Abstract
Background
The purpose of this study was to determine the developmental course of marijuana use among adolescents based on their history of cigarette and e-cigarette use among a national U.S. sample of adolescents who were followed over a four year time-period.
Methods
The data for this study used four waves of the Population Assessment of Tobacco and Health (PATH) Study provided by a panel of 12 to 17-year-olds at Wave 1 (n=11,059) who completed each of the four annual waves of the adolescent/adult survey. We examined recent use (i.e., past 30-day) of e-cigarettes, cigarettes, and marijuana use at each of the four waves.
Results
Respondents who had a history of non-concurrent dual use (AOR = 1.67, 95% CI = 1.24, 2.24) and a history of concurrent dual use (AOR = 1.67, 95% CI = 1.40, 1.99) had greater odds of past 30-day marijuana use when compared to respondents who had a history of past 30-day e-cigarette use only. Interaction effect models found that e-cigarette only users were at lower risk for past 30-day marijuana use at Wave 1, however, the risk of past 30-day marijuana use increased at a faster rate across the four waves for e-cigarette only users when compared to their peers who used cigarettes or a combination of cigarettes and e-cigarettes.
Conclusion
While concurrent and non-concurrent dual use was strongly associated with marijuana use over the study period, marijuana use increased at a faster rate across the four-year span of the study among e-cigarette only users.
Keywords: e-cigarettes, cigarettes, marijuana
1. Introduction
Prior longitudinal studies have found that both cigarette and e-cigarette use is associated with marijuana use among adolescents and young adults (Audrain-McGovern et al., 2018; Dai et al., 2018; Evans-Polce et al., 2020; Ramo et al., 2012; Unger et al., 2016). Studies also indicate that the risk of marijuana use is highest among adolescents who engage in dual use (i.e., individuals who use both cigarettes and e-cigarettes) (Dai and Hao, 2017; Kristjansson et al., 2015; McCabe et al., 2019; McCabe et al., 2017). For instance, more than 80% of U.S. high school students with a lifetime history of dual use had also used marijuana (McCabe et al., 2017). These findings derived from cross-sectional and short-term data (e.g., one year) indicate the need for long-term prospective studies to examine the longitudinal association between marijuana use and different combinations of cigarette and e-cigarette use among adolescents (Evans-Polce et al., 2020). Given recent increases in both marijuana and e-cigarette/e-product use among U.S. adolescents (Cullen et al., 2018; Johnston et al., 2020; Miech et al., 2019), it is necessary to understand how the history of cigarette and e-cigarette use during adolescence (i.e., no use, e-cigarette use only, cigarette use only, and concurrent and non-concurrent dual use) is associated with the developmental course of marijuana use over a longer time period. Accordingly, the purpose of this study was to determine the different trajectories of marijuana use among adolescents based on their history of cigarette and e-cigarette use among a national U.S. sample of adolescents who were followed over a four-year time-period.
2. Methods
2.1. Sample
This study used data from the Population Assessment of Tobacco and Health (PATH) Study, a nationally representative panel of youth (ages 12 to 17 at Wave 1) who were assessed at four separate time points, Wave1: September/2013-December/2014; Wave 2: October/2014-October/2015; and Wave 3: October/2015-October/2016; and Wave 4: December/2016-January/2018 (United States Department of Health and Human Services, 2020). The PATH Study used a four-stage stratified area probability sample design. Audio computer-assisted selfinterviewing (ACASI) was conducted and on-screen displays and flashcards were used to aid adolescent respondents. The retention rate within the adolescent sample was 79.5% by Wave 4. The retained youth sample (including those who aged into the adult sample) included 11,059 respondents.
2.2. History of past 30-day cigarette and e-cigarette use
Past 30-day cigarette and e-cigarette use were measured with two variables across each wave of the survey (Waves 1 through 4): “In the past 30 days, on how many days did you smoke cigarettes?”, and “In the past 30 days, on how many days did you use an e-cigarette?” Response options ranged from 0 to 30 days. These measures were then recoded as a binary variable ‘past 30-day use’ versus ‘no past 30-day use’. In order to best isolate the patterns of past 30-day use across the four waves of the survey, we combined these two measures at each wave (i.e., Wave 1 through 4) to construct a mutually exclusive variable with five unique categories: (1) no cigarette or e-cigarette use during the past 30-days during the four waves of the study (2) only e-cigarette use (only indicated past 30-day e-cigarette use during at least one wave of the survey), (3) history of non-concurrent cigarette use and e-cigarette use (indicated past 30-day use of either e-cigarette or cigarette use during at least one wave of the survey, but not both during the same wave [i.e., non-concurrent dual use]), (4) only cigarette use (only indicated past 30-day cigarette use during at least one wave of the survey), and (5) dual use (concurrent past 30-day cigarette and e-cigarette use during at least one wave). For the analyses, this independent variable capturing history of cigarette and e-cigarette use across the four waves was treated as a time invariant variable in order to clearly assess the different trajectories of marijuana use during this time-period.
2.3. Frequency of Past 30-day cigarette and e-cigarette use
Two additional independent variables measured at each wave assessed past 30-day cigarette frequency (i.e., 0-30 days) and e-cigarette frequency (i.e., 0-30 days). For the analysis both of these measures were treated as continuous measures that were time-varying.
2.4. Past 30-day marijuana use
The major outcome variable, marijuana use, was assessed with one item at each wave. The question asked if respondents “used marijuana in the past 30 days”. Response options included “Yes” or “No”. For the analyses the variable was coded as a binary outcome (i.e., Yes = 1, No = 0) and was treated as a time-varying outcome.
2.5. Control variables
In order to account for potentially confounding factors, all of the multivariable analyses controlled for sex, race, Hispanic ethnicity, respondent’s age at Wave 1, household income, and U.S. region. Each of these sociodemographic were treated as time-invariant characteristics. Additionally, a single time-invariant composite measure was constructed to assess lifetime use of other tobacco products.
2.6. Analyses
Binary logistic regression models were fitted using the generalized estimating equations (GEE) methodology with an autoregressive correlation structure to assess the association between history of past 30-day use of cigarettes/e-cigarettes and the time-varying outcome for past 30-day marijuana use (Hanley et al., 2003; Zeger et al., 1988). Models with and without covariates are provided along with the unadjusted odds ratio (OR), adjusted odds ratio (AOR) and 95% confidence intervals. GEE models were estimated with the full sample using respondents with no past 30-day cigarette or e-cigarette use during the four waves of the study as the reference group. Moreover, models using a smaller subset of respondents that indicated any past 30-day cigarette or e-cigarette use during the study period were estimated using respondents that only engaged in past 30-day e-cigarette use during the study period as the reference group. Finally, in order to assess potential differences in trajectories of marijuana use between each of the four cigarette/e-cigarette groups, interaction effects were estimated based on history of past 30-day cigarette/e-cigarette use and wave (i.e., time) of the survey. These models treated wave as a continuous time-varying variable; interaction effects were assessed by taking the product of history of past 30-day cigarette/e-cigarette use (time-invariant) and wave of survey (time-varying). All conducted analyses used weights and designated variables to account for the complex sampling design. Stata 15.0 was used for all analyses. Sample sizes may vary given that listwise deletion was used when estimating these models in Stata.
3. Results
Table 1 provides the descriptive statistics for the full panel sample (n = 11059) and the sample who indicated past 30-day use of either cigarette or e-cigarette use during at least one wave (n = 2902). Among the longitudinal sample, 27% of the sample indicated past 30-day use of either cigarettes or e-cigarettes during the study period. Table 1 also shows that the prevalence of e-cigarette use only, cigarette use only, dual use, and marijuana use increased over the four waves of the study. With respect to history of past 30-day use of cigarettes and e-cigarettes over the four waves among past 30-day cigarette and/or e-cigarette users (n = 2902), the largest group was concurrent dual users (34.5%), followed by only e-cigarette users (34.4%), only cigarette users (25.3%), and non-concurrent dual users (5.8%).
Table 1.
Sample characteristics with the adolescent panel sample
| Total Sample | Past 30-day Nicotine/Tobacco Users | |||
|---|---|---|---|---|
| Total n | n = 11059 | Total n | n = 2902 | |
| Sexa | ||||
| Males | 5610 | 51.2% | 1594 | 55.5% |
| Females | 5385 | 48.8% | 1305 | 44.5% |
| Raceb | ||||
| White | 7077 | 69.6% | 2015 | 75.2% |
| Black | 1706 | 16.1% | 318 | 11.6% |
| Other | 1623 | 14.3% | 444 | 13.3% |
| Hispanic ethnicityb | ||||
| Non-Hispanic | 7671 | 77.4% | 2122 | 79.5% |
| Hispanic | 3156 | 22.6% | 742 | 20.5% |
| Age (Wave 1)c | ||||
| 12 to 14 years of age | 5341 | 47.5% | 777 | 26.2% |
| 15 to 17 years of age | 5678 | 52.5% | 2125 | 73.8% |
| Household incomed | ||||
| $24,999 or lower | 1700 | 13.4% | 437 | 13.3% |
| $25,000 to $49,999 | 2128 | 17.4% | 459 | 14.2% |
| $50,000 to $99,000 | 2451 | 22.2% | 541 | 18.8% |
| $100,000 or higher | 2349 | 24.1% | 482 | 18.4% |
| Missing | 2392 | 22.9% | 983 | 35.4% |
| U.S. regione | ||||
| Northeast | 1587 | 16.3% | 444 | 17.1% |
| Midwest | 2440 | 22.2% | 725 | 24.5% |
| South | 4127 | 37.0% | 1027 | 35.4% |
| West | 2866 | 24.6% | 706 | 23.0% |
| Past 30 day marijuana use | ||||
| Marijuana use (Wave 1) | 561 | 5.2% | 443 | 15.5% |
| Marijuana use (Wave 2) | 962 | 9.1% | 725 | 26.3% |
| Marijuana use (Wave 3) | 1209 | 11.7% | 845 | 31.2% |
| Marijuana use (Wave 4) | 1626 | 16.0% | 1036 | 39.8% |
| Past 30 day cigarette/e-cigarette use | ||||
| No Use (Wave 1) | 10307 | 93.8% | 2227 | 77.1% |
| No Use (Wave 2) | 9590 | 89.7% | 1679 | 60.7% |
| No Use (Wave 3) | 9001 | 84.7% | 1171 | 42.5% |
| No Use (Wave 4) | 8397 | 80.3% | 692 | 25.2% |
| E-cigarette use only (Wave 1) | 170 | 1.6% | 170 | 6.1% |
| E-cigarette use only (Wave 2) | 308 | 3.1% | 308 | 11.6% |
| E-cigarette use only (Wave 3) | 576 | 5.6% | 576 | 20.9% |
| E-cigarette use only (Wave 4) | 689 | 7.0% | 689 | 26.5% |
| Cigarette use only (Wave 1) | 344 | 3.2% | 344 | 11.8% |
| Cigarette use only (Wave 2) | 489 | 4.5% | 489 | 17.2% |
| Cigarette use only (Wave 3) | 536 | 5.1% | 536 | 19.3% |
| Cigarette use only (Wave 4) | 780 | 7.4% | 780 | 28.1% |
| Dual use (Wave 1) | 151 | 1.3% | 151 | 5.0% |
| Dual use (Wave 2) | 282 | 2.7% | 282 | 10.4% |
| Dual use (Wave 3) | 480 | 4.6% | 480 | 17.3% |
| Dual use (Wave 4) | 510 | 5.3% | 510 | 20.1% |
Notes: n = unweighted sample size; Percentages and means incorporate baseline survey weights for the longitudinal sample; SE = standard error; Sample sizes may vary due to missing data.
Sex of respondent was a derived variable (i.e., PATH constructed the variable) from the interview and included either ‘Male’ or ‘Female’.
Race/Ethnicity of respondent was a derived variable from the interview and included either ‘White alone’, ‘Black alone’, and ‘Other’. Hispanic was derived from the interview and included either ‘Hispanic’ or ‘Not Hispanic’.
Age of respondent at Wave 1was a derived variable from the interview and included either ’12 to 14 years old’ and ‘15 to 17 years old’. It should be noted that the public use files only provide dichotomous age ranges for the adolescent sample (i.e., 12 to 14, 15 to 17) and seven for the adult sample (18 to 24, etc… adolescent who age into the adult sample will only fall into his category. Below we provide the age ranges of respondents between Wave 1 and Wave 4 based on the possible three age brackets from the adolescent public use sample (unweighted estimates are provided): Wave 1, 12-14 (52.7%), 15-17 (47.3%), 18-24 (0%); Wave 2, 12-14 (35.7%), 15-17 (49.2%), 18-24 (15.1%); Wave 3, 12-14 (18.2%, 15-17 (51.4%), 18-24 (30.3%); Wave 4, 12-14 (1.9%), 15-17 (50.6%), 18-24 (47.5%).
Household income was a derived variable from the interview and include five categories: ‘less than $10,000’, ‘$10,00 to $24,999’, ‘$25,000 to $49,999’, ‘$50,000 to $99,999’, and ‘$100,000 or more’. The maximum income indicated in either Wave 2 through Wave 4 was used for the analysis. A derived variable for household income is not included at Wave 1. Missing data on this variable is due to the youth sample in Wave 1 moving to the adult sample in Wave 2.
Table 2 provides the results from the GEE analyses assessing history of past 30-day cigarette/e-cigarette use and past 30-day marijuana use. Assessing the full sample (see Analysis 1) shows that any history of past 30-day use of cigarettes or e-cigarettes over the study period was associated with greater odds of past 30-day marijuana use when compared to peers who did engage in any past 30-day use of cigarette/e-cigarette use. For instance, respondents who indicated any history of concurrent dual use had roughly four times greater odds of indicating past 30-day marijuana use when compared to respondents who had no history of past 30-day cigarette/e-cigarette use (AOR = 4.56, 95% CI = 3.82, 5.45) when adjusting for both frequency of use, wave of survey, and other sociodemographic factors. The odds of past 30-day marijuana use were greater at later waves when compared to Wave 1; odds of past 30-day marijuana use significantly increased across each wave (refer to 95% CI’s for the analyses using the full models).
Table 2.
Assessing how the history of e-cigarette and cigarette use is associated with the developmental course of marijuana use
| Past 30-day marijuana use (time-varying) | Past 30-day marijuana use (time-varying) | |
|---|---|---|
| Analysis 1 (Full Sample) | Full Sample Models (n=9,263) | |
| Time Invariant Variables | Unadjusted OR (95% CI) | Adjusted aOR (95% CI) |
| History of past 30 day use of cigarettes and e-cigarettes across all 4 waves | ||
| Did not use cigarettes or e-cigarettes | Reference | Reference |
| E-cigarette use only | 5.46***(4.78,6.24) | 2.96***(2.50,3.51) |
| Non-concurrent dual use | 11.3***(8.91,14.4) | 4.66***(3.45,6.29) |
| Cigarette use only | 7.93***(6.88,9.14) | 3.24***(2.69,3.89) |
| Any history of concurrent dual use | 13.8***(12.2,15.6) | 4.56***(3.82,5.45) |
| Time Varying Variables | ||
| Number of days used (past 30 days) | ||
| Number of days used e-cigarettes during the past 30 days | 1.05***(1.04,1.06) | 1.01** (1.00,1.02) |
| Number of days used cigarettes during the past 30 days | 1.05***(1.04,1.06) | 1.02***(1.01,1.03) |
| Time | ||
| Wave 1 | Reference | Reference |
| Wave 2 | 1.84***(1.69,2.02) | 2.07***(1.85,2.32) |
| Wave 3 | 2.44***(2.21,2.68) | 2.90***(2.56,3.27) |
| Wave 4 | 3.53***(3.20,3.88) | 4.90***(4.34,5.52) |
| Past 30-day marijuana use (time-varying) | Past 30-day marijuana use (time-varying) | |
| Analysis 2 (Only respondents with a history of 30-day cigarette/e-cigarette use) | Subgroup Sample Models (n=2,351) | |
| Time Invariant Variables | Unadjusted OR (95% CI) | Adjusted aOR (95% CI) |
| History of past 30 day use of cigarettes and e-cigarettes across all 4 waves | ||
| E-cigarette use only | Reference | Reference |
| Non-concurrent dual use | 2.07***(1.60,2.67) | 1.67***(1.24,2.24) |
| Cigarette use only | 1.44***(1.23,1.70) | 1.17 (.969,1.41) |
| Any history of concurrent dual use | 2.52***(2.18,2.91) | 1.67***(1.40,1.99) |
| Time Varying Variables | ||
| Number of days used (past 30 days) | ||
| Number of days used e-cigarettes during the past 30 days | 1.02***(1.01,1.03) | 1.01** (1.00,1.02) |
| Number of days used cigarettes during the past 30 days | 1.03***(1.02,1.04) | 1.02***(1.00,1.02) |
| Time | ||
| Wave 1 | Reference | Reference |
| Wave 2 | 1.92***(1.72,2.15) | 2.02***(1.78,2.30) |
| Wave 3 | 2.45***(2.17,2.77) | 2.57***(2.23,2.97) |
| Wave 4 | 3.63***(3.19,4.11) | 4.05***(3.51,4.68) |
| Past 30-day marijuana use (time-varying) | ||
| Analysis 3 (Interaction effect model) | Interaction Effect Model (n=2,351) | |
| Time Invariant Variables | Adjusted aOR (95% CI) | |
| History of past 30 day use of cigarettes and e-cigarettes | ||
| E-cigarette use only | Reference | |
| Non-concurrent dual use | 3.06***(2.04,4.61) | |
| Cigarette use only | 1.66***(1.22,2.26) | |
| Any history of concurrent dual use | 2.78***(2.11,3.66) | |
| Time Varying Variables | ||
| Time | ||
| Continuous measure for Wave of PATH (Wave 1 = 0 to Wave 4 = 3) | 1.87***(1.70,2.05) | |
| Interaction Effects | ||
| Non-concurrent dual use X Time | 0.716***(.591,.866) | |
| Cigarette use only X Time | 0.833***(.732,.948) | |
| Any history of concurrent dual use X Time | 0.759***(.676,.852) | |
Notes: n = unweighted sample size; Analyses incorporate baseline survey weights for the longitudinal sample; SE = standard error; Sample sizes may vary due to missing data; All analyses control for sex, race, Hispanic ethnicity, respondent’s age at Wave 1, household income, U.S. region, and lifetime use of other tobacco products.
Table 2 also provides the results assessing only respondents who had a history of past 30-day cigarette/e-cigarette use (see Analysis 2). Accordingly, respondents who had a history of non-concurrent dual use (AOR = 1.67, 95% CI = 1.24, 2.24) and a history of any concurrent dual use (AOR = 1.67, 95% CI = 1.40, 1.99) had greater odds of past 30-day marijuana use when compared to respondents who had a history of past 30-day e-cigarette use only. No differences in the odds of past 30-day marijuana use were found between respondents with a history of past 30-day cigarette use only and respondents with a history of past 30-day e-cigarette use only.
The interaction effect model provided in Table 2 found statistically significant differences in the linear increase between e-cigarette only users and the three other groups that had a history of cigarette use only or a history of both cigarette and e-cigarette use (see Analysis 3). The main effects show that at baseline, respondents who indicated a history of cigarette use only or a combination of cigarette and e-cigarette use had higher odds of past 30-day marijuana use when compared to respondents who had a history of e-cigarette use only. However, the interaction effects show that when compared to e-cigarette only users, the positive linear increase in past 30-day marijuana use is weaker for non-concurrent dual users, cigarette use only users, and concurrent dual users. In other words, while e-cigarette only users were at lower risk for past 30-day marijuana use at Wave 1 when compared to their peers who used cigarettes or a combination of cigarettes and e-cigarettes, the risk of past 30-day marijuana use increased at a faster rate across the four waves for e-cigarette only users. Additional analyses found no differences in the linear increase between the other three groups of respondents with a history of cigarette use or combination of cigarette and e-cigarette use (supplemental figure A provides the observed results to show this interaction graphically).
4. Discussion
This is one of the first studies to assess the developmental course of marijuana use based on the history of cigarette and e-cigarette use during adolescence over a four-year time period. The findings from the present study extends prior research that has found an association between marijuana use and different combinations of cigarette and e-cigarette use over shorter time periods (Dai and Hao, 2017; Kristjansson et al., 2015; McCabe et al., 2019; McCabe et al., 2017). In particular, adolescents who reported concurrent dual use or non-concurrent dual use of cigarettes and e-cigarettes were at the greatest risk of marijuana use during the four waves of the study. Moreover, while adolescents who had a history of only e-cigarette use were at lower risk of marijuana use when compared to their peers who used cigarettes or some combination of cigarette and e-cigarettes, the odds of marijuana use during the study period among e-cigarette only users were roughly three times higher when compared to adolescents who did not engage in any past 30-day cigarette or e-cigarette use.
Additionally, the results of this study also provide new evidence that the developmental course of marijuana use is significantly different among adolescents who have a history of e-cigarette use only when compared to their peers who use cigarettes or some combination of cigarette and e-cigarettes. In particular, e-cigarette only users had a lower risk of marijuana use at the first wave when compared to adolescents who used cigarettes or some combination of cigarettes and e-cigarettes. However, marijuana use increased at a faster rate across the four-year span of the study among e-cigarette only users that could be missed in shorter term studies.
The results of this study suggest that the use of e-cigarettes alone, regardless of frequency of use, was associated with a steeper increase in the risk of using marijuana when compared to their peers who only used cigarettes or a combination of cigarettes and e-cigarettes over the study period. This is particularly concerning given the increase in e-cigarette use among adolescents (Cullen et al., 2018; Johnston et al., 2020; Miech et al., 2019), and the growing proportion of adolescent marijuana users who report vaping as their route of administration (Knapp et al., 2019; Trivers et al., 2018). Based on the results of this study, greater effort needs to be focused on how e-cigarette/e-product use (e.g., flavorings) may be an initial pathway to marijuana use. While this study expands upon our knowledge of how e-cigarettes are associated with the developmental course of marijuana use among adolescents, several limitations should be noted. First, this study relied on self-reported data and may be subject to various types of respondent bias. Moreover, there are a number of confounding factors not included in the analysis that may account for the association between cigarette/e-cigarette use and marijuana use (e.g., availability, peer use/norms). Finally, this study could only assess a general measure of past 30-day marijuana use and could not determine whether adolescents used some type of e-product to smoke this substance. Despite these issues, this study provides needed epidemiological information to understand how e-cigarettes are associated with the increased risk of marijuana use as adolescents’ age into young adulthood. While dual users are at an increased risk for marijuana use, adolescent’s with only a history of e-cigarette use appear to be a vulnerable group whose risk of marijuana use sharply increases during this phase of development. These findings reinforce the need to target adolescents who use e-cigarettes on either an experimental or frequent basis given that this type of substance use behavior is a potential marker for later risk behaviors like marijuana use.
Supplementary Material
Highlights.
Dual users are at an increased risk for marijuana use.
Marijuana use increased at a faster rate among e-cigarette only users.
E-cigarette use is a significant marker for later risk behaviors like marijuana use.
Acknowledgments
Funding source:
Supported by grants R01 DA044157 (Dr. Boyd) and R01 CA203809 (Dr. McCabe) from the National Institutes of Health (NIH), National Institute on Drug Abuse (NIDA), and National Cancer Institute (NCI).
Footnotes
Notation of Research Letter status:
This paper contains original material, not submitted, in press, or published elsewhere.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Disclosure
All of the authors have nothing to disclose.
Conflict of interest:
The authors of this Research Letter have no conflicts of interest to report.
References
- Audrain-McGovern J, Stone MD, Barrington-Trimis J, Unger JB, Leventhal AM, 2018. Adolescent e-cigarette, hookah, and conventional cigarette use and subsequent marijuana use. Pediatrics 142 (3),e20173616. doi: 10.1542/peds.2017-3616 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cullen KA, Ambrose BK, Gentzke AS, Apelberg BJ, Jamal A, King BA, 2018 Notes from the field: use of electronic cigarettes and any tobacco product among middle and high school students - United States, 2011-2018. MMWR Morb Mortal Wkly Rep. 67(45), 1276–1277. Published 2018. November 16. doi: 10.15585/mmwr.mm6745a5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Dai H, Catley D, Richter KP, Goggin K, Ellerbeck EF, 2018. Electronic cigarettes and future marijuana use: a longitudinal study. Pediatrics 141(5), e20173787. doi: 10.1542/peds.2017-3787 [DOI] [PubMed] [Google Scholar]
- Dai H, Hao J, 2017. Electronic cigarette and marijuana use among youth in the United States. Addict Behav. 66, 48–54. doi: 10.1016/j.addbeh.2016.11.005 [DOI] [PubMed] [Google Scholar]
- Evans-Polce RJ, Veliz PT, Boyd CJ, McCabe SE, 2020. E-cigarette and cigarette use among U.S. adolescents: longitudinal associations with marijuana use and perceptions. Am J Prev Med 58(6), 854–857. doi: 10.1016/j.amepre.2020.01.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hanley JA, Negassa A, Edwardes MD, Forrester JE, 2003. Statistical analysis of correlated data using generalized estimating equations: an orientation. Am J Epidemiol 157(4), 364–375. doi: 10.1093/aje/kwf215 [DOI] [PubMed] [Google Scholar]
- Johnston LD, Miech RA, O’Malley PM, Bachman JG, Schulenberg JE, Patrick ME, 2020. Monitoring the Future national survey results on drug use 1975-2019: Overview, key findings on adolescent drug use. Ann Arbor: Institute for Social Research, University of Michigan. [Google Scholar]
- Knapp AA, Lee DC, Borodovsky JT, Auty SG, Gabrielli J, Budney AJ, 2019. Emerging trends in cannabis administration among adolescent cannabis users. J Adolesc Health 64(4), 487–493. doi: 10.1016/j.jadohealth.2018.07.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kristjansson AL, Mann MJ, Sigfusdottir ID, 2015. Licit and illicit substance use by adolescent e-cigarette users compared with conventional cigarette smokers, dual users, and nonusers. J Adolesc Health 57(5), 562–564. doi: 10.1016/j.jadohealth.2015.07.014 [DOI] [PubMed] [Google Scholar]
- McCabe SE, Veliz PT, McCabe VV, Boyd CJ, 2019. Initiation sequence of e-cigarette and cigarette smoking among US adolescents: a national study. Am J Addict 28(4), 285–294. doi: 10.1111/ajad.1288 [DOI] [PMC free article] [PubMed] [Google Scholar]
- McCabe SE, West BT, Veliz PT, Boyd CJ, 2017. E-cigarette use, cigarette smoking, dual use, and problem behaviors among U.S. adolescents: results from a national survey. J Adolesc Health 61(2), 155–162. doi: 10.1016/j.jadohealth.2017.02.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Miech R, Johnston LD, O’Malley PM, Bachman JG, Patrick ME, 2019. Trends in adolescent vaping, 2017-2019. N Engl J Med 381(15), 1490–1491. doi: 10.1056/NEJMc1910739 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ramo DE, Liu H, Prochaska JJ, 2012. Tobacco and marijuana use among adolescents and young adults: a systematic review of their co-use. Clin Psychol Rev 32(2), 105–121. doi: 10.1016/j.cpr.2011.12.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Trivers KF, Phillips E, Gentzke AS, Tynan MA, Neff LJ, 2018. Prevalence of cannabis use in electronic cigarettes among US youth. JAMA Pediatr 172(11), 1097–1099. doi: 10.1001/jamapediatrics.2018.1920 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Unger JB, Soto DW, Leventhal A, 2016. E-cigarette use and subsequent cigarette and marijuana use among Hispanic young adults. Drug Alcohol Depend 163, 261–264. doi: 10.1016/j.drugalcdep.2016.04.027 [DOI] [PMC free article] [PubMed] [Google Scholar]
- United States Department of Health and Human Services. National Institutes of Health. National Institute on Drug Abuse, and United States Department of Health and Human Services. Food and Drug Administration. Center for Tobacco Products. Population Assessment of Tobacco and Health (PATH) Study [United States] Public-Use Files. 2020. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor]. [Google Scholar]
- Zeger SL, Liang KY, Albert PS, 1988. Models for longitudinal data: a generalized estimating equation approach. Biometrics 44(4), 1049–1060. [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.