 38.9% (95% CI: 32.9 to 45.3) did not smoke in the P30D at follow-up (online supplementary table 4). GEE models failed to converge; thus, correlates for quitting among attempters are not reported, although percentages, stratified by correlate, are shown in online supplementary table 4.

## Discussion

PATH Study W1–W3 data show that, among the US population of adult P30D users of each type of tobacco product examined here (cigarettes, ENDS, cigars, hookah and smokeless tobacco), frequent users were less likely to discontinue use at approximately 1 year follow-up than infrequent users. Among cigarette smokers, non-Hispanic white smokers (compared with Hispanic smokers), those with lower educational attainment and those with lower household income were less likely to discontinue cigarette use, whereas among ENDS users, those with lower income were more likely to discontinue ENDS use than those with higher income, and among cigar smokers, non-Hispanic white smokers were more likely to discontinue cigar use than non-Hispanic black smokers.

For cigarette cessation behaviours, tobacco dependence was associated with a lower likelihood of cigarette quitting among those who attempted to quit, as shown before.8 Additionally, use of ENDS was associated with a greater likelihood of attempting to quit, which is consistent with prior research showing that some cigarette smokers use ENDS to help them quit smoking.19 30 31 However, we did not see an association between ENDS use and discontinuing cigarette smoking or quitting among attempters as others have found,18 though this is not surprising given that our descriptive analysis did not consider features of ENDS use that would be important when assessing ENDS use as a potential cessation aid, such as daily versus non-daily ENDS use.32

We did, however, find that smokeless tobacco use was associated with greater cigarette quitting among attempters, which differs from the Messer et al 21 finding and may reflect substitution of cigarette use with smokeless tobacco use.33–35 However, here we also found that cigarette smoking was associated with greater smokeless tobacco quitting among quit attempters. Future research can investigate possible explanations for these results by considering whether use of one type of tobacco product is intended to help quit use of another type.

While demographic correlates of cessation behaviours found here generally align with the limited prior national data on cigarette cessation behaviours among adults,8–10 the PATH Study data additionally show that age, sex, race/ethnicity and sexual orientation are significant correlates of discontinuing tobacco use among youth. Further, these data show that youth cigarette smokers with higher tobacco dependence were less likely to attempt to quit than those with lower tobacco dependence, though this finding may reflect those with lower dependence being less established smokers/experimental smokers who are no longer smokers at follow-up rather than those with lower dependence being more likely to attempt to quit per se.

### Limitations

The findings reported here cover broad cessation behaviours across a range of tobacco products, but this necessarily limits the depth of correlates that could be examined. Future studies can build on these foundational data; for example, studies can consider previous quit attempts, mental health and other substance use correlates, motivations for product use, intentions to quit product use, device features of ENDS, features of quit attempts such as use of cessation medications, and future studies can consider more stringent definitions of quitting such as abstinence for 12+ months.

Another limitation is the way in which quit attempts were queried; those who used multiple tobacco products are sometimes asked whether they tried to quit ‘tobacco’ rather than whether they tried to quit specific types of tobacco products, so in some cases, we were unable to attribute a quit attempt to a specific product. Indeed, we find that about 36% of adult cigarette smokers made a past-year quit attempt, while the NHIS data show that about 55% of adult cigarette smokers (defined the same way) made a past-year quit attempt in 2015.36 As with survey designs in general, it is also possible that some participants failed to recall having made a quit attempt.37

Despite these limitations, and taken together with correlates of tobacco product initiation38 and tobacco product relapse behaviours,39 findings underscore tobacco use disparities in the USA. For example, adults with low income are less likely to quit cigarette smoking when they try and are more likely to relapse after they quit.39 We also see that a history of tobacco product use predicts transitions both toward and away from using another tobacco product, and additional waves of data will track the stability/transience of tobacco use behaviours over longer periods of time.1–5 40

## Summary and implications

Demographic correlates of tobacco cessation behaviours underscore tobacco use disparities in the USA, such as socioeconomic disparities in cigarette use, which calls for renewed efforts to reduce tobacco use disparities. Further, findings show that the use of ENDS and use of smokeless tobacco products are positively related to some adult cigarette cessation behaviours.

• Among adult cigarette smokers, those who are younger, have higher educational attainment and lower tobacco dependence are more likely to attempt to quit than their counterparts, and lower tobacco dependence is also associated with greater quitting among quit attempters.

• Nationally representative longitudinal data on correlates of cessation behaviours across other tobacco products are limited.

• This paper reports on demographic and tobacco use correlates of cessation behaviours across cigarettes, electronic nicotine delivery systems (ENDS), cigars, hookah and smokeless tobacco among the US population of adults and youth.

• Findings indicate that among adult users of each type of tobacco product, frequent users were less likely to discontinue use than infrequent users; among adult cigarette smokers, those who used ENDS were more likely to attempt to quit but were not more likely to quit when they attempted compared with those who did not use ENDS, while those who used smokeless tobacco were more likely to quit when they attempted compared with those who did not use smokeless tobacco.

• Among adult ENDS users, those with lower income were more likely to discontinue ENDS use than those with higher income; among cigar smokers, non-Hispanic white smokers were more likely to discontinue cigar use than non-Hispanic black smokers.

• Among youth cigarette smokers, those with higher tobacco dependence were less likely to attempt to quit than those with lower tobacco dependence.

## Footnotes

• Contributors KAK and KCE led the conceptual design of the work. KAK drafted the initial manuscript and all authors critically revised it. ZT conducted the statistical analyses and all authors contributed to the interpretation of results. All authors approved the work for journal publication and agree to be accountable for all aspects of the work.

• Funding This manuscript is supported with Federal funds from the National Institute on Drug Abuse, National Institutes of Health, and the Center for Tobacco Products, Food and Drug Administration, Department of Health and Human Services, under a contract to Westat (Contract No: HHSN271201100027C).

• Disclaimer The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the U.S. Department of Health and Human Services or any of its affiliated institutions or agencies.

• Competing interests WMC reports long-term stock holdings in General Electric Company, 3M Company, and Pfizer Incorporated, unrelated to this manuscript. No financial disclosures were reported by the other authors of this paper.

• Patient consent for publication Not required.

• Ethics approval The study was conducted by Westat and approved by the Westat Institutional Review Board. All respondents aged 18 and older provided informed consent, with youth respondents ages 12 to 17 providing assent and each youth’s parent/legal guardian providing consent.

• Provenance and peer review Not commissioned; externally peer reviewed.

• Data availability statement Data from the PATH Study Wave 1 to Wave 3 may be obtained from a third party and are not publicly available (https://www.icpsr.umich.edu/icpsrweb/NAHDAP/studies/36231). Applications instructions and conditions of use are available at the website above.