Article Text
PDF
Abstract
Introduction The use of cigarettes and electronic nicotine delivery system (ENDS) has likely changed since 2019 with the rise of pods and disposables, the lung injuries outbreak, flavour bans, Tobacco 21 and the COVID-19 pandemic.
Methods Using the Population Assessment of Tobacco and Health Study, we applied a multistate transition model to 28 061 adults in waves 4–5 (2017–2019) and 24 584 adults in waves 5–6 (2019–2021), estimating transition rates for initiation, cessation and switching products for each period overall and by age group.
Results Cigarette initiation among adults who never used either product decreased from 2017–2019 to 2019–2021, but ENDS initiation did not significantly change. The persistence of ENDS-only use remained high (75%–80% after 1 year). Cigarette-only use transitions remained similar (88% remaining, 7% to non-current use and 5% to dual or ENDS-only use). In contrast, dual use to ENDS-only transitions increased from 9.5% (95% CI 7.3% to 11.7%) to 20.0% (95% CI 17.4% to 22.6%) per year, decreasing the persistence of dual use. The dual to cigarette-only use transition remained at about 25%. These changes were qualitatively similar across adult age groups, though adults ages 18–24 years exhibited the highest probability of switching from cigarette-only use to dual use and from dual use to ENDS-only use.
Conclusions The persistence of ENDS use among adults remained high in 2019–2021, but a larger fraction of dual users transitioned to ENDS-only use compared with 2017–2019. Because the fraction of cigarette-only users switching to dual use remained low, especially among older adults, the public health implications of this change are minimal.
- Cessation
- COVID-19
- Electronic nicotine delivery devices
- Surveillance and monitoring
Data availability statement
Data are available in a public, open access repository. Public Use Files from the Population Assessment of Tobacco and Health Study are available for download from an open access repository (https://doi.org/10.3886/ICPSR36498.v20). Conditions of use are available on the aforementioned website.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Recent changes in the marketplace, as well as events including the lung injury outbreak, COVID-19 pandemic and electronic nicotine delivery system (ENDS) flavour restrictions, may have impacted real-world product transition patterns.
Underlying transition hazard rates can explain changes in observed transition probabilities between different patterns of cigarettes and ENDS use.
WHAT THIS STUDY ADDS
We found decreased cigarette initiation, increased cessation from ENDS-only use and increased transitions from dual to ENDS-only use among adults overall.
ENDSs uptake among adult cigarette users remained low (<5% overall) so the population public health impact of the increased dual to ENDS-only use transition is likely to be limited.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Policies are needed to support transitions from dual use to non-current or ENDS-only use.
Policies are needed to reduce cigarette-related harm among older adults without increasing ENDS-related harm among people who would not have initiated ENDS otherwise.
Introduction
The landscape of tobacco and nicotine products in the USA and other countries has evolved quickly over the past decade. Electronic nicotine delivery systems (ENDSs), including e-cigarettes, have transformed from cig-a-likes, refillable and tank systems—which initially used freebase nicotine that was unpalatable at higher nicotine concentrations—to pod mods and disposables that use nicotine salts, often perceived as less harsh.1–5 As ENDS products have changed, so too have patterns of transitions between ENDS and cigarette products.6 7 Moreover, recent events and policies, including the lung injury outbreak in 2019,8 the COVID-19 pandemic, the federal Tobacco 21 policy and the increasing restriction of ENDS flavours,9 may have further shifted patterns of use and transition between ENDS and cigarettes.
ENDS products cause cardiopulmonary harm10–12 but also appear to be less harmful than cigarettes.13–15 Hence, understanding product transitions is key to determining the likely public health benefit or harm of ENDS.16–18 The living Cochrane review currently suggests that ENDS aid cigarette cessation in a clinical trial context,19 but studies have suggested that the effectiveness of ENDS for cigarette cessation may be less outside of clinical trials.20 21 If ENDS use promoted cigarette cessation, reduced smoking or diverted those who would have otherwise smoked, there would likely be a benefit to public health.22–25 However, there remain concerns about youth initiation and subsequent negative health impacts, particularly as many flavours are targeted to youth,26 27 as well as concerns that ENDS may interfere with long-term cigarette cessation because of continued or enhanced nicotine addiction.28–30 ENDS products continue to evolve and so continued analysis of how actual transition rates change over time is important for tracking real-world associations between ENDS and other product use and providing the information necessary to project future public health outcomes.
In previous work, we analysed transitions in the Population Assessment of Tobacco and Health (PATH) Study waves 1–5 (2013–2019), a nationally representative longitudinal survey of the USA, using a multistate transition model.6 7 Multistate transition models estimate the transition rates that underly observed transitions between product.31–40 In this analysis, we extend our models to compare transitions observed in PATH in 2017–2019 (waves 4–5) to 2019–2021 (waves 5–6).
Methods
Data
The PATH Study is a nationally representative longitudinal cohort study of tobacco and nicotine product use behaviours in the USA among the civilian, non-institutionalised adult population. Our analysis compared two samples. The first sample was 28 061 adults in the wave 4 cohort who completed wave 5; we analysed transitions between waves 4 and 5 (December 2016–November 2019, abbreviated as 2017–2019) for this sample. The second sample was 24 584 adults in the wave 4 cohort who completed both waves 5 and 6; we analysed transitions between waves 5 and 6 (December 2018–November 2021, abbreviated as 2019–2021) for this sample. Although this analysis primarily focuses on the differences between 2017–2019 and 2019–2021, we include previous results in 2015–2017 in figure 1 and the supplementary material for comparison purposes.7
Transition hazard rates among adults in 2015–2017 (waves 2–4; estimates from Brouwer et al7), 2017–2019 (waves 4–5) and 2019–2021 (waves 5–6). ENDS, electronic nicotine delivery system.
Follow-up time for participants was approximately 2 years between waves 4 and 5 and between waves 5 and 6, and follow-up time was treated as exactly 2 years in the model. The multistate transition framework explicitly accounts for the time between observations and for potential unobserved transitions, so 1-year transition probabilities can be estimated from 2-year data. We categorised participants as ages 18–24, 25–34 or 35–90 years. This set of groups was chosen because our transition estimates were not well powered for participants ages 55–90 alone in 2019–2021. We also used information on race and ethnicity (categorising participants as Hispanic, non-Hispanic black, non-Hispanic white and other/unknown) and gender.
Each participant’s product use was categorised as never established use (of either product), non-current use, cigarette-only use, ENDS-only use or dual use of cigarettes and ENDS, as in previous work.6 7 Established cigarette use was defined as having smoked at least 100 cigarettes in one’s lifetime and established ENDS use was defined as ever ‘fairly regularly’ using ENDS. Current use of cigarettes and ENDS was defined as any past 30-day use by an established user of that product, and current dual use was defined as the current use of both cigarettes and ENDS by established users of both products. Non-current use was defined as no past 30-day use of either product by a participant who had established use of one or both products. Daily use of each product was defined as use in 30 of the past 30 days.
Characteristics of the populations are given in online supplemental table S1.
Supplemental material
Transition modelling
We applied our multistate transition model to analyse the underlying transition hazard rates between product use for adults overall and in each of the three age categories (18–24, 25–34 or 35–90 years) in each of the two time periods (2017–2019 and 2019–2021). Multistate transition models are finite-state, stochastic process models that assume that transition hazard rates depend only on the current state and not on past states or transition history. Technical details of the multistate transition model are provided in online supplemental material. Using discrete-time observations, the model estimated the instantaneous risk of transition from one state to another, that is, transition hazard rates, which collectively define the probability of transitioning from one state to any other at a future time. Because all rates are estimated simultaneously, there is no need to adjust for multiple comparisons. We incorporated wave 4 longitudinal survey weights into the model. Using the estimated transition rates, we estimated the 1-year transition probabilities for each group for the two time periods separately. Additionally, two-sided p values for whether transition rates were significantly different between periods were calculated as described in online supplemental material using the replicate weights provided by PATH. We also estimated covariate HRs for the whole adult population in the two time periods, adjusting for gender, age group, race/ethnicity and daily versus non-daily cigarette and ENDS use. Multistate transition models were estimated using the wmsm function6 in R (publicly available at https://tcors.umich.edu/Resources_Research.php), which is an extension of the msm function41 modified to incorporate participant weights.
Results
In 2019 (wave 5), the weighted prevalence of never established use was 57.7%, the prevalence of non-current use was 23.7%, the prevalence of cigarette-only use was 14.2%, the prevalence of ENDS-only use was 2.5% and the prevalence of dual cigarette and ENDS use was 1.9%.
Transition hazard rates
There were fewer changes between 2017–2019 and 2019–2021 than there were between 2015–2017 and 2017–2019. Three transitions statistically significantly changed between 2017–2019 and 2019–2021 (figure 1; p values given in online supplemental figure S1). The transition rate of cigarette initiation declined for adults in 2019–2021 (p=0.02). The transition rate from ENDS-only use to non-current use increased in 2019–2021 (p=0.005). The dual use to ENDS-only use transition also increased in 2019–2021 rate (p<0.001).
Transition probabilities
The changing transition rates resulted in changes in the 1-year transition probabilities (figure 2). The transition patterns for those with never established use, non-current use and cigarette-only use remained largely similar between the two periods, with the large majority remaining in the same use state (never established use persistence: 98.8% (95% CI 98.6% to 98.9%) and 98.6% (95% CI 98.5% to 98.8%) in the two periods; non-current use persistence: 96.3% (95% CI 95.9% to 96.7%) and 96.8% (95% CI 96.4% to 97.3%) and cigarette-only use: 88.2% (95% CI 87.3% to 89.9%) and 88.3% (95% CI 87.4% to 89.1%)). ENDS uptake among adults who only use cigarettes remained low, with 3.8% (95% CI 3.4% to 4.2%) and 4.0% (95% CI 3.5% to 4.4%) transitioning.
One-year transition probabilities for adults in A. 2017–2019 (waves 4–5) and B. 2019–2021 (waves 5–6). CIs, including previous results from 2015 to 2017 (Waves 2–4), are shown in online supplemental figure S2. The transitions weighted by each use state’s prevalence are given in figure 3. ENDS, electronic nicotine delivery system.
The ENDS cessation probability increased slightly from 13.9% (95% CI 11.9% to 16.0%) to 18.1% (95% CI 16.2% to 19.9%). The transition patterns for those who dual used both products changed substantially over this period, driven by the increase in cigarette cessation. Specifically, from 2017–2019 to 2019–2021, the transition probability from dual to cigarette-only use did not statistically significantly change, but the transition probability from dual to ENDS-only use nearly doubled, increasing from 9.5% (95% CI 7.3% to 11.7%) to 20.0% (95% CI 17.4% to 22.6%). This increase was at the expense of the persistence of dual use, which decreased from 63.8% (95% CI 59.9% to 67.6%) to 50.1% (95% CI 46.2% to 54.1%) between the periods. There was no analogous increase in the cigarette-only to non-current use transition, which remained low at 6.8% (95% CI 6.2% to 7.5%) and 7.1% (95% CI 6.4% to 7.7%) in the two periods.
These transition probabilities weighted by prevalence demonstrate the larger population-level effect of these changes (figure 3). The decrease in cigarette initiation by 0.07% of the population translates to only about 180 000 fewer adults initiating cigarettes given an adult population 258.3 million in the 2020 US Census. The increase in ENDS-only prevalence from 1.5% in 2017 to 2.5% in 2019 (about 2.6 million additional adults) dwarfed the increase in the fraction of the population that were ENDS-only users transitioning to non-current use (an increase of about 620 000 adults), and the fraction of the population persisting in ENDS-only use increased by about 1.9 million. The population impact of the decrease in the persistence of dual use—with the fraction of the population that were dual users transitioning to either cigarette-only or ENDS-only use increasing by about 336 000 and 620 000 adults, respectively—was largely offset by the moderate increase in dual use prevalence from 1.5% to 1.9% (about 1.0 million adults), so that the fraction of the population persisting in dual use remained about the same.
One-year transition probabilities weighted by the prevalence of each use state at the beginning of that time period (given in online supplemental table S1), so that all transition probabilities add up to 100% and each row sums to the prevalence of that use type, for adults in A. 2017–2019 (waves 4–5) and B. 2019–2021 (waves 5–6). ENDS, electronic nicotine delivery system.
Considering the age-stratified transition probabilities (figure 4), most of the changes in transitions for those using ENDS only or dual using both products were qualitatively similar across the age groups. The persistence of e-cigarette-only use remained similar in 2017–2019 and 2019–2021, decreasing somewhat for ages 25–34 only. Persistence of dual use decreased in 2019–2021 for all three groups, accompanied by an increase in the transition probability of dual to ENDS-only.
One-year transition probabilities for adults ages 18–24 (A–B), ages 25–34 (C–D) and ages 35–90 (E–F). Each pair of plots compares periods 2017–2019 (waves 4–5) and 2019–2021 (waves 5–6). CIs, including previous results from 2015 to 2017 (waves 2–4), are shown in online supplemental figure S3). ENDS, electronic nicotine delivery system.
The fraction of young adults ages 18–24 transitioning from cigarette-only use to dual use increased from 2017 to 2019 (10.2% (95% CI 8.3% to 12.0)) to 2019–2021 (15.1% (95% CI 11.9% to 18.3%)). Young adults ages 18–24 were more likely to transition from cigarette-only use to dual use than those ages 25–34 (6.8% (95% CI 5.7% to 8.0%) in 2019–2021) or ages 35–90 (2.3% (95% CI 1.9 to 2.7%) in 2019–2021). Accordingly, a larger population fraction transitioned from cigarette-only use to dual use in 2019–2021 than from dual use to cigarette-only use among ages 18–24 and ages 25–34 while the reverse was true for ages 35–90 (online supplemental figure S4). Additionally, while the dual use to ENDS-only transition increased for all age groups, it was higher for young adults (27.6% (95% CI 21.9% to 33.3%) than for the other two ages groups (19.2% (95% CI 14.2% to 23.9%) and 16.8% (95% CI 13.0% to 20.7%), respectively).
Covariate HRs
HRs for sociodemographic and non-daily versus daily product use for 2017–2019 and 2019–2021 are given in online supplemental table S2. The largest HRs are by age group, particularly for transitions to ENDS use from never or non-current use, and the largest changes in HRs across this period are also by age (with decreasing HRs for never to ENDS use and dual to ENDS use transitions). HRs for gender, race and ethnicity, and non-daily versus daily use of each product largely remained similar or had modest changes over time.
Discussion
Changing marketplaces, policy environments, public health crises and evolving attitudes and perceptions of products have made it difficult to predict short-term patterns of use of cigarettes and ENDS, let alone their longer-term impacts on public health. In previous work, we had identified dramatic increases in the persistence of ENDS-only and dual use of cigarettes and ENDS around the time of the marketplace shift to pod mod products, including JUUL.7 Here, we found that many of the transition rates that significantly changed from 2015–2017 to 2017–2019 did not significantly change again in 2017–2019 to 2019–2021 (figure 1), including the non-current to cigarette-only use, cigarette-only to non-current use, ENDS-only to dual use and dual to cigarette-only use transitions. We did find three significant changes in transition rates from 2017–2019 to 2019–2021: a decline in never-established to cigarette-only use, an increase in ENDS-only to non-current use, and an increase in dual to ENDS-only use. From the perspective of reducing cigarette use, these changes appear to be beneficial, but the bigger picture underscores the overall public health challenge (figure 3), with high persistence of cigarette, ENDS and dual use.
The reduction in adult cigarette initiation did not make much of a difference at the population level. Indeed, in previous decades, about 90% of initiation took place before age 18.42 However, the proportion of young adults among those who have recently initiated smoking—possibly not as their first nicotine product—may be increasing, although the evidence is mixed.43 44 This trend merits further study. Additionally, the increase in the dual to ENDS-only use transition is likely to have a limited impact on the population-level health of those smoking cigarettes: dual use remained uncommon among adults smoking cigarettes (12% of those using cigarettes also used ENDS) and ENDS uptake among adults who smoke cigarettes remained low (<5% switching to dual or ENDS-only use in 1 year), particularly for those ages 25 and older.
While dual use remained persistent, potentially fuelling concerns that continued nicotine addiction through ENDS use may make cigarette cessation more difficult, there is evidence that an increased fraction of adults who dual used both products transitioned to ENDS-only use. That fraction remained small (overall, we estimated 20.0% of those who used both products transitioned to ENDS-only use in 1 year, less than the 26.3% of dual users transitioning to cigarette-only use), but it exceeded the fraction of participants who transitioned from cigarette-only to non-current use (7.1% overall).
However, we cannot attribute the increased cigarette cessation from dual use to participants’ use of ENDS; the pattern may be a result of demographic differences between dual and cigarette users or of ENDS use being an indicator of trying to quit. Additionally, our results continue to highlight the importance of understanding the different patterns among different stages of adulthood. Adoption of ENDS among those who currently used cigarettes was much greater among those ages 18–24 (15.1%) than among those ages 25–34 (6.8%) and was nearly negligible among older adults (2.3%) in 2019–2021.
There were multiple major public health events and policy changes during the 2019–2021 period that could have affected public perception and behaviours around cigarette and ENDS use. Following sharply increased rates of ENDS use among youth in 2018, the US Surgeon General declared youth ENDS use to be an ‘epidemic’45 Less than a year later, an outbreak of vaping-related lung injuries began in August 2019 and continued through the end of the year.8 Although the injuries are now thought to largely be associated with the use of vitamin E acetate as a diluent for tetrahydrocannabinol-containing liquids, early Centers for Disease Control and Prevention reports suggested nicotine vaping as the cause.8 10 46 The consequent emphasis on the dangers of vaping in US news resulted in more negative public perceptions of ENDS use and more of the public discourse focusing on their potential dangers to youth,47–51 in contrast to countries like the UK, where there has been more emphasis on ENDS’ potential benefits for adults.52 53 Additionally, in the USA, many states and localities enacted restrictions or bans on flavoured ENDS sales in 2020 and 2021.9 Moreover, legislation raising the minimum age of sale of tobacco products to 21 years, which has been associated with reduced smoking prevalence among 18–20 years at the local level,54 was federally enacted in late 2019.
Early 2020 marked the start of the COVID-19 pandemic in the USA, as SARS-CoV-2 spread rapidly across the country. Cigarette smoking was quickly found to be a risk factor for severe COVID-19 outcomes,55–57 and many studies linked greater perceived risk from COVID-19 to increased motivation to quit smoking.58–60 At the same time, elevated stress, particularly during lockdown or similarly restrictive periods, may have led to increased nicotine use as a coping mechanism.60 Ultimately, COVID-19 appears to have resulted in a combination of conflicting pressures for and against tobacco and nicotine product use.61
Given the period’s many public health events and policy changes, it is perhaps surprising that we do not see more changes in transition patterns for adults who use cigarettes only or ENDS only between 2017–2019 and 2019–2021. We did find a small increase in the ENDS-only to non-current use rate, but we found no change in the initiation rate of ENDS among never established users of cigarettes or ENDS nor in the ENDS cessation rate among those using both cigarettes and ENDS. Because the 2019 PATH data collection period encompassed December 2018 to November 2019, the changes we previously found between 2015–2017 and 2017–2019 may already reflect some of the changes discussed above. Also, because of the local nature of flavour restrictions, the impacts of those restrictions may not be observed at the national scale. Further, any changes in perceptions and behaviour may have been short-lived, for example, if lung-injury-related hesitance to use ENDS in 2019 or COVID-19-related motivation to quit smoking in early 2020 became irrelevant by the time PATH collected data in 2021. Or, the lack of change may represent the disruption of existing trends, for example, if more adults who smoke cigarettes would have taken up ENDS in the absence of the lung injury outbreak and flavour restrictions. Because we cannot know what would have happened without these events, we cannot be certain about the extent to which any individual public health event or policy change was associated with the observed transition changes.
The strengths of our analysis include the high-quality, nationally representative data collected by PATH and our multistate transition analysis allowing us to analyse the population-level changes in the transition rates that underly observed changes in transition probabilities over time. Additionally, our results continue to highlight the importance of understanding the different patterns of nicotine product use at different stages of adulthood, given the very different rates of adoption and transition between product use categories between younger and older adults. One strength of the study is the inclusion of a multivariable analysis accounting for gender, age, race and ethnicity, and daily versus non-daily use of each product; however, one limitation is that we did not include a mediation analysis to understand how much of the changes over time were associated with changes in sociodemographic or use HRs (eg, the increasing fraction of non-daily ENDS use (online supplemental table S1)). Another limitation of this study is that we did not incorporate data from the Adult Telephone Survey conducted in 2020; we chose to omit these data for this analysis because of the sample size and change in data collection methods. Additionally, our work is limited by our focus only on adults and by a reduced set of age categories compared with previous work.7 These changes were made so that the subgroups had sufficient sizes to support inferences about all the transitions of interest. We plan to analyse youth transitions in separate work with models using a set of product use categories and transitions tailored to the youth population. Finally, our work does not account for the use of other tobacco products, such as cigars, cigarillos or oral nicotine pouches.
As we discussed above, although many of the changes identified in this analysis appear to be positive from the point of view of reducing cigarette use, the overall public health implications are likely minimal. The US Food and Drug Administration (FDA) is currently determining which ENDS should be authorised and potentially marketed as reduced risk products. Their primary regulatory challenge is to make decisions that encourage cigarette cessation or reduce cigarette use among adults who use cigarettes while limiting ENDS use among youth and young adults who would never have initiated tobacco use in the absence of ENDS.27 While some tobacco-flavoured ENDSs have received FDA marketing authorisation, it is not clear if or when other flavoured ENDS could be authorised. In the meantime, a regulation banning non-tobacco characterising flavours, including menthol, in cigarettes and cigars, has been proposed.62 It will be important to continue to monitor how transitions in product use continue to change in response to regulatory changes, changes in the marketplace and future public health events.
Data availability statement
Data are available in a public, open access repository. Public Use Files from the Population Assessment of Tobacco and Health Study are available for download from an open access repository (https://doi.org/10.3886/ICPSR36498.v20). Conditions of use are available on the aforementioned website.
Ethics statements
Patient consent for publication
Ethics approval
This analysis was not regulated as human subjects research (University of Michigan Institutional Review Board HUM00162265).
Acknowledgments
This project was funded through National Cancer Institute (NCI) and Food and Drug Administration (FDA) grant U54CA229974, which supported all authors. Authors JJ, TRH, ASF, JT, DTL and RMe also acknowledge support from U01CA253858. JT acknowledges support from K01DA056424.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
X @jamie_tam
Contributors Conceptualisation and methodology: AFB and RMe; data curation: JJ and EJ-M; analysis and original draft preparation: AFB; review and editing: AFB, JJ, EJ-M, SRL, TRH, ASF, JT, RMi, DTL and RMe; funding acquisition: RMe and DTL; guarantor: AFB.
Funding This study was funded by Center for Tobacco Products (U54CA229974) and National Cancer Institute (K01DA056424), (U01CA253858), (U54CA229974).
Disclaimer The opinions expressed in this article are the authors’ own and do not reflect the views of the National Institutes of Health, the Department of Health and Human Services, or the US government.
Competing interests DTL has presented virtually at the tobacco-industry-sponsored Global Tobacco and Nicotine Forum conference but received no funds for his participation. All other authors declare that they have no conflicts of interest.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.