Objective Electronic nicotine delivery systems (ENDS; including e-cigarettes) are rapidly evolving in the US marketplace. This study reports cross-sectional prevalence and longitudinal pathways of ENDS use across 3 years, among US youth (12–17 years), young adults (18–24 years) and adults 25+ (25 years and older).
Design Data were 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 and adults. Respondents with data at all three waves (youth, n=11 046; young adults, n=6478; adults 25+, n=17 188) were included in longitudinal analyses.
Results Weighted cross-sectional ever use of ENDS increased at each wave. Across all three waves, young adults had the highest percentages of past 12-month, past 30-day (P30D) and daily P30D ENDS use compared with youth and adults 25+. Only about a quarter of users had persistent P30D ENDS use at each wave. Most ENDS users were polytobacco users. Exclusive Wave 1 ENDS users had a higher proportion of subsequent discontinued any tobacco use compared with polytobacco ENDS users who also used cigarettes.
Conclusions ENDS use is most common among young adults compared with youth and adults 25+. However, continued use of ENDS over 2 years is not common for any age group. Health education efforts to reduce the appeal and availability of ENDS products might focus on reducing ENDS experimentation, and on reaching the smaller subgroups of daily ENDS users to better understand their reasons for use.
- electronic nicotine delivery devices
- non-cigarette tobacco products
- surveillance and monitoring
Statistics from Altmetric.com
- electronic nicotine delivery devices
- non-cigarette tobacco products
- surveillance and monitoring
Electronic nicotine delivery systems (ENDS), including e-cigarettes, are battery-operated products that produce an aerosolised mixture containing nicotine, flavourings, propylene glycol, glycerin and/or other additives that the user inhales.1 In the USA, the tobacco marketplace has seen a rapidly evolving array of ENDS products.2 US representative cross-sectional surveys have shown ever use of e-cigarettes is significantly increasing among youth (12–17 years).3–8 For both high school students and middle school students, ENDS have been the most frequently used tobacco products since 2014.3 6 7 9 10 The 2011–2018 National Youth Tobacco Survey data show significant non-linear increases in past 30-day (P30D) ENDS use among high school students from 1.5% to 20.8% and among middle school students from 0.6% to 4.9%.7 Aggregated initiation rates of past 12-month (P12M) ENDS use across the approximate 3 years’ time span (2013–2016) of Wave 1 to Wave 3 (W1–W3) of the Population Assessment of Tobacco and Health (PATH) Study indicate that most ENDS initiation occurs among youth (ages 12–17) and young adults (ages 18–24) (youth, 22.7%; young adults, 28.4%; adults 25+, 6.7%) and that compared with other tobacco products, ENDS have shown the highest initiation rates during this time period.11 P30D ENDS use has been reported to be less frequent compared with combustible cigarette use, with the vast majority of ENDS use being non-daily.12–14
Most US youth and adult ENDS users are dual or polytobacco product users.1 15 16 Among US youth,15 17 P30D use of cigarettes and ENDS is the most common product combination (15.0%).16 Use of ENDS with combustible cigarettes can result in continued exposure to the toxic combustion products of traditional cigarettes and exposure to unique constituents in ENDS (eg, flavours, humectants).18 19 Tobacco users who use both ENDS and combustible cigarettes may be using ENDS to gradually substitute ENDS for cigarettes with the intention to reduce or quit cigarettes, or may be adding ENDS to existing cigarette use as a way to cope with smoke-free policies that create discomfort due to nicotine dependence.18 Data on the long-term health effects of dual combustible cigarette and ENDS use are limited, with some studies showing no reduced risk from dual use compared with cigarette smoking alone,20–22 and others suggesting dual use may be associated with more risk of negative health effects.23–25
The first aim of this study is to examine differences between each of the first three waves of cross-sectional weighted estimates of ENDS use from the PATH Study (2013–2016). Differences over time are reported for different definitions of ENDS use, such as ever, P12M, P30D and daily P30D ENDS use for US youth, young adults and adults 25+. Drawing from the first three waves of longitudinal within-person data from the PATH Study, the second aim is to examine age group differences in W1-W2-W3 pathways of persistent use, discontinued use and reuptake of ENDS among W1 P30D ENDS users. The final aim of these analyses is to compare longitudinal transition pathways among W1 exclusive ENDS users, W1 ENDS polytobacco users including cigarettes (ENDS polytobacco use w/CIGS) and W1 ENDS polytobacco users who do not use cigarettes (ENDS polytobacco use w/o CIGS) to understand product transitions. Comparing longitudinal transitions among exclusive versus polytobacco ENDS users will advance our understanding of patterns of polytobacco use in the USA and critical product transitions, such as switching and complete tobacco cessation. These analyses will lay the groundwork for more robust evaluations of the potential health risks and benefits of ENDS use at the population level.
Study design and population
The PATH Study is an ongoing, nationally representative, longitudinal cohort study of youth (ages 12–17) and adults (ages 18 or older) in the USA. Self-reported data were collected using Audio Computer-Assisted Self-Interview administered in English and Spanish. Further details regarding the PATH Study design and W1 methods are published elsewhere.26 27 At W1, the weighted response rate for the household screener was 54.0%. Among screened households, the overall weighted response rate was 78.4% for youth and 74.0% for adults at W1, 87.3% for youth and 83.2% for adults at W2 and 83.3% for youth and 78.4% for adults at W3. Details on interview procedures, questionnaires, sampling, and weighting and information on accessing the data are available at https://doi.org/10.3886/Series606.The study was conducted by Westat and approved by the Westat Institutional Review Board. All participants ages 18 and older provided informed consent, with youth participants ages 1217 providing assent while their parent/legal guardian provided consent.
The current study reports cross-sectional estimates from 13 651 youth and 32 320 adults who participated in W1 (data collected 12 September 2013 through 14 December 2014), 12 172 youth and 28 362 adults at W2 (23 October 2014 through 30 October 2015) and 11 814 youth and 28 148 adults at W3 (19 October 2015 to 23 October 2016). The differences in the number of completed interviews between W1, W2 and W3 reflect attrition due to non-response, mortality and other factors, as well as youth who enrol in the study at W2 or W3.26 We also report longitudinal estimates from W1 youth (n=11 046), W1 young adults (n=6478) and W1 adults 25+ (n=17 188) with data collected at all three waves. See online supplementary figure 1 for a detailed description of the analytic sample for longitudinal analysis.
At each wave, adults and youth were asked about their tobacco use behaviours for cigarettes, ENDS, traditional cigars, cigarillos, filtered cigars, pipe tobacco, hookah, snus pouches, other smokeless tobacco (loose snus, moist snuff, dip, spit or chewing tobacco) and dissolvable tobacco. Participants were asked about ‘e-cigarettes’ at W1 and ‘e-products’ (e-cigarettes, e-cigars, e-pipes and e-hookah) at W2 and W3; all electronic products are referred to as ENDS in this paper. In addition, youth were asked about their use of bidis and kreteks but these data were not included in the analyses due to small sample sizes.
At W1, ENDS were described as ‘e-cigarettes that look like regular cigarettes, but are battery-powered and produce vapor instead of smoke. Some common brands include NJOY, Blu, and Smoking Everywhere’. At W2 and W3, ENDS were described as ‘electronic nicotine products such as e-cigarettes, e-cigars, e-pipes, e-hookahs, and personal vaporizers, as well as vape pens and hookah pens that are battery-powered, use nicotine fluid rather than tobacco leaves, and produce vapor instead of smoke. Some common brands include Fin, NJOY, Blu, e-Go and Vuse’. Participants were shown generic pictures of the product at all three waves.
Cross-sectional definitions of use included ever, P12M, P30D and daily P30D use. Longitudinal outcomes included persistent ENDS use, discontinued ENDS use and reuptake of ENDS use, as well as transitions among exclusive and polytobacco ENDS users. The definition of each outcome is included in the footnote of the table/figure in which it is presented.
To address Aim 1, weighted cross-sectional prevalence of ENDS use was compared across waves for each age group for ever, P12M, P30D and daily P30D use. For Aim 2, irrespective of other tobacco product use, longitudinal W1-W2-W3 transitions in P30D ENDS use were compared by age group within three separate user groups persistent any P30D ENDS use (defined as continued P30D ENDS use at W2 and W3), discontinued any P30D ENDS use (stopped ENDS use at W2 and W3 or just W3) and reuptake of any P30D ENDS use (used ENDS at W1, discontinued ENDS use at W2 and used ENDS again at W3). Finally, to address Aim 3, longitudinal W1-W2-W3 ENDS use pathways that flow through seven mutually exclusive and exhaustive transition categories were examined for W1 P30D exclusive ENDS use, W1 P30D ENDS polytobacco use w/CIGS and W1 P30D ENDS polytobacco use w/o CIGS (see online supplementary figure 2). For each aim, weighted t-tests were conducted on differences in proportions to assess statistical significance. To correct for multiple comparisons, Bonferroni post hoc tests were conducted. Given that cigarettes are the most commonly used tobacco product with the most robust evidence base of potentially harmful health consequences,5 two polytobacco use groups were examined separately to compare longitudinal transitions among polytobacco users who use and do not use cigarettes. These pathways represent building blocks that may be aggregated to reflect higher level behavioural transitions.
Cross-sectional estimates (Aim 1) were calculated using PATH Study cross-sectional weights for W1 and single-wave (pseudo-cross-sectional) weights for W2 and W3. The weighting procedures adjusted for complex study design characteristics and non-response. Combined with the use of a probability sample, the weighted data allow these estimates to be representative of the non-institutionalised, civilian, resident U.S population aged 12 or older at the time of each wave. Longitudinal estimates (Aims 2 and 3) were calculated using the PATH Study W3 all-waves weights. These weighted estimates are representative of the resident US population aged 12 and older at the time of W3 (other than those who were incarcerated) who were in the civilian, non-institutionalised population at W1.
All analyses were conducted using SAS Survey Procedures V.9.4 (SAS Institute). Variances were estimated using the balanced repeated replication method28 with Fay’s adjustment set to 0.3 to increase estimate stability.29 Analyses were run on the W1–W3 Public Use Files (https://doi.org/10.3886/ICPSR36498.v8). Estimates with low precision (fewer than 50 observations in the denominator or with a relative standard error greater than 0.30) were flagged and are not discussed in the Results section.
Cross-sectional weighted prevalence
As shown in figure 1, across age groups, ever use of ENDS significantly increased at each wave. Prevalence of P12M ENDS use did not change in young adults across waves, although among youth it increased 2.5% from W1 to W3, and among adults 25+ it dropped 2.8% at W3 from W1 and W2. Prevalence of P30D use did not change in youth or adults 25+ between W2 and W3, but among young adults P30D use increased by 4.9% at W3 compared with W2. Daily P30D ENDS use among young adults increased from 1.4% (95% CI 1.2 to 1.7) at W1 to 2.5% (95% CI 2.1 to 3.1) at W2 and to 3% (95% CI 2.6 to 3.4) at W3. Across all three waves, young adults had the highest percentages of ever, P12M, P30D and daily P30D ENDS use. Described differences are absolute per cent differences, not relative per cent differences.
Longitudinal weighted W1-W2-W3 pathways
Among any P30D ENDS users at W1
Among those with data at all three waves, 3.0% (95% CI 2.6 to 3.4) of youth, 12.5% (95% CI 11.6 to 13.5) of young adults and 5.8% (95% CI 5.5 to 6.1) of adults 25+ had P30D ENDS use at W1. As illustrated in figure 2, persistent P30D ENDS use, defined as P30D ENDS use at all three waves, irrespective of concurrent use of other products, was similar across each age group. Discontinued ENDS use, defined as stopping ENDS use at W2 or W3 among W1 P30D ENDS users, was higher among young adults (62.1% (95% CI 58.7 to 65.5)) and adults 25+ (63.0% (95% CI 60.4 to 65.4)) compared with youth (54.1% (95% CI 48.4 to 59.7)). ENDS reuptake, defined as ENDS use at W1, no ENDS use at W2 and ENDS use again at W3, was lowest among adults 25+ (10.1% (95% CI 8.8 to 11.6)) compared with youth (18.0% (95% CI 14.3 to 22.5)) and young adults (15.0% (95% CI 12.8 to 17.4)).
Among Wave 1 P30D ENDS user types (exclusive, polytobacco users w/CIGS and polytobacco users w/o CIGS)
Among the longitudinal sample of W1 P30D ENDS users, most were ENDS users who also used another tobacco product: 63.8% (95% CI 58.8 to 68.5) of youth, 87.4% (95% CI 85.0 to 89.35) of young adults and 86.2% (95% CI 84.5 to 87.7) of adults 25+. Among ENDS polytobacco users, 70.6% (95% CI 63.1 to 77.1) of youth, 84.7% (95% CI 82.1 to 86.9) of young adults and 94.8% (95% CI 93.4 to 95.9) of adults 25+ also used cigarettes. Less than half of youth and young adult ENDS users who also used cigarettes used only cigarettes and e-cigarettes, but most adults 25+ (69.4% (95% CI 66.6 to 72.0)) with ENDS polytobacco use w/CIGS used only cigarettes and e-cigarettes.
To address the third aim and compare user types, 49 possible W1-W2-W3 pathways were examined across seven mutually exclusive categories (see conceptual map in online supplementary figure 2) among three separate W1 user type categories: (1) P30D exclusive ENDS users (online supplementary table 1a), (2) P30D ENDS polytobacco users w/CIGS (online supplementary table 1b), and (3) P30D ENDS polytobacco users w/o CIGS (online supplementary table 1c). Pathways from online supplementary table 1a–c estimate broad behavioural transitions such as persistent use type, discontinued all tobacco use and tobacco use reuptake which were compared across the three user types.
Among youth (table 1), W1 exclusive P30D ENDS users had higher rates of discontinued all tobacco use (53.9% (95% CI 44.6 to 63.0)) compared with W1 P30D ENDS polytobacco users w/CIGS (17.9% (95% CI 12.2 to 25.6)) and W1 P30D ENDS polytobacco users w/o CIGS (28.3% (95% CI 16.6 to 43.8)).
Among young adults (table 1), W1 exclusive P30D ENDS users had higher rates of discontinued all tobacco use (45.8% (95% CI 35.3 to 56.5)) compared with W1 P30D ENDS polytobacco users w/CIGS (11.9% (95% CI 9.5 to 14.9)) and W1 P30D ENDS polytobacco users w/o CIGS (27.2% (95% CI 18.8 to 37.6)). Discontinuing ENDS use but continuing other tobacco use was the highest among W1 ENDS polytobacco users w/CIGS (44.8% (95% CI 40.8 to 48.8)) than ENDS polytobacco users w/o CIGS (28.0% (95% CI 19.9 to 37.9)) and exclusive ENDS users who switch to other tobacco products by W3 (22.5% (95% CI 15.6 to 31.2)).
Among adults 25+ (table 1), W1 P30D ENDS polytobacco users w/CIGS had the lowest rates of discontinued all tobacco use (8.3% (95% CI 6.9 to 9.9)) compared with W1 exclusive P30D ENDS users (34.0% (95% CI 28.3 to 40.1)) and polytobacco users w/o CIGS (28.8% (95% CI 18.7 to 41.4)). Similar to young adults, switching from exclusive ENDS to other tobacco products was the lowest among W1 exclusive ENDS users (9.3% (95% CI 6.4 to 13.5)) compared with W1 ENDS polytobacco users w/CIGS (54.1% (95% CI 51.2 to 57.0)) and ENDS polytobacco users w/o CIGS (40.1% (95% CI 27.5 to 54.2)). Persistent use was higher among W1 exclusive ENDS users (34.3% (95% CI 29.2 to 39.9)) compared with ENDS polytobacco users w/CIGS (18.8% (95% CI 16.9 to 21.0)).
As rates of smoking cigarettes continue to decline among US youth and adults,6 30 nationally representative cross-sectional surveys have shown an increase in both initiation (new ever use) and P30D ENDS use over the past 5 years, especially among youth and young adults.6 20 31 Consistent with other reports, data from the first three waves of the PATH Study found that ever use of ENDS increased at each wave for youth, young adults and adults 25+. ENDS product availability, device sophistication, e-liquid variability (eg, nicotine salts) and marketing also increased from 2013 to 2016, making access and product awareness a potential reason for increased new ever use over this time period.32–36
Differences in trends of ENDS use based on the definition of use were also noted. For example, prevalence of P12M ENDS use did not change among young adults across waves, although among youth it increased 2.5% from W1 to W3, and among adults 25+ it dropped 2.8% at W3. Prevalence of P30D use did not change in youth or adults 25+ from W2 to W3, but among young adults, there was a 4.9% increase in P30D ENDS use at W3 compared with W2. While generally less than 3% of P30D ENDS use is daily, daily P30D ENDS use among young adults increased incrementally at each wave (W1, 1.4% (95% CI 1.2 to 1.7); W2, 2.5% (95% CI 2.1 to 3.1); W3, 3.0% (95% CI 2.6 to 3.4)); prevalence did not change in youth or adults 25+. Extending previous reports1 compared with youth and older adults, young adults had the highest percentages of P12M, P30D and daily P30D ENDS use at each of the three waves. Thus, young adults are a subpopulation of significant concern if these rising rates of use lead to the potential harms of early exposure to e-cigarette toxins and the well-established harms of subsequent combustible tobacco use.18
Longitudinal patterns across the three waves revealed that P30D ENDS use is not stable, with only about a quarter of users having persistent P30D ENDS use at each wave. The most common pattern of use across the three waves was marked by discontinued use, with about half of youth and about 60% of all adults discontinuing W1 ENDS use for the next 2 years at W2 and W3 or for 1 year at W3. Less than 20% (18.0% (95% CI 14.3 to 22.5)) of youth, 15.0% (95% CI 12.8 to 17.4) of young adults and 10.1% (95% CI 8.8 to 11.6) of adults 25+ who were W1 P30D ENDS users discontinued ENDS at W2 and then returned to P30D ENDS use at W3. These findings are similar to Coleman et al’s findings, in which half of adult e-cigarette users at W1 had discontinued their use of e-cigarettes at W2 and approximately half of W1 dual e-cigarette and cigarette users had discontinued ENDS use at W2.37 Thus, despite rising rates of ever use of ENDS, patterns over this 3-year period (2013–2016) showed that ENDS use does not persist among most who try the product.
The majority of ENDS users in each age group also used another tobacco product. There were notably different patterns of ENDS use for exclusive versus polytobacco users (table 1). Among all age groups, more W1 exclusive ENDS users stopped using all tobacco products compared with those with W1 ENDS polytobacco use w/CIGS. Patterns of use among W1 ENDS polytobacco users w/CIGS were marked by inconsistent use across the waves and discontinued ENDS use but continued use of other tobacco products. Other studies using PATH Study data have shown similar trends in discontinued use between dual users of ENDS and cigarettes compared with exclusive ENDS users.37 Moreover, persistent use of ENDS polytobacco use w/CIGS was more common than ENDS polytobacco use w/o CIGS (table 1). These results suggest that ENDS users also using other tobacco products have patterns of use that are not likely to include quitting all tobacco and that these users are more likely to persist either using ENDS with cigarettes or switching to other tobacco products.
Longitudinal patterns of ENDS use over the first three waves of the PATH Study (2013–2016) differ notably from trends presented in parallel reports of PATH Study data focused on other tobacco products.38–41 The most common ENDS use patterns over 3 years were either discontinued use, inconsistent use or switching to other products. In contrast, combustible cigarettes and smokeless tobacco products, products that have been on the market in the USA for decades, were used persistently across the waves. Products with rising market presence in the USA, such as hookah and cigars (including little cigars and cigarillos), have more inconsistent patterns of use similar to those of ENDS, particularly among young adults. The literature reports that product characteristics (eg, flavours),42 cost and taxation,43 44 access in social settings such as vape shops and hookah bars,45 as well as differences in targeted marketing and risk perceptions46–48 may be driving differences between traditional products and these emerging products, which are poised for intervention to reduce young adults’ attraction to these emerging products.
These data predate the rapid increase in ENDS from 2017 to 2018 among youth2 7 20 and transition patterns among youth may be different over time. Specifically, data collection occurred before the explosive growth of Juul, which delivers a high dose of nicotine and therefore has a greater potential for addictiveness compared with other ENDS brands, and may reduce rates of discontinued use and increase rates of reuptake.49–52 The potential for recall bias from a self-report questionnaire is noted. Additionally, the PATH Study asked about ‘e-cigarettes’ (the predominant e-product on the market) at W1 and ‘e-products’ at W2 and W3 which may have resulted in misclassification of ENDS product-specific use between W1 and the subsequent waves. Weighted longitudinal analyses excluded participants who were missing data at one of the waves. The extent of missing data and the small number of observations for low-prevalence pathways may limit interpretation. Other reports suggest that the frequency and intensity of ENDS use may be a critical factor in pathways such as discontinued use and switching.17 20 31 Future studies can examine adjusted models to determine which factors predict priority pathways, including frequency of use and other factors that may drive different patterns of use. Kasza et al 43 44 and Edwards et al 45 examine demographic correlates of initiation, cessation and relapse to further explore predictors of these critical outcomes.
Summary and implications
Evidence to date suggests that ever use of ENDS among youth may contribute to ever combustible tobacco use. This report identifies patterns of P30D ENDS use that are unstable, with common longitudinal pathways of quitting all tobacco, switching to other products or ENDS polytobacco use with cigarettes. While ever use of ENDS has increased among US youth and young adults,11 only a small percentage of the population (less than 1.0% of youth, 3.0% of young adults and 1.6% of adults 25+) were using ENDS daily. Differences in patterns of use among exclusive ENDS users and ENDS polytobacco users were identified, suggesting it is important to make these distinctions. Health education efforts to reduce the appeal and availability of ENDS products might focus on reducing ENDS experimentation, and on reaching the smaller subgroups of ENDS users who are using daily to better understand their reasons for use.
What this paper adds
This study includes a three-wave examination of electronic nicotine delivery systems (ENDS) use in the USA across multiple definitions of use for three different age groups. Across all three waves, young adults had the highest percentages of past 12-month, past 30-day (P30D) and daily P30D ENDS use compared with youth and adults 25+.
While rates of ever use of ENDS increased within each age group, only young adults increased P30D use between W2 and W3. Daily P30D use remained very low (less than 3%) for all age groups.
Longitudinal pathways indicate that P30D ENDS use is not stable, as only about a quarter of users showed persistent P30D ENDS use across the three waves.
The majority of ENDS use is polytobacco use, and ENDS polytobacco users who also use cigarettes are less likely to stop using tobacco 2 or 3 years later compared with exclusive ENDS users.
Contributors CAS and ES led the conceptual design. CAS drafted initial manuscript and all authors critically revised it. ES and MJH conducted statistical analysis and all authors contributed to 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 US 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 participants ages 18 and older provided informed consent, with youth participants ages 12 to 17 providing assent while their parent/legal guardian provided consent.
Provenance and peer review Not commissioned; internally peer reviewed.
Data availability statement Data from the PATH Study Wave 1 to Wave 3 are available for download as Public Use Files in a public, open access repository (https://www.icpsr.umich.edu/icpsrweb/NAHDAP/studies/36498). Conditions of use are available at the website above.
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.