You are here

Article Text

Article menu

Download PDFPDF

Original research
Polytobacco use among young adult smokers: prospective association with cigarette consumption
  1. Angela Petersen1,2,
  2. Mark G Myers1,2,
  3. Lyric Tully1,
  4. Kristin Brikmanis3,
  5. Neal Doran1,2
  1. 1 Department of Psychiatry, University of California, San Diego, California, USA
  2. 2 Psychology Service, VA San Diego Healthcare System, San Diego, California, USA
  3. 3 Department of Psychology, University of Illinois, Chicago, Illinois, USA
  1. Correspondence to Dr Neal Doran, Department of Psychiatry, University of California, San Diego, CA 92161, USA; nmdoran{at}ucsd.edu

Abstract

Background. The risks of polytobacco use among young adults are unclear because we know relatively little about the consistency of multiproduct patterns over time and how these patterns impact cigarette smoking. The purpose of this study was to examine changes in multiple tobacco product use over time and associations with cigarette smoking quantity.

Methods. Participants (n=335; 55% male) were 18–24 years old non-daily cigarette smokers living in California. Polytobacco use patterns were assessed quarterly for 2 years.

Results. Transition analyses showed that while the number of products that had been used recently was volatile, the most common pattern was stability between timepoints. A longitudinal negative binomial regression model indicated that those who used more non-cigarette products also reported greater cigarette quantity. The strength of this relationship increased over time.

Conclusions. Findings suggest that individuals who use more tobacco products are at greater risk for increased cigarette smoking and maintaining a multiple product use pattern.

  • nicotine
  • non-cigarette tobacco products
  • priority/special populations

https://doi.org/10.1136/tobaccocontrol-2018-054625

Statistics from Altmetric.com

    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

    Polytobacco use (PTU) is increasingly common among US adults,1 2 particularly those aged 18–24 years.3–5 Data from the Population Assessment of Tobacco and Health study indicate PTU prevalence of 21.7% among participants aged 18–24 years, compared with 3.6% of those aged 12–17 years and 15.8% of those aged 24–34 years.3 Another study of adults aged 18–34 years tracked tobacco use over three waves in 2011–2012. They found 57% reported lifetime use at baseline, compared with 67% 1 year later, and that those who had tried multiple products were the modal group at wave 3, accounting for 45% of participants.5 Among young adults, it is increasingly clear that PTU is more common than single product use,6–8 and that this change is a function of escalating use of non-cigarette products, particularly e-cigarettes and hookah tobacco.5–8

    While other tobacco products, such as e-cigarettes and hookah, may not lead to nicotine dependence as readily as regular cigarette consumption, there is substantial evidence that repeated exposure induces dependence.9–12 In addition, the increasing prevalence of PTU may signal an impending population-level problem. PTU provides exposure to multiple sources of nicotine, potentially increasing risk of dependence,13 and thus of continued or progressive use of multiple products and increased likelihood of cigarette smoking. Our current understanding is largely limited to prevalence of multiple product use in a specified time frame, and correlates of current or ever use of multiple products. This is a significant limitation given recent evidence that for young adults, transitions among use states are highly dynamic.14 Further, studies have used varying PTU definitions, making comparisons difficult.2 15 16 It is unclear whether PTU should be defined as use of multiple products within a particular time frame or whether the specific products matter. As a result, the risk of PTU for development of chronic tobacco use is unknown. However, initial studies suggest other tobacco products are associated with cigarette uptake17–20 and progression.13 21 Preliminary evidence suggests users of 3+ products may be more dependent and less motivated to quit cigarettes,22 but whether this predicts cigarette use is unknown. These studies suggest PTU is likely to confer risk for chronic use of multiple products, for cigarette uptake and progression, and for tobacco-related morbidity and mortality.

    The most notable gap in studies of young adults is the lack of longitudinal evaluations of PTU. To date, understanding is largely limited to prevalence of use within a specified time frame, and correlates of current or ever PTU. As a result, the consistency of PTU among young adults is not well understood. Additionally, the extent to which PTU confers risk for development of nicotine dependence and progression of cigarette smoking is unknown. The present study addresses these gaps by examining stability of PTU over time and testing the hypotheses that PTU predicts heavier cigarette use over time, and of a dose–response relationship between PTU and cigarette quantity, such that cigarettes would increase in concert with the total number of products used. Use of cigarettes was chosen as the primary outcome because of their status as the most dangerous and prevalent combustible tobacco product.23 24

    Materials and methods

    Participants

    Participants (n=335) were young adults participating in a longitudinal study of non-daily cigarette smoking. Eligibility criteria included being 18–24 years old, smoking cigarettes at least monthly for the past 6 months or longer, and never smoking cigarettes daily for 1 month or longer. Because assessments were completed online, consistent internet access was required. Due to regional and state differences in cigarette prevalence25 and legal restrictions,26 the parent study was limited to California residents.

    The sample was 45% female, with a racial/ethnic composition of 39% non-Hispanic white, 25% Asian American, 24% Hispanic/Latino and 11% from other or multiple backgrounds. Participants were on average 20.4 (SD 1.8) years of age, and 58% were full-time university or community college students.

    Procedure

    Participants were recruited via paid Facebook posts. Interested individuals completed a brief online screening to determine eligibility. Eligible respondents completed an online baseline assessment. Participants completed online assessments quarterly over the course of 2 years. They received $25 gift cards for completing assessments at baseline (BL) and 1 (Y1) and 2 (Y2) years later. At the other six timepoints (3 months, 6 months, 9 months, 15 months, 18 months, 21 months), participants completed daily assessments for 9 consecutive days, receiving $4 gift cards for each day completed, plus an additional $4 if they completed all 9 days. Data were collected in 2015–2017.

    Measures

    Demographic characteristics assessed included sex, race/ethnicity and student status. Race was collapsed into non-Hispanic white (n=132), Asian American (n=85), Hispanic or Latino (n=81), and other or multiple backgrounds (n=37). Student status was dichotomised as full-time students (58%) or part-time and non-students (42%).

    Tobacco use: Use of tobacco was assessed at each timepoint for cigarettes, hookah, e-cigarettes, cigars/cigarillos and smokeless tobacco. Tobacco use was assessed for the past 14 days at BL, Y1 and Y2, and for the past 9 days for 3 months, 6 months, 9 months, 15 months, 18 months and 21 months. For BL, Y1 and Y2, participants indicated whether each product was used. Positive responses triggered detailed assessment via the Timeline Followback,27–29 which included number of cigarettes smoked on each of the past 14 days, and whether each of the other products was used on each day. For 3 months, 6 months, 9 months, 15 months, 18 months and 21 months, participants reported the number of cigarettes smoked in the past 24 hours, and whether each other product was used during the same period, on each of 9 consecutive days. These measures have been shown to be strongly associated with biological measures of cigarette use, including measures of nicotine, cotinine and carbon monoxide.30

    Polytobacco use: Prior to defining PTU we examined the distribution of baseline use patterns reported for the prior 14 days: 46% reported use of a single tobacco product and 52% multiple products. Other than PTU consisting only of cigarettes and e-cigarettes (11%), the remainder reflected multiple combinations that precluded a simple classification. Thus, consistent with our goal of estimating cumulative nicotine intake, we operationalised PTU in two ways. First, to examine transitions in PTU patterns we created an ordinal variable (PTUord) representing the number of products, including cigarettes, reported in the past 9 or 14 days at each assessment interval (0, 1, 2 or 3+ products). Second, to model the cumulative impact of PTU on cigarette outcomes, we created a count variable (PTUcount) reflecting the time-varying, cumulative sum of the number of days on which each product was used prior to each assessment point. For example, at 3 months, the value of PTUcount was the sum of the number of days of use of each product at baseline. At 6 months, the value was the sum of the number of days of use of each product at baseline and 3 months. Two versions of PTUcount were generated, one including the sum of days used for all products, and a second excluding cigarettes. To account for differences in the number of days assessed we calculated a days variable that reflected the cumulative number of days on which use had been assessed to that point in the study.

    Analytical plan

    Data were tabulated to describe quarterly transitions in number of products used from BL to Y2 (see figure 2) using PTUord. The hypothesis that PTU predicts greater cigarette use was tested by fitting separate longitudinal models for each time-varying predictor (ie, PTUcount with and without cigarettes) with final models selected to examine overall strength of predictors. To test whether changes in number of products predicted cigarette use, we fit a third longitudinal model using PTUord as an ordinal predictor. For all models, the outcome variable was total cigarettes at each of eight assessments (3 months, 6 months, 9 months, Y1, 15 months, 18 months, 21 months and Y2 post-BL). To account for potential pre-existing differences in cigarette consumption, models accounted for total cigarettes at baseline, sex, racial/ethnic background and student status. The days variable was a covariate. Because the outcome variables were overdispersed, negative binomial models were a better fit compared with alternative choices (eg, multi-level or Poisson models).31 No assumptions or imputations were made for missing data. Models initially included terms for time, time2, sex, race/ethnicity, and student status and their interactions with PTU. Non-significant interactions were removed. All analyses were conducted using Stata IC V.15.1 (StataCorp LP), with α=0.05.

    Results

    Preliminary analyses

    Eighteen participants (5.1%) completed no assessments after baseline and were excluded from the analyses. Dropouts did not differ from responders with regard to age, race/ethnicity, sex, student status, education or PTU. Overall, the 335 participants included in the analyses below completed 94% of possible assessments. Preliminary analyses indicated that male participants and those who were not full-time students reported higher levels of PTU (p<0.05).

    Transitions in use over time

    Initial analyses examined transitions across PTU categories. Results indicated substantial instability across product categories, but as seen in figure 1 stability was the modal state from one timepoint to the next. The proportion of the sample that reported stability between timepoints gradually increased from 32% between BL and 3 months to 55% between 21 months and Y2. Additionally, figure 2 shows the proportion of the sample using 0, 1, 2 or 3+ products at each timepoint. This plot indicates that recent use of one product was most common over time, and that non-use increased from 2% at BL to approximately 30% at Y2. Across timepoints, 92.8% of single-product users smoked cigarettes, compared with 95.2% of users of 2 products and 99.5% of users of 3+ products.

    Figure 1
    Figure 1

    Proportion of participants whose number of tobacco products increased, decreased or remained the same between timepoints. m, months; Y1, year 1, Y2, year 2.

    Figure 2
    Figure 2

    Proportion of sample using different numbers of products over time. BL, baseline; m, months; Y1, year 1; Y2, year 2.

    PTU predicting cigarettes smoked

    The initial model examined PTU including cigarettes (table 1). Some demographic categories predicted smoking fewer cigarettes, including being a full-time student (z=−2.82, p<0.001), being female (z=3.34, p<0.001) and reporting Asian American ethnicity (z=2.55, p=0.011). PTU x demographics interaction terms were not significant and were not retained. In terms of the impact of PTU on cigarette consumption, we found significant linear (z=3.46, p=0.001) and quadratic (z=−4.37, p<0.001) interactions between PTU and time. To better understand this interaction, we plotted pairwise correlations between PTU and total cigarettes at each timepoint, as seen in figure 3. The pattern of correlations indicates that greater PTU frequency was associated with more cigarettes smoked, and that the strength of this association increased over time during the 2 years of the study. More specifically, correlations ranged from r=0.02 at 3 months to r>0.43 from Y1 onwards.

    View this table:
    Table 1

    Longitudinal model of association between polytobacco use (PTU) and total cigarettes

    Figure 3
    Figure 3

    Correlations between polytobacco use (PTU) and cigarettes smoked over time. m, months.

    Next we examined the association between PTUcount excluding cigarettes and cigarettes smoked over time. This model (table 2) produced similar results, including significant linear (z=2.11, p=0.035) and quadratic (z=−2.35, p=0.019) PTU x time interactions. We again plotted correlations between non-cigarette PTU and cigarettes smoked at each timepoint. Correlations indicated a pattern in which the association between frequency of use of non-cigarette products and quantity of cigarettes consumed grew stronger over time. Removing cigarettes from the PTU calculation weakened these associations which ranged from near-zero at 3 months and 6 months to r=0.16 at 9 months and Y1, and r>0.22 for each of the last four assessments. From 9 months onwards, all correlations were significant at p<0.05.

    View this table:
    Table 2

    Longitudinal model of association between polytobacco use (PTU) (excluding cigarettes) and total cigarettes

    Finally, we used PTUord as an ordinal predictor to examine whether number of products used at each timepoint predicted cigarettes smoked over time. Because previous research suggested use of 3+ products may be particularly predictive,22 we coded this as the reference category. Differences between number of products used did not vary over time. Participants who used 3+ products reported significantly more cigarettes over time compared with those who used two products (z=−4.57, p<0.001), one product (z=−8.40, p<0.001) and no products (z=21.74, p<0.001). Over each 9–14-day period, 3+ product users reported 0.27 more cigarettes than two-product users and 0.51 more than single-product users. Refitting the model with two products as the reference indicated that this group smoked significantly more than single product users (z=−5.26, p<0.001). After accounting for covariates, the average difference between these groups was 0.24 cigarettes per assessment period.

    Discussion

    Use of multiple tobacco products is increasingly common among youth and young adults, but little is known regarding the potential consequences of this behaviour. We examined PTU over time in a sample of young adult non-daily cigarette smokers. Findings revealed considerable instability over time, both within and across categories reflecting the number of tobacco products used. In addition, we found that greater PTU predicted greater cigarette use, and that this relationship grew stronger over time. The pattern held whether or not cigarettes were included in the PTU predictor. We also found incremental effects, whereby users of 3+ products smoked significantly more cigarettes than users of fewer products, and that two-product users smoked more cigarettes than those who used a single product. These findings are consistent with initial studies and suggest PTU confers a significant risk for progression of cigarette use and therefore may increase long-term risk for negative health effects. Findings also suggest that risk for chronic cigarette smoking increases in tandem with the number of products used.

    These results mirror prior work14 32 identifying frequent transitions among different categories of tobacco use. Changes occurred both across the number of products used and in the proportion of the sample within a given category over time. Despite this instability, the most common transition pattern was to remain in the same category of number of products used from one timepoint to the next, and the likelihood of stability increased over time. The proportion of those reporting recent use of two or more products declined over time, and the number reporting no recent use increased. These reductions in use could reflect that recent cigarette smoking was a study entry criterion, and potentially captured a period of heightened use for all participants. The fluctuations in tobacco use patterns observed here also indicate that these behaviours have not yet become routinised, suggesting these natural transitions may represent opportunities for disruptive interventions to prevent progression to long-term tobacco use.

    Findings were consistent with our hypothesis that greater PTU would predict greater cigarette smoking. Importantly, the finding of a positive association between PTU frequency and cigarette quantity that grew stronger over time was consistent regardless of whether cigarette frequency was included in the PTU predictor. Thus, while likelihood of PTU across the sample was unstable and declined over time, those who did use multiple products also reported increasing cigarette consumption across 2 years. This finding adds to a growing body of evidence that PTU presents a risk for progression to nicotine dependence and chronic cigarette use. This is a concern given that PTU now appears to be the most common pattern of tobacco use in younger populations,6–8 and suggests that without intervention this pattern could lead to an increase in the prevalence of adult use of tobacco products generally and cigarettes particularly. Beyond concerns related to tobacco-related health consequences, PTU is also associated with problem use of alcohol, marijuana and other drugs6 33 and may also increase risk for chronic use of these substances. This suggests that young adult PTU may reflect a higher risk of negative consequences from tobacco as well as other drugs of abuse.

    The finding of a dose–response relationship between number of products used and cigarette quantity is consistent with prior work.22 This suggests using more products is associated with greater likelihood of chronic cigarette smoking and attendant negative consequences. It is notable that users of two products smoked significantly more than participants who used a single product, suggesting use of non-cigarette products was not primarily intended to reduce cigarette harms or aid cessation. This is consistent with studies suggesting young adult PTU is associated with recreation and experimentation.34 35 We also found participants who were female, full-time students and identified as Asian American tended to smoke fewer cigarettes, and that women and full-time students were less likely to use multiple products. However, none of these moderated associations between PTU and cigarette smoking, indicating that PTU predicted cigarette use across demographic categories.

    The present findings must be interpreted in light of a number of limitations. The parent study recruited non-daily cigarette smokers in California and may not generalise to other populations of young adults in the USA or elsewhere. However, it is clear that PTU is prevalent among young adults across the USA. Recent data also suggest that smoking initiation is now more common among young adults than adolescents.36 Together these findings underscore the value and likely generalisability of studying non-daily smokers in the USA in the early stages of smoking progression. While regional variations in product use exist, the present findings, by indicating that use of more products predicts greater cigarette use, suggest the risks associated with PTU here may generalise to other young adults in the USA. Whether these patterns hold outside the USA is unknown and bears examination. In addition, assessment of non-cigarette products was limited to number of days used. It may be that more detailed assessment of quantity or nicotine content would yield different results. However, it is noteworthy that this relatively gross assessment of tobacco product use during brief assessment periods significantly predicted increased cigarette use.

    In sum, these findings suggest use of multiple nicotine-containing products is associated with increasing cigarette consumption over time. This is alarming as PTU is more common than use of any individual product among younger US populations,6–8 and in conjunction with evidence that multiple product users are more receptive to tobacco company marketing15 37 raises the possibility that it could cause increased tobacco-related illness and death. The possibility that high PTU prevalence could increase cigarette smoking over time highlights the need to address this issue and counter factors driving PTU use, particularly because public health education efforts designed to reduce tobacco prevalence have been effective.38 For example, antismoking campaigns on college campuses have been successful in reducing cigarette consumption39; broadening such campaigns to include content about other products and about PTU, and to reach young adults who are not students, could reduce the long-term impact of PTU on nicotine dependence and its health consequences. Additionally, new regulatory policies may be an effective way to mitigate PTU growth among youth and young adults.40 For example, current restrictions on tobacco advertising are of limited effectiveness in preventing young people’s exposure to such advertisements.41 Additional restrictions may reduce exposure and thus long-term risk for PTU and cigarette smoking.

    What this paper adds

    • The prevalence of polytobacco use has increased substantially in the past decade, particularly among younger populations.

    • Cross-sectional data suggest polytobacco use may be associated with greater risk of nicotine dependence.

    • Previous polytobacco studies are generally limited to cross-sectional surveys identifying prevalence and correlates. Little is known about whether polytobacco use predicts tobacco outcomes over time.

    • Our findings indicate that after accounting for baseline use, greater use of multiple tobacco products over time may be associated with increasing consumption of cigarettes and thus with greater risk for negative health outcomes.

    References

    1. ↵
      2. Lee YO ,
      3. Hebert CJ ,
      4. Nonnemaker JM , et al
      . Multiple tobacco product use among adults in the United States: cigarettes, cigars, electronic cigarettes, hookah, smokeless tobacco, and snus. Prev Med 2014;62:14–19.doi:10.1016/j.ypmed.2014.01.014
      OpenUrlCrossRefPubMed
    2. ↵
      2. Sung HY ,
      3. Wang Y ,
      4. Yao T , et al
      . Polytobacco use of cigarettes, cigars, chewing tobacco, and snuff among US Adults. Nicotine Tob Res 2016;18:817–26.doi:10.1093/ntr/ntv147
      OpenUrlCrossRefPubMed
    3. ↵
      2. Osibogun O ,
      3. Taleb ZB ,
      4. Bahelah R , et al
      . Correlates of poly-tobacco use among youth and young adults: Findings from the Population Assessment of Tobacco and Health study, 2013-2014. Drug Alcohol Depend 2018;187:160–4.doi:10.1016/j.drugalcdep.2018.02.024
      OpenUrl
    4. ↵
      2. Osibogun O ,
      3. Jaber R ,
      4. Bahelah R , et al
      . Poly-tobacco use among young adults in the united states: findings from the population assessment on tobacco and health study, 2013–2014. Ann Epidemiol 2017;27:535.doi:10.1016/j.annepidem.2017.07.053
      OpenUrl
    5. ↵
      2. Richardson A ,
      3. Williams V ,
      4. Rath J , et al
      . The next generation of users: prevalence and longitudinal patterns of tobacco use among US young adults. Am J Public Health 2014;104:1429–36.doi:10.2105/AJPH.2013.301802
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      2. Collins LK ,
      3. Villanti AC ,
      4. Pearson JL , et al
      . Frequency of youth e-cigarette, tobacco, and poly-use in the United States, 2015: update to villanti et al., "frequency of youth e-cigarette and tobacco use patterns in the united states: measurement precision is critical to inform public health". Nicotine Tob Res 2017;19:1253–4.doi:10.1093/ntr/ntx073
      OpenUrl
    7. ↵
      2. Fix BV ,
      3. O’Connor RJ ,
      4. Vogl L , et al
      . Patterns and correlates of polytobacco use in the United States over a decade: NSDUH 2002-2011. Addict Behav 2014;39:768–81.doi:10.1016/j.addbeh.2013.12.015
      OpenUrlCrossRefPubMed
    8. ↵
      2. Rath JM ,
      3. Villanti AC ,
      4. Abrams DB , et al
      . Patterns of tobacco use and dual use in US young adults: the missing link between youth prevention and adult cessation. J Environ Public Health 2012;2012:1–9.doi:10.1155/2012/679134
      OpenUrl
    9. ↵
      2. Case KR ,
      3. Mantey DS ,
      4. Creamer MR , et al
      . E-cigarette- specific symptoms of nicotine dependence among Texas adolescents. Addict Behav 2018;84:57–61.doi:10.1016/j.addbeh.2018.03.032
      OpenUrlPubMed
    10. ↵
      2. Etter JF ,
      3. Bullen C
      . Electronic cigarette: users profile, utilization, satisfaction and perceived efficacy. Addiction 2011;106:2017–28.doi:10.1111/j.1360-0443.2011.03505.x
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      2. Maziak W ,
      3. Ward KD ,
      4. Eissenberg T
      . Factors related to frequency of narghile (waterpipe) use: the first insights on tobacco dependence in narghile users. Drug Alcohol Depend 2004;76:101–6.doi:10.1016/j.drugalcdep.2004.04.007
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      2. St Helen G ,
      3. Benowitz NL ,
      4. Dains KM , et al
      . Nicotine and carcinogen exposure after water pipe smoking in hookah bars. Cancer Epidemiol Biomarkers Prev 2014;23:1055–66.doi:10.1158/1055-9965.EPI-13-0939
      OpenUrlAbstract/FREE Full Text
    13. ↵
      2. Doran N ,
      3. Godfrey KM ,
      4. Myers MG
      . Hookah use predicts cigarette smoking progression among college smokers. Nicotine Tob Res 2015;17:1347–53.doi:10.1093/ntr/ntu343
      OpenUrlCrossRefPubMed
    14. ↵
      2. Hair E ,
      3. Bennett M ,
      4. Williams V , et al
      . Progression to established patterns of cigarette smoking among young adults. Drug Alcohol Depend 2017;177:77–83.doi:10.1016/j.drugalcdep.2017.03.040
      OpenUrl
    15. ↵
      2. El-Toukhy S ,
      3. Choi K
      . A risk-continuum categorization of product use among us youth tobacco users. Nicotine Tob Res 2016;18:1596–605.doi:10.1093/ntr/ntw008
      OpenUrlCrossRefPubMed
    16. ↵
      2. Soneji S ,
      3. Sargent J ,
      4. Tanski S
      . Multiple tobacco product use among US adolescents and young adults. Tob Control 2016;25:174–80.doi:10.1136/tobaccocontrol-2014-051638
      OpenUrlAbstract/FREE Full Text
    17. ↵
      2. Barrington-Trimis JL ,
      3. Urman R ,
      4. Berhane K , et al
      . E-cigarettes and future cigarette use. Pediatrics 2016;138:e20160379–8.doi:10.1542/peds.2016-0379
      OpenUrlAbstract/FREE Full Text
    18. ↵
      2. Leventhal AM ,
      3. Strong DR ,
      4. Kirkpatrick MG , et al
      . Association of electronic cigarette use with initiation of combustible tobacco product smoking in early adolescence. JAMA 2015;314:700–7.doi:10.1001/jama.2015.8950
      OpenUrlCrossRefPubMed
    19. ↵
      2. Primack BA ,
      3. Soneji S ,
      4. Stoolmiller M , et al
      . Progression to traditional cigarette smoking after electronic cigarette use among us adolescents and young adults. JAMA Pediatr 2015;169:1018–23.doi:10.1001/jamapediatrics.2015.1742
      OpenUrl
    20. ↵
      2. Spindle TR ,
      3. Hiler MM ,
      4. Cooke ME , et al
      . Electronic cigarette use and uptake of cigarette smoking: A longitudinal examination of U.S. college students. Addict Behav 2017;67:66–72.doi:10.1016/j.addbeh.2016.12.009
      OpenUrlCrossRefPubMed
    21. ↵
      2. Doran N ,
      3. Brikmanis K ,
      4. Petersen A , et al
      . Does e-cigarette use predict cigarette escalation? A longitudinal study of young adult non-daily smokers. Prev Med 2017;100:279–84.doi:10.1016/j.ypmed.2017.03.023
      OpenUrlPubMed
    22. ↵
      2. Ali M ,
      3. Gray TR ,
      4. Martinez DJ , et al
      . Risk profiles of youth single, dual, and poly tobacco users. Nicotine Tob Res 2016;18:1614–21.doi:10.1093/ntr/ntw028
      OpenUrlCrossRefPubMed
    23. ↵
      2. Zeller M ,
      3. Hatsukami D
      ; The Strategic Dialogue on Tobacco Harm Reduction: a vision and blueprint for action in the US. Tob Control 2009;18:324–32.doi:10.1136/tc.2008.027318
      OpenUrlAbstract/FREE Full Text
    24. ↵
      2. Ng M ,
      3. Freeman MK ,
      4. Fleming TD , et al
      . Smoking prevalence and cigarette consumption in 187 countries, 1980-2012. JAMA 2014;311:183–92.doi:10.1001/jama.2013.284692
      OpenUrlCrossRefPubMedWeb of Science
    25. ↵
      2. Kasza KA ,
      3. Ambrose BK ,
      4. Conway KP , et al
      . Tobacco-product use by adults and youths in the United States in 2013 and 2014. N Engl J Med 2017;376:342–53.doi:10.1056/NEJMsa1607538
      OpenUrlPubMed
    26. ↵
      2. Morain SR ,
      3. Winickoff JP ,
      4. Mello MM
      . Have tobacco 21 laws come of age? N Engl J Med 2016;374:1601–4.doi:10.1056/NEJMp1603294
      OpenUrlCrossRefPubMed
    27. ↵
      2. Brown RA ,
      3. Burgess ES ,
      4. Sales SD , et al
      . Reliability and validity of a smoking timeline follow-back interview. Psychology of Addictive Behaviors 1998;12:101–12.doi:10.1037/0893-164X.12.2.101
      OpenUrlCrossRefWeb of Science
    28. ↵
      2. Sobell LC ,
      3. Sobell MB
      . Timeline follow-back: A technique for assessing self-reported alcohol consumption. In: Litten RZ , Allen JP , eds. Measuring alcohol consumption: Psychosocial and biochemical methods. Totowa, NJ: Humana Press, 1992:41–72.
    29. ↵
      2. Sobell MB ,
      3. Sobell LC ,
      4. Klajner F , et al
      . The reliability of a timeline method for assessing normal drinker college students' recent drinking history: utility for alcohol research. Addict Behav 1986;11:149–61.doi:10.1016/0306-4603(86)90040-7
      OpenUrlCrossRefPubMedWeb of Science
    30. ↵
      2. Wray JM ,
      3. Gass JC ,
      4. Miller EI , et al
      . A comparative evaluation of self-report and biological measures of cigarette use in nondaily smokers. Psychol Assess 2016;28:1043–50.doi:10.1037/pas0000227
      OpenUrl
    31. ↵
      2. Ismail N ,
      3. Jemain A
      . Handling overdispersion with negative binomial and generalized Poisson regression models. Casualty Actuarial Society 2007:103–58.
    32. ↵
      2. Schweizer CA ,
      3. Roesch SC ,
      4. Khoddam R , et al
      . Examining the stability of young-adult alcohol and tobacco co-use: A latent transition analysis. Addict Res Theory 2014;22:325–35.doi:10.3109/16066359.2013.856884
      OpenUrl
    33. ↵
      2. Haardörfer R ,
      3. Berg CJ ,
      4. Lewis M , et al
      . Polytobacco, marijuana, and alcohol use patterns in college students: A latent class analysis. Addict Behav 2016;59:58–64.doi:10.1016/j.addbeh.2016.03.034
      OpenUrlPubMed
    34. ↵
      2. Wong EC ,
      3. Haardörfer R ,
      4. Windle M , et al
      . Distinct motives for use among polytobacco versus cigarette only users and among single tobacco product users. Nicotine Tob Res 2017;20:117–23.doi:10.1093/ntr/ntw284
      OpenUrl
    35. ↵
      2. Yu E ,
      3. Saddleson M ,
      4. Murphy J , et al
      . Patterns of polytobacco use among college students. Health Behav Policy Rev 2017;4:173–81.doi:10.14485/HBPR.4.2.8
      OpenUrl
    36. ↵
      2. Perry CL ,
      3. Pérez A ,
      4. Bluestein M , et al
      . Youth or young adults: Which group is at highest risk for tobacco use onset? J Adolesc Health 2018;63:413–20.doi:10.1016/j.jadohealth.2018.04.011
      OpenUrlPubMed
    37. ↵
      2. Lee YO ,
      3. Pepper JK ,
      4. MacMonegle AJ , et al
      . Examining youth dual and polytobacco use with e-cigarettes. Int J Environ Res Public Health 2018;15:699.doi:10.3390/ijerph15040699
      OpenUrl
    38. ↵
      2. Farrelly MC ,
      3. Davis KC ,
      4. Haviland ML , et al
      . Evidence of a dose-response relationship between "truth" antismoking ads and youth smoking prevalence. Am J Public Health 2005;95:425–31.doi:10.2105/AJPH.2004.049692
      OpenUrlCrossRefPubMedWeb of Science
    39. ↵
      2. Lechner WV ,
      3. Meier E ,
      4. Miller MB , et al
      . Changes in smoking prevalence, attitudes, and beliefs over 4 years following a campus-wide anti-tobacco intervention. J Am Coll Health 2012;60:505–11.doi:10.1080/07448481.2012.681816
      OpenUrlCrossRefPubMed
    40. ↵
      2. Fowler CD ,
      3. Gipson CD ,
      4. Kleykamp BA , et al
      . Basic science and public policy: Informed regulation for nicotine and tobacco products. Nicotine Tob Res 2018;20:789–99.doi:10.1093/ntr/ntx175
      OpenUrl
    41. ↵
      2. Pierce JP ,
      3. Sargent JD ,
      4. White MM , et al
      . Receptivity to tobacco advertising and susceptibility to tobacco products. Pediatrics 2017;139:e20163353.doi:10.1542/peds.2016-3353

    Footnotes

    • Contributors AP, MGM and ND conceptualised the study together. AP wrote the first draft, and MGM wrote the final draft. KB and LT collected data and conducted analyses. ND obtained funding and oversaw data analyses. All authors provided feedback and edited the manuscript.

    • Funding Funding for this study was provided by NIDA grant R01 DA037217.

    • Competing interests None declared.

    • Patient consent Not required.

    • Ethics approval UC San Diego IRB.

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

    • Data sharing statement All data from this study are available for sharing. Requests should be directed to ND at nmdoran@ucsd.edu