Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2018 Jan 1.
Published in final edited form as: Tob Regul Sci. 2017 Jan 1;3(1):115–127. doi: 10.18001/TRS.3.1.12

Youth E-cigarette, Blunt, and Other Tobacco Use Profiles: Does SES Matter?

Patricia Simon 1, Deepa R Camenga 2, Grace Kong 3, Christian M Connell 4, Meghan E Morean 5, Dana A Cavallo 6, Suchitra Krishnan-Sarin 7
PMCID: PMC5654631  NIHMSID: NIHMS874258  PMID: 29082301

Abstract

Objective

We examined the associations between socioeconomic status (SES) and adolescent polytobacco use profiles (including e-cigarettes and blunts).

Methods

Adolescents (N = 1932) completed surveys conducted in 2014 in 3 Connecticut high schools. Within a Latent Class Analysis (LCA) framework, logistic regressions examined associations between SES and polytobacco use profiles comprising never, ever, and current e-cigarette, blunt, cigarette, cigar, cigarillo, hookah, and smokeless tobacco use.

Results

We identified 5 classes: (1) current polytobacco users; (2) ever polytobacco users; (3) current e-cigarette, blunt, and cigarette users; (4) ever e-cigarette and blunt users; and (5) never users. Low SES, relative to high SES, was associated with greater likelihood of being: (1) an ever polytobacco user; (2) a current e-cigarette, blunt, and cigarette user; and (3) an ever e-cigarette and blunt user, relative to a never user.

Conclusions

Low SES is associated with membership in distinct polytobacco use latent classes. Regulatory initiatives that focus exclusively on cigarette use may miss the opportunity to influence adolescent use of other products, which may be especially relevant to low-income individuals. Future research should examine whether targeting a broader range of products reduces SES-related tobacco use disparities.

Keywords: polytobacco use, adolescents, socioeconomic status, latent class analysis

Tobacco use is a leading contributor to preventable morbidity and mortality in the United States (US). Although effective tobacco control policies have contributed to steady declines in cigarette use among youth,1 emerging trends for use of alternative tobacco products are concerning. National data indicate increases in current use of alternative tobacco products from 2011 to 2015; e-cigarette use increased from 1.5% to 16%, and hookah use increased from 4.1% to 7.2%.2 Moreover, a survey among adolescents in Connecticut showed high current use rates for blunts (ie, cigars filled with marijuana; 11.6%)3 and national trends show increased rates for blunt use in combined samples of adolescents and adults.4 Additionally, national surveys show high current use rates for cigar use (inclusive of cigarillo, 8.6%) and smokeless tobacco use (6.0%).2

Low socioeconomic status (SES) is associated with cigarette smoking initiation and progression to regular smoking in adolescents.5,6 Specifically, low SES, as documented by participation in free/reduced lunch, is associated with increased likelihood of adolescent cigarette smoking.7 Unlike the established associations between low SES and cigarette smoking in adults and adolescents, the association between low SES and alternative tobacco product use has not been well-studied. The emerging evidence has been contradictory; one study found that lower SES is associated with increased e-cigarette use in adults,8 but other studies have reported no association between SES and adolescent e-cigarette use.9-11

Low SES also appears to be associated with increased use of cigars,12 cigarillos,12 hookah,13 and smokeless tobacco14 in adults. However, there is less consistent literature examining the associations between SES and use of alternative tobacco products among adolescents. One study found no association between parent education and adolescent cigar/cigarillo use.15 Another study found that adolescents' weekly spending income was positively associated with both cigarillo and cigar use.16 Concerning smokeless tobacco use, whereas parent education was not related to smokeless tobacco use in a nationally representative sample of US high school students,15 it was positively associated with smokeless tobacco use in Mexican-origin adolescents in Texas.17 With reference to blunt use, parental education and plans to attend college were not associated with blunt use in adolescents.18 Blunt use typically is examined as part of marijuana use, which is associated with higher SES among young adults potentially due to greater access on college campuses or stress/anxiety about achieving academic goals.19-22 Although there has been much debate about whether blunt use is best characterized as cigar use or marijuana use, we thought it was important to include use of blunts in our examination because youth create blunts from hollowed out cigars or tobacco leaf wrappers. Finally, pertaining to hookah, in nationally representative samples in the US, research has shown that greater parent education, greater adolescent earned income, and greater adolescent weekly spending money are associated with ever, past-year, and current hookah use.23,24

Although examining the association between SES and individual tobacco products is useful, important relationships may be obscured because such analyses may not reflect adolescents' polytobacco use. Indeed, studies have shown that the use of one tobacco product is associated with concurrent use of other tobacco products.25-30 Nationally, 12.7% of adolescents report using 2 or more tobacco products.31 The emergence of new alternative tobacco products (eg, e-cigarettes) and the resurgence of blunt use necessitate new investigations of poly-tobacco use in adolescents. Prior research that has relied on latent class analysis32,33 and latent transition analysis34 to identify tobacco use classes that include e-cigarettes and traditional tobacco products indicates that there are at least 3 tobacco use classes: nonusers, users that have a high probability of polytobacco use, and users that primarily use e-cigarettes.32-34 Although they contributed to knowledge of how e-cigarette use relates to other tobacco product use, these studies did not assess how factors like SES may contribute to class membership.

To our knowledge, no studies have examined the association between SES and polytobacco use latent classes (inclusive of e-cigarettes and blunts) in adolescents. Understanding which new polytobacco use profiles have emerged and how SES is associated with these new polytobacco profiles use could help local and national policy makers develop more nuanced approaches to prevent adolescent tobacco use, especially with regard to reducing SES-based tobacco use disparities.

Thus, the objective of this study is to describe the association between SES and polytobacco use profiles that include e-cigarettes, blunts, cigarettes, cigarillos, cigars, hookah, and smokeless tobacco, through the use of latent class analysis (LCA). Importantly, we assessed SES using the Family Affluence Scale (FAS), a composite measure of adolescents' family resources (ie, family car and computer ownership; frequency of family vacationing; individual bedrooms for children). FAS items are easier for adolescents to answer than common SES measures like family financial status and parental education because adolescents may not be aware of these factors.32 As such, adolescent responses to the FAS are more accurate and have fewer missing responses than more common measures.32 Consistent with prior research, we developed latent classes based on never, ever, and current use of tobacco products.33-35 We anticipated identifying latent classes characterized by no use and polytobacco product use, respectively. Due to the novelty of this research, we did not make hypotheses about additional tobacco use profiles. We also hypothesized that low SES, relative to high SES, would be associated with membership in polytobacco user profiles relative to the non-user profile.

Methods

Participants

Participants were 1932 adolescents from 3 high schools in southeastern Connecticut. In spring 2014, participants completed a school-wide, paper-and-pencil survey examining attitudes toward and use patterns of e-cigarettes and other tobacco products.

Procedures

After obtaining IRB and school administrator approvals to conduct the study, we obtained passive parental permission for participants younger than 18 years old and participant consent for participants 18 years old or older. Prior to administering the survey in homeroom, we informed participants that their participation was voluntary and that their data would remain confidential.36,37

Measures

Demographics

The current survey was part of a larger, longitudinal study. Participants reported their age and sex in the current survey and in a prior survey conducted in fall 2013. Participants reported their race/ethnicity only on the prior survey. As such, data on race were available only for students who completed both surveys and whose surveys could be linked.38 Using this procedure we recovered race/ethnicity for 72.4% of participants. Race was coded as white (88.0%) versus non-white (12.0%) due to relatively small sample sizes for racial/ethnic minorities (black = 1.4%; Hispanic/Latino = 3.4%; Asian = 2.9%; biracial = 1.9%; Other = 0.7%). Sex was coded as female = 1 and male = 0.

Socioeconomic status (SES)

Participants completed the Family Affluence Scale (FAS),39 a 4-item measure developed by the World Health Organization (WHO). The FAS has been shown to be a reliable and valid measure of SES among adolescents in Europe and North America.32,39,40 Participants indicated whether their family owns a car, van, or truck (no = 0, yes = 1) and whether they have their own bedroom (no = 0, yes = 1). In addition, participants reported the number of laptops/computers their family owns (none = 0, 1= 1, 2 = 2, more than 2 = 3). Participants also indicated how often their family had vacationed in the past 12 months (not at all = 0, once = 1, twice = 2, more than twice = 3). Based on established thresholds,39,41 responses were used to create composite scores, which resulted in a trichotomous SES measure, whereby a 3-point ordinal scale was created, such that scores 0–4 were classified as low, 5–6 as medium, and 7–8 as high. Dummy-coded variables were created to contrast Low SES with Medium SES; Low SES with High SES; and Medium SES with High SES.

Tobacco product use

We assessed ever use of e-cigarette, blunt (cigar filled with marijuana), cigarette, cigarillo (short, narrow cigar), cigar, hookah, and smokeless tobacco with the question stem: “Have you ever tried [product]?” Participant responses were coded as either yes or no. Current use of these products was assessed with the question: “How many days out of the past 30 days did you use [product]?”. A report of one or more days of use of each product was coded as current use of that product. Separately for each product, we created a trichotomized variable, wherein never use = 0, ever use = 1 (a report of ever use but not current use), and current use = 2. All tobacco products, with the exception of cigarettes, were described with a photograph and short definition of the product. We also created 4 variables (ever tried one or more tobacco products; ever tried 2 or more tobacco products; current use of one or more products; current use of 2 or more products) to help describe the extent of “any product use” and “multiple product use” in preliminary screening.

Data Analysis

We completed preliminary data screening (eg, examination of descriptive data, missing data analysis) in SPSS 21.42 Data regarding race were missing for 27.6% of the sample. All other variables were missing data for fewer than 5% of cases. We conducted logistic regression to identify differences between participants who were missing data for race and participants who had data available for race. Participants who were missing data on race were more likely to be male (OR 0.54; 95% CI 0.43, 0.68), have tried cigarettes (OR 2.02; 95% CI 1.44, 2.83) and cigarillos (OR 2.02; 95% CI 1.28, 3.17) in their life time, but were less likely to have tried cigars (OR 0.64; 95% CI 0.411, 0.99). To address missing data, we conducted multiple imputation in Mplus 7.2.43 We requested 100 multiple imputation datasets to reduce bias in estimates.44 We compared results using multiple imputation to results not using multiple imputation (with and without a category for race = unknown). We found all 3 approaches yielded similar results. Therefore, we only present results from analyses that used multiple imputation to handle missing data. We used chi-square tests followed by Bonferroni-corrected z-tests (for dichotomous and ordinal variables) and analysis of variance (ANOVA; for continuous and categorical/ordinal variables) to determine bivariate associations between SES and demographic variables and tobacco use.

In Mplus 7.2,43 we used LCA to identify polytobacco product use profiles.45 We identified the optimal number of classes by starting with a one-class solution and comparing the effect of adding additional classes on multiple indicators of model fit; in the LCA, optimal fit was determined by identifying models with a lower Bayesian Information Criteria (BIC) statistic, a lower sample size–adjusted BIC statistic, a lower Akaike Information Criterion (AIC) statistic, and a higher entropy score. Within the same Mplus model, we used multinomial regression to examine whether SES was associated with the identified latent classes, after controlling for race, age, sex and school. Modeling the latent classes in tandem with the multinomial regression accounts for the effects of covariates on the probability of class membership.

Results

Sample Characteristics and Bivariate Associations

Sample characteristics and bivariate associations are presented in Table 1. In our sample, 18.1% of participants were from a low SES background, 48.8% of participants were from a medium SES background, and 33.1% were from a high SES background. The distribution of responses to the Family Affluence Scale (ie, SES) items is presented in Table 2. Most adolescents from low SES backgrounds reported having one to 2 computers at home, whereas adolescents from medium and high SES backgrounds tended to report having more than 2 computers at home. Rates of family car ownership were high in our sample (98%), though rate of car ownership was lowest among adolescents from low SES backgrounds. Rates for having one's own room were above 90% for adolescents from medium and high SES backgrounds, but at 69.5% for adolescents from low SES backgrounds.

Table 1. Demographics and Prevalence of Tobacco Product Use by Socioeconomic Status (SES) (N = 1932).

Total Low SES (18.1%, N = 349) Medium SES (48.8%, N = 943) High SES (33.1%, N = 640) χ2 F
Demographics
Age (Mean, SD) 16.0 (1.3) 16.2 (1.3)a 16.1 (1.3)a 15.9 (1.3)b 5.29**
Sex (%)
 Female 50.5 40.1a 50.7b 55.8b 22.21***
Race (%)
 White 88.4 81.4a 90.2b 89.5b 20.73***
Tobacco Use
E-cigarettes (%)
 Never 66.0 58.7a 66.1b 70.0b 14.36**
 Ever 18.7 21.2a 19.4a 16.3a
 Past month 15.3 20.1a 14.5b 13.8b
Blunts (%)
 Never 76.8 70.8a 78.4b 77.7b 9.88*
 Ever 12.1 15.2a 11.8a 10.8a
 Past month 11.2 14.0a 9.9a 11.6a
Cigarettes (%)
 Never 77.2 64.2a 78.8b 81.9b 43.83***
 Ever 11.9 18.1a 11.6b 8.9b
 Past month 11.0 17.8a 9.7b 9.2b
Cigarillos (%)
 Never 86.6 83.1a 86.4a,b 88.8b 6.63
 Ever 8.2 10.6a 8.5a 6.6a
 Past month 5.2 6.3a 5.1a 4.7a
Cigars (%)
 Never 83.5 81.7a 82.7a 85.6a 12.11*
 Ever 10.3 13.2a 11.3a 7.2b
 Past month 6.2 5.2a 5.9a 7.2a
Hookahs (%)
 Never 80.8 75.6a 81.9b 82.0a,b 7.85
 Ever 13.0 16.0a 12.6a 11.9a
 Past month 6.2 8.3a 5.5a 6.1a
Smokeless Tobacco (%)
 Never 90.0 84.5a 90.6b 92.2b 17.86**
 Ever 5.8 8.6a 6.0a,b 3.9b
 Past month 4.2 6.9a 3.4b 3.9a,b
Ever Tried One or More Product 41.5 52.4a 40.6b 36.7b 23.53***
Current Use of One or More Product 23.8 31.5a 22.5b 21.4b 14.41**
Ever Tried 2 or More Product 31.4 39.8a 30.8b 27.7b 15.87***
Current Use of 2 or More Product 13.5 18.6a 12.5b 12.2b 9.57**
Open in a new tab

Note. Within rows, superscripts reflect result of Bonferroni-corrected z-tests; Cells with different letters significantly differ from one another at p < .05;

*

p < .05,

**

p < .01,

***

p < .001

Table 2. Distribution of Responses to Family Affluence Scale Items across SES Category (N = 1932).

Total Low SES (18.1%, N = 349) Medium SES (48.8%, N = 943) High SES (33.1%, N = 640)
Number of Computers at Home
 None 1.6 8.1 0.2 0.0
 One 9.4 39.2 4.9 0.0
 2 22.9 41.3 26.7 7.4
 More than 2 66.1 11.3 68.2 92.6
Family Has Vehicle
 No 2.0 7.6 1.2 0.3
 Yes 98.0 92.4 98.8 99.7
Has Own Bedroom
 No 11.5 30.5 9.8 3.9
 Yes 88.5 69.5 90.2 96.1
Vacations in Past Year
 Not at all 21.9 68.3 19.7 0.0
 Once 34.5 28.5 60.3 0.0
 Twice 23.8 2.0 17.4 45.2
 More than Twice 19.7 1.2 2.6 54.8
Open in a new tab

As Table 1 shows, among adolescents from low SES backgrounds, ever use ranged from 8.6% (smokeless tobacco) to 21.2% (e-cigarette), and current use ranged from 6.9% (smokeless tobacco) to 20.1% (e-cigarette). For adolescents from medium SES backgrounds, ever use rates ranged from 6.0% (smokeless tobacco) to 19.4% (e-cigarettes) and current use rates ranged from 3.4% (smokeless tobacco) to 14.5% (e-cigarettes). For adolescents from high SES backgrounds, ever use rates ranged from 3.9% (smokeless tobacco) to 16.3% (e-cigarettes) and current use rates ranged from 3.9% (smokeless tobacco) to 13.8% (e-cigarettes). Note, across all 3 SES groups, ever and current use rates were lowest for smokeless tobacco and highest for e-cigarette.

Lower SES was associated with high rates of use for e-cigarettes, [χ2(4) = 14.36, p < .01], blunts [χ2(4) = 9.88, p < .05], cigarettes [χ2(4) = 43.83, p < .001], cigars, [χ2(4) = 12.11, p < .05], and smokeless tobacco [χ2(4) = 17.86, p < .01] (Table 2). SES was not significantly associated with use of cigarillos [χ2(4) = 6.63, p = .16] or hookah [χ2(4) = 7.85, p = .10]. Lower SES also was associated with greater likelihood of having ever tried one or more tobacco products [χ2(2) = 23.53, p < .01] and current use of one or more tobacco products [χ2(2) = 14.41, p < .01]. In addition, lower SES was associated with greater likelihood of having ever tried 2 or more tobacco products [χ2(2) = 15.87, p < .001] and current use of 2 or more tobacco products [χ2(2) = 9.57, p < .01].

SES groups also differed by age, sex and race: (1) adolescents of low and medium SES backgrounds tended to be younger than adolescents of high SES background [F(2, 1929) = 5.29, p < .01]; (2) adolescents of low SES background were less likely to be female than male, compared to adolescents of medium and high SES background [χ2(2) = 22.21, p < .001]; and (3) adolescents of low SES background were less likely to be white than adolescents of medium and high SES background [χ2(2) = 20.73, p < .001]. Therefore, to address potential confounding effects, we controlled for these variables in our analyses.

E-cigarette and Tobacco Use Latent Classes

Table 3 shows that all fit statistics improved as we progressed from testing a one-class model to a 3-class model. In the 4-class model, all fit statistics improved except for entropy. As BIC tends to perform well for categorical data,46 we continued to fit additional models (5-class model, 6-class model) until BIC showed a distinct minimum value (which occurred in the 5-class model). We opted for the 5-class solution because it showed improved fit relative to prior models and had theoretically valid latent classes. Response probabilities for class membership are shown in Table 4. We named classes based on highest response probability for each tobacco product use category. Class 1 (4.2%; current polytobacco use group) included youth with high probability of current use for all of the tobacco products (e-cigarette, blunt, cigarettes, cigar, cigarillo, hookah, and smokeless tobacco) studied. Class 2 (6.8%; ever polytobacco use group) included youth with high probability of ever use for all of the tobacco products studied. Class 3 (6.9%; current e-cigarette, blunt and cigarette use group) included youth with high probability of current use for e-cigarettes, blunts and cigarettes. Class 4 (14.3%; ever e-cigarette, blunt use group) included youth with high probability of ever use for e-cigarettes and blunts. Class 5 (67.8%; never use group) included youth with low probability of ever or current use for all of the tobacco products (e-cigarette, blunt, cigarettes, cigar, cigarillo, hookah, and smokeless tobacco) studied.

Table 3. Fit Indices of Latent Class Analyses of the Tobacco and Modified Tobacco Products (N = 1932).

Number of Classes Log-Likelihood AIC BIC Adjusted BIC Entropy
1 -8339.128 16706.256 16784.184 16739.706 N/A
2 -6333.486 12724.972 12886.395 12794.261 0.918
3 -5949.734 11987.469 12232.386 12092.597 0.923
4 -5765.489 11648.977 11977.390 11789.946 0.896
5 -5685.566 11519.131 11931.040 11695.939 0.907
6 -5656.149 11490.298 11985.700 11702.945 0.916
Open in a new tab

Note. Text in bold indicates the optimal class solution chosen based on empirical and theoretical considerations; AIC, Akaike Information Criterion; BIC, Bayesian Information Criterion; N/A, not applicable.

Table 4. Demographics and Probability of Class Membership for Tobacco Product Use by Latent Classes (N = 1932).

Total Class 1-Current Polytobacco Users (4.2%, N = 82) Class 2-Ever Polytobacco Users (6.8%, N = 134) Class 3-Current E-cigarette, Blunt and Cigarette users (6.9%, N = 133) Class 4-Ever E-cigarette and Blunt Users (14.3%, N = 274) Class 5-Never Users (67.8%, N = 1309)
Demographics
Age (Mean, SD) 16.0 (1.3) 16.9 (1.0) 16.7(1.2) 16.0 (1.2) 16.4 (1.1) 15.8 (1.3)
Sex (%)
 Female 50.5 20.7 27.6 61.7 55.8 52.4
Race (%)
 White 88.4 84.1 91.0 85.7 88.3 88.7
SES (%)
 Low SES 18.1 35.4 23.9 25.6 31.8 35.0
 Medium 48.8 47.6 50.0 48.9 46.4 49.3
 High SES 33.1 17.1 26.1 25.6 21.9 15.7
E-cigarettes (%)
 Never 66.0 4.0 1.5 4.9 26.2 92.4
 Ever 18.7 9.3 71.3 2.7 67.4 4.2
 Past Month 15.3 86.7 27.1 92.4 6.4 3.4
Blunts (%)
 Never 76.8 17.1 8.6 36.8 44.8 99.2
 Ever 12.1 6.3 61.8 4.8 45.2 0.4
 Past Month 11.2 76.6 29.6 58.3 10.0 0.4
Cigarettes (%)
 Never 77.2 20.9 13.3 29.8 57.6 96.9
 Ever 11.9 5.3 44.6 29.0 32.3 2.4
 Past Month 11.0 73.7 42.1 41.2 10.1 0.8
Cigarillos (%)
 Never 86.6 0.0 6.6 88.9 87.3 99.9
 Ever 8.2 5.7 87.3 2.4 12.5 0.0
 Past Month 5.2 94.3 6.2 8.7 0.2 0.1
Cigars (%)
 Never 83.5 2.9 5.8 75.7 73.7 99.7
 Ever 10.3 7.4 86.2 6.0 24.2 0.0
 Past Month 6.2 89.7 8.0 18.3 2.1 0.3
Hookahs (%)
 Never 80.8 26.6 19.7 45.9 60.4 99.0
 Ever 13.0 19.3 76.6 10.4 37.8 0.6
 Past Month 6.2 54.1 3.7 43.6 1.9 0.4
Smokeless Tobacco (%)
 Never 90.0 30.7 34.8 87.4 89.2 99.9
 Ever 5.8 8.9 58.6 2.1 8.5 0.0
 Past Month 4.2 60.4 6.6 10.5 2.2 0.1
Open in a new tab

SES and Polytobacco Use Latent Classes

Table 5 shows results from the multinomial regression examining the association between SES and tobacco use profiles, with the never use group (class 5) as the reference group. Results using other classes as reference groups are not presented because no statistically significant associations were observed in those comparisons

Table 5. Odds Ratios and Confidence Intervals from Multinomial Regression Examining the Association Between SES and Tobacco Use Latent Classes (N = 1932).

Class 1-Current Polytobacco Use Group (ref: Class 5 – Never User Group) Class 2-Ever Polytobacco Use Group (ref: Class 5 – Never User Group)
OR 95% CI OR 95% CI
Age 2.06 1.65 2.57 1.79 1.50 2.12
Female (ref: Male) 0.25 0.11 0.53 0.37 0.22 0.61
Low SES (ref: High SES) 0.88 0.35 2.20 2.23 1.22 4.07
Low SES (ref: Med SES) 1.04 0.42 2.54 1.59 0.91 2.77
Med SES (ref: High SES) 0.85 0.49 1.49 1.40 0.84 2.33
White (ref: other) 0.68 0.30 1.53 1.48 0.66 3.32
School 1 0.62 0.35 1.12 0.85 0.52 1.38
School 2 0.50 0.23 1.09 0.94 0.55 1.61
Class 3-Current E-cigarette, Blunt and Cigarette Use Group (ref: Class 5 – Never User Group) Class 4-Ever E-cigarette and Blunt Use Group (ref: Class 5 – Never User Group)
OR 95% CI OR 95% CI
Age 1.08 0.91 1.27 1.38 1.20 1.58
Female (ref: Male) 1.65 1.01 2.69 1.16 0.84 1.61
Low SES (ref: High SES) 2.21 1.19 4.10 1.58 1.01 2.46
Low SES (ref: Med SES) 1.77 0.98 3.17 1.55 1.02 2.34
Med SES (ref: High SES) 1.25 0.72 2.17 1.02 0.71 1.46
White (ref: other) 0.84 0.46 1.53 0.93 0.58 1.48
School 1 1.75 0.98 3.15 1.01 0.71 1.44
School 2 0.95 0.49 1.83 0.89 0.59 1.34
Open in a new tab

Note. Bold indicates statistical significance; p < .05.

Class 1: current polytobacco use group

There was no association between low SES, relative to high SES, and likelihood of being in the current polytobacco use group (OR 0.88; 95% CI 0.35, 2.20) relative to the never use group (class 5). Low SES relative to medium SES, was not associated with membership in the current polytobacco use group (OR 1.04; 95% CI 0.42, 2.54). There was no association between medium SES, relative to high SES, and membership in the current polytobacco use group (OR 0.85; 95% CI 0.49, 1.49). Females were less likely to belong to the current polytobacco use group (OR 0.25; 95% CI 0.11, 0.53) than males. Older youth were more likely to be in the current polytobacco use group (OR 2.06; 95% CI 1.65, 2.57) than younger youth. White race was not associated with membership in the current polytobacco use group (OR 0.68; 95% CI 0.30, 1.53).

Class 2: ever polytobacco use group

Low SES, relative to high SES, was associated with greater likelihood of being in the ever polytobacco use group (OR 2.23; 95% CI 1.22, 4.07). However, low SES, relative to medium SES, was not associated with membership in the ever polytobacco use group (OR 1.59; 95% CI 0.91, 2.77). There was no association between medium SES, relative to high SES, and membership in the ever polytobacco use group (class 2; OR 1.40; 95% CI 0.84, 2.33). Females were less likely to belong to the ever poly-tobacco use group (OR: 0.37; 95% CI 0.22, 0.61) than males. Older youth were more likely to be in the ever polytobacco use group (OR 1.79; 95% CI 1.50, 2.12) than younger youth. White race was not associated with membership in the ever poly-tobacco use group (OR 1.48; 95% CI 0.66, 3.32)

Class 3: current e-cigarette, blunt and cigarette use group

Low SES, relative to high SES, was associated with greater likelihood of being in the current e-cigarette, blunt and cigarette use group (OR 2.21; 95% CI 1.19, 4.10). Relative to medium SES, low SES was not associated with membership in the current e-cigarette, blunt and cigarette user group (OR 1.77; 95% CI 0.98, 3.17). There was no association between medium SES, relative to high SES, and membership in the current e-cigarette, blunt, and cigarette use group (OR 1.25; 95% CI 0.72, 2.17). Females were more likely to belong to the current e-cigarette, blunt, and cigarette use group (OR: 1.65; 95% CI 1.01, 2.69) than males. In contrast to the other classes, age was not associated with membership in the current e-cigarette, blunt, and cigarette use group (OR 1.08; 95% CI 0.91, 1.27). White race was not associated with membership in the current e-cigarette, blunt, and cigarette use group (OR 0.84; 95% CI 0.46, 1.53).

Class 4: ever e-cigarette and blunt use group

Low SES, relative to high SES was associated with greater likelihood of being in the ever e-cigarette, blunt use group (OR 1.58; 95% CI 1.01, 2.46). Low SES, relative to medium SES, was associated with membership in the ever e-cigarette and blunt use group (OR 1.55; 95% CI 1.02, 2.34). There was no association between medium SES, relative to high SES, and membership in the ever e-cigarette, blunt use group (OR 1.02; 95% CI 0.71, 1.46). Sex was not associated with likelihood of membership in this class (OR 1.16; 95% CI 0.84, 1.61). Older youth were more likely to be in the ever e-cigarette, blunt use group (OR 1.38; 95% CI 1.20, 1.58) than younger youth. White race was not associated with membership in the ever e-cigarette, blunt use group (OR 0.93; 95% CI 0.58, 1.48).

Discussion

We examined the association between SES and polytobacco use profiles that included e-cigarettes, blunts, cigarettes, cigarillos, cigars, hookah, and smokeless tobacco. We identified 5 latent classes: (1) current polytobacco users; (2) ever polytobacco users; (3) current e-cigarette, blunt, and cigarette users; (4) ever e-cigarette and blunt users; and (5) never users (ie, abstainers). These classes generally were consistent with previous adolescent research.33,47 However, we identified more latent classes (5 classes), identified classes with e-cigarette and blunt use as predominate features, and found that our polytobacco use profiles included a substantial proportion of e-cigarette users, suggesting that those who have tried e-cigarettes are likely to have tried other tobacco products (or vice versa). We also found that relative to the never user group, low SES, compared to high SES, was associated with greater likelihood of membership in 3 classes: (1) ever polytobacco use group; (2) current e-cigarette, blunt and cigarette use group; and (3) ever e-cigarette and blunt use group.

Our study is unique because, in contrast to prior LCAs that are inclusive of e-cigarettes, we described the association between latent classes and SES.33,47,48 Low SES was associated with greater likelihood of membership in 3 of 5 latent classes, and was not associated with the current polytobacco use class or never-user group. Previous research has suggested a number of factors that may contribute to greater probability of tobacco product use among youth from low SES backgrounds.49 For example, adolescents from low SES backgrounds may be more likely to have parents or friends who smoke,50,51 be targeted for tobacco related advertisements,52 and use tobacco as a way to cope with stress associated with their low SES status.53 Thus, the lack of association between low SES and current polytobacco use is somewhat unexpected. Perhaps, these adolescents lack financial resources to use all of the studied tobacco products at once. It is also possible factors such as parental substance use, biologic predispositions or cultural norms supersede the effects of SES in this population. Given that we did not examine the aforementioned factors, it is essential that future studies examine them as potential mediators of the relationship between SES and alternative tobacco product use. It is also surprising that we did not consistently observe significant associations for comparisons involving medium SES, as other studies have shown that each level of increase in SES (ie, income) is associated with corresponding protective health effects for youth.54

Another unique finding in the current study is that the polytobacco use profiles included blunts, a tobacco product that largely has not been included in previous studies of polytobacco use. It is surprising that latent classes defined by blunt use would emerge for an overwhelmingly white sample, as blunt use tends to occur more often among African Americans.18 However, more recent trends have suggested blunt use has spread to other racial groups.4

A secondary finding of this study is that girls were more likely than boys to be current e-cigarette, blunt and cigarette users, but less likely than boys to be current polytobacco users and ever polytobacco users. Whereas boys have traditionally been at risk for greater tobacco use, a recent nationally representative study has shown an increasing trend in concurrent cigarette and cigar use for girls, but a decreasing trend for boys from 1999 to 2013.55 Extending this prior work, our findings suggest that girls may be uniquely at risk for combinations of polytobacco use that include blunts as a unique cigar product, e-cigarettes, and cigarettes.

Our study had several limitations. Due to the cross-sectional nature of our study, we cannot make causal inferences or determine whether e-cigarette use preceded use of the other products. The generalizability of the findings may be limited by the inclusion of a convenience sample of students from predominantly white, middle-to-high income backgrounds in Connecticut. In particular, we may not have been able to detect unique findings in black students due to sample size limitations, as we did not identify a class of blunt/cigar use (which has been shown previously to be a typical use pattern in this population).56-59 Also, a substantial portion of the race/ethnicity data were missing (27.6%). We minimized the effects of missing data by using multiple imputation. This approach helped mitigate the effects of missing data bias by allowing us to retain cases that would have otherwise been excluded from the analysis via list wise deletion. The imputation resulted in demographics that are consistent with the demographics of the schools' towns. It is important to point out that our results were consistent across 3 missing data approaches (ie, complete case analysis, multiple imputation, and including an indicator for missing race). Lastly, we observed differences in age, sex and race across levels of SES. However, we controlled for potential confounding effects by including these demographic factors in our analysis.

Despite these limitations, our study extends the existing literature on polytobacco use among adolescents by including e-cigarettes and blunts (along with 5 other tobacco products) in an LCA, and showing that low SES, relative to high SES, was associated with greater likelihood of being: (1) an ever polytobacco user; (2) a current e-cigarette, blunt, and cigarette user; and (3) an ever e-cigarette and blunt user, relative to a never user. Future research should examine how polytobacco use that is inclusive of e-cigarettes emerges among low-in-come adolescents and how such use is associated with age of initiation and progression of tobacco use over time. There also is a need for research that aims to develop a better understanding of why low SES youth are more susceptible to polytobacco use.

Implications for Tobacco Regulation

Our findings have significant implications for tobacco regulatory policies that aim to prevent poly-tobacco use and tobacco-related health disparities in low SES adolescents. In particular, the higher probability of certain polytobacco use patterns in low SES youth may widen tobacco-related health disparities by conferring heightened risk of chronic health problems in a group that may not have the resources to access and engage in appropriate healthcare. The current study suggests that regulatory initiatives should focus on polytobacco use, as those that focus exclusively on cigarettes miss the opportunity to have a direct influence on low-income individuals. Future research should examine whether targeting a broader range of products helps reduce SES-related tobacco use disparities. Tobacco regulatory efforts have contributed to steady declines in tobacco use over several decades; however, high SES adolescents have experienced greater declines than low SES adolescents.60 Overall, our findings suggest that socioeconomic differences in tobacco use persist.

Acknowledgments

This study was supported by a grant to Dr Krishnan-Sarin through the National Institute on Drug Abuse (NIDA) and the Food and Drug Agency (FDA) Center for Tobacco Products (P50DA036151). Dr Patricia Simon's efforts were supported by NIDA/FDA grant P50DA036151 and NIDA grant T32DA019426. Dr Grace Kong's and Dr Deepa Camenga's efforts also were supported in part by K12DA033012, CTSA grants UL1 TR000142, and KL2 TR000140 from the National Center for Advancing Translational Science (NCATS), components of the National Institutes of Health (NIH), and NIH roadmap for Medical Research. The content is solely the responsibility of the authors and does not represent the views of the funding agencies.

Footnotes

Human Subjects Statement: The Yale University Institutional Review Board and each school's local administrators/superintendents approved this study.

Conflict of Interest Disclosure Statement: All authors of this article declare they have no conflicts of interest.

Contributor Information

Patricia Simon, Yale School of Medicine, Department of Psychiatry & The Consultation Center, New Haven, CT.

Deepa R. Camenga, Yale School of Medicine, Department of Emergency Medicine, New Haven, CT.

Grace Kong, Yale School of Medicine, Department of Psychiatry, New Haven, CT.

Christian M. Connell, Yale School of Medicine, Department of Psychiatry & The Consultation Center, New Haven, CT.

Meghan E. Morean, Department of Psychology, Oberlin, OH.

Dana A. Cavallo, Yale School of Medicine, Department of Psychiatry, New Haven, CT.

Suchitra Krishnan-Sarin, Yale School of Medicine, Department of Psychiatry, New Haven, CT.

References

  • 1.Johnston LD, O'Malley PM, Bachman JG, et al. Monitoring the Future Occasional Paper No 81. Ann Arbor, MI: The University of Michigan, Institute for Social Research; 2014. [Accessed October 28, 2016]. Demographic subgroup trends among adolescents in the use of various licit and illicit drugs, 1975–2013. Available at: http://www.monitoringthefuture.org/pubs/occpapers/mtf-occ81.pdf. [Google Scholar]
  • 2.Singh T. Tobacco use among middle and high school students – United States, 2011–2015. MMWR Morb Mortal Wkly Rep. 2016;65(14):361–367. doi: 10.15585/mmwr.mm6514a1. [DOI] [PubMed] [Google Scholar]
  • 3.Morean ME, Kong G, Camenga DR, et al. Latent class analysis of current e-cigarette and other substance use in high school students. Drug Alcohol Depend. 2016;161:292–297. doi: 10.1016/j.drugalcdep.2016.02.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Timberlake DS. The changing demographic of blunt smokers across birth cohorts. Drug Alcohol Depend. 2013;130(1-3):129–134. doi: 10.1016/j.drugalcdep.2012.10.022. [DOI] [PubMed] [Google Scholar]
  • 5.Green MJ, Leyland AH, Sweeting H, et al. Socioeconomic position and early adolescent smoking development: evidence from the British Youth Panel Survey (1994-2008) Tob Control. 2016;25(2):203–210. doi: 10.1136/tobaccocontrol-2014-051630. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Hiscock R, Bauld L, Amos A, et al. Socioeconomic status and smoking: a review. Ann N Y Acad Sci. 2012;1248(1):107–123. doi: 10.1111/j.1749-6632.2011.06202.x. [DOI] [PubMed] [Google Scholar]
  • 7.Scarinci IC, Robinson LA, Alfano CM, et al. The relationship between socioeconomic status, ethnicity, and cigarette smoking in urban adolescents. Prev Med. 2002;34(2):171–178. doi: 10.1006/pmed.2001.0967. [DOI] [PubMed] [Google Scholar]
  • 8.Huang J, Kim Y, Vera L, et al. Electronic cigarettes among priority populations: role of smoking cessation and tobacco control policies. Am J Prev Med. 2015;50(2):199–209. doi: 10.1016/j.amepre.2015.06.032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Pentz MA, Shin H, Riggs N, et al. Parent, peer and executive function relationships to early adolescent e-cigarette use: a substance use pathway? Addict Behav. 2015;42:73–78. doi: 10.1016/j.addbeh.2014.10.040. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Barrington-Trimis JL, Berhane K, Unger JB, et al. Psychosocial factors associated with adolescent electronic cigarette and cigarette use. Pediatrics. 2015;136(2):308–317. doi: 10.1542/peds.2015-0639. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Moore G, Hewitt G, Evans J, et al. Electronic-cigarette use among young people in Wales: evidence from 2 cross-sectional surveys. BMJ Open. 2015;5(4):e007072. doi: 10.1136/bmjopen-2014-007072. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Borawski EA, Brooks A, Colabianchi N, et al. Adult use of cigars, little cigars, and cigarillos in Cuyahoga County, Ohio: a cross-sectional study. Nicotine Tob Res. 2010;12(6):669–673. doi: 10.1093/ntr/ntq057. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Jamil H, Elsouhag D, Hiller S, et al. Sociodemographic risk indicators of hookah smoking among white Americans: a pilot study. Nicotine Tob Res. 2010;12(5):525–529. doi: 10.1093/ntr/ntq026. [DOI] [PubMed] [Google Scholar]
  • 14.White TJ, Redner R, Bunn JY, et al. Do socioeconomic risk factors for cigarette smoking extend to smokeless tobacco use? Nicotine Tob Res. 2015;18(5):869–873. doi: 10.1093/ntr/ntv199. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Melnick MJ, Miller KE, Sabo DF, et al. Tobacco use among high school athletes and nonathletes: results of the 1997 Youth Risk Behavior Survey. Adolescence. 2001;36(144):727–747. [PubMed] [Google Scholar]
  • 16.Leatherdale ST, Rios P, Elton-Marshall T, et al. Cigar, cigarillo, and little cigar use among Canadian youth: are we underestimating the magnitude of this problem? J Prim Prev. 2011;32(3-4):161–170. doi: 10.1007/s10935-011-0248-6. [DOI] [PubMed] [Google Scholar]
  • 17.Wilkinson AV, Koehly LM, Vandewater EA, et al. Demographic, psychosocial, and genetic risk associated with smokeless tobacco use among Mexican heritage youth. BMC Med Genet. 2015;16(43):1–8. doi: 10.1186/s12881-015-0188-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Soldz S, Huyser DJ, Dorsey E. The cigar as a drug delivery device: youth use of blunts. Addiction. 2003;98(10):1379–1386. doi: 10.1046/j.1360-0443.2003.00492.x. [DOI] [PubMed] [Google Scholar]
  • 19.Hanson M, Chen E. Socioeconomic status and health behaviors in adolescence: a review of the literature. J Behav Med. 2007;30(3):263–285. doi: 10.1007/s10865-007-9098-3. [DOI] [PubMed] [Google Scholar]
  • 20.Patrick ME, Wightman P, Schoeni RF, et al. Socioeconomic status and substance use among young adults: a comparison across constructs and drugs. J Stud Alcohol Drugs. 2012;73(5):772–782. doi: 10.15288/jsad.2012.73.772. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Zapata Roblyer MI, Betancourth S, Grzywacz JG. Risk and protective factors for lifetime marijuana use among Colombian emergent adults attending college. ISSBD Bull. 2015;2015(1):5–9. [PMC free article] [PubMed] [Google Scholar]
  • 22.Galea S, Ahern J, Tracy M, et al. Neighborhood income and income distribution and the use of cigarettes, alcohol, and marijuana. Am J Prev Med. 2007;32(6):S195–S202. doi: 10.1016/j.amepre.2007.04.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Amrock SM, Gordon T, Zelikoff JT, et al. Hookah use among adolescents in the United States: results of a national survey. Nicotine Tob Res. 2014;16(2):231–237. doi: 10.1093/ntr/ntt160. [DOI] [PubMed] [Google Scholar]
  • 24.Palamar JJ, Zhou S, Sherman S, et al. Hookah use among U.S. high school seniors. Pediatrics. 2014;134(2):227–234. doi: 10.1542/peds.2014-0538. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Fix BV, O'Connor RJ, Vogl L, et al. Patterns and correlates of polytobacco use in the United States over a decade: NSDUH 2002-2011. Addict Behav. 2014;39(4):768–781. doi: 10.1016/j.addbeh.2013.12.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Bombard JM, Rock VJ, Pederson LL, et al. Monitoring polytobacco use among adolescents: do cigarette smokers use other forms of tobacco? Nicotine Tob Res. 2008;10(11):1581–1589. doi: 10.1080/14622200802412887. [DOI] [PubMed] [Google Scholar]
  • 27.Lee YO, Hebert CJ, Nonnemaker JM, et al. Youth tobacco product use in the United States. Pediatrics. 2015;135(3):409–415. doi: 10.1542/peds.2014-3202. [DOI] [PubMed] [Google Scholar]
  • 28.Arrazola RA, Kuiper NM, Dube SR. Patterns of current use of tobacco products among U.S. high school students for 2000-2012--findings from the National Youth Tobacco Survey. J Adolesc Health. 2014;54(1):54–60. doi: 10.1016/j.jadohealth.2013.08.003. e59. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Soneji S, Sargent J, Tanski S. Multiple tobacco product use among US adolescents and young adults. Tob Control. 2016;25:174–180. doi: 10.1136/tobaccocontrol-2014-051638. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Saunders C, Geletko K. Adolescent cigarette smokers' and non-cigarette smokers' use of alternative tobacco products. Nicotine Tob Res. 2012;14(8):977–985. doi: 10.1093/ntr/ntr323. [DOI] [PubMed] [Google Scholar]
  • 31.Arrazola RA, Singh T, Corey CG, et al. Tobacco use among middle and high school students – United States, 2011-2014. MMWR Morb Mortal Wkly Rep. 2015;64(14):381–385. [PMC free article] [PubMed] [Google Scholar]
  • 32.Currie C, Molcho M, Boyce W, et al. Researching health inequalities in adolescents: the development of the Health Behaviour in School-Aged Children (HBSC) family affluence scale. Soc Sci Med. 2008;66(6):1429–1436. doi: 10.1016/j.socscimed.2007.11.024. [DOI] [PubMed] [Google Scholar]
  • 33.Gilreath TD, Leventhal A, Barrington-Trimis JL, et al. Patterns of alternative tobacco product use: emergence of hookah and e-cigarettes as preferred products amongst youth. J Adolesc Health. 2016;58(2):181–185. doi: 10.1016/j.jadohealth.2015.10.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Gilreath TD, Astor RA, Estrada JN, et al. Substance use among adolescents in California: a latent class analysis. Subst Use Misuse. 2014;49(1-2):116–123. doi: 10.3109/10826084.2013.824468. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Tomczyk S, Isensee B, Hanewinkel R. Latent classes of polysubstance use among adolescents – a systematic review. Drug Alcohol Depend. 2016;160:12–29. doi: 10.1016/j.drugalcdep.2015.11.035. [DOI] [PubMed] [Google Scholar]
  • 36.Krishnan-Sarin S, Morean ME, Camenga DR, et al. E-cigarette use among high school and middle school adolescents in Connecticut. Nicotine Tob Res. 2015;17(7):810–818. doi: 10.1093/ntr/ntu243. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Kong G, Morean ME, Cavallo DA, et al. Reasons for electronic cigarette experimentation and discontinuation among adolescents and young adults. Nicotine Tob Res. 2015;17:847–854. doi: 10.1093/ntr/ntu257. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Bold KW, Kong G, Cavallo DA, et al. Reasons for trying e-cigarettes and risk of continued use. Pediatrics. 2016;138(3):e20160895. doi: 10.1542/peds.2016-0895. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Boyce W, Dallago L. Socioeconomic inequality. In: Currie C, Roberts C, Morgan A, et al., editors. Young People's Health in Context. Health Behaviour in School-aged Children (HBSC) Study: International Report from the 2001/2002 Survey. Copenhagen, Denmark: World Health Organization; 2004. [Accessed October 28, 2016]. pp. 13–25. Available at: http://www.who.int/immunization/hpv/target/young_peoples_health_in_context_who_2011_2012.pdf. [Google Scholar]
  • 40.Boyce W, Torsheim T, Currie C, et al. The family affluence scale as a measure of national wealth: validation of an adolescent self-report measure. Soc Indic Res. 2006;78(3):473–487. [Google Scholar]
  • 41.Botello-Harbaum MT, Haynie DL, Iannotti RJ, et al. Tobacco control policy and adolescent cigarette smoking status in the United States. Nicotine Tob Res. 2009;11(7):875–885. doi: 10.1093/ntr/ntp081. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.IBM. Version 21. IBM; Armonk, NY: 2012. [Google Scholar]
  • 43.Muthén B, Muthé L. Mplus Version 7: User's Guide. Vol. 2013 Los Angeles, CA: Muthén and Muthén;; 1998-2013. [Google Scholar]
  • 44.Graham JW, Olchowski AE, Gilreath TD. How many imputations are really needed? some practical clarifications of multiple imputation theory. Prev Sci. 2007;8(3):206–213. doi: 10.1007/s11121-007-0070-9. [DOI] [PubMed] [Google Scholar]
  • 45.Collins LM, Lanza ST. Latent Class and Latent Transition Analysis: With Applications in the Social, Behavioral, and Health Sciences. Hoboken, NJ: John Wiley & Sons; 2013. [Google Scholar]
  • 46.Nylund KL, Asparouhov T, Muthén BO. Deciding on the number of classes in latent class analysis and growth mix- ture modeling: a monte carlo simulation study. Struct Equ Modeling. 2007;14(4):535–569. [Google Scholar]
  • 47.Nădăşan V, Foley KL, Pénzes M, et al. Use of electronic cigarettes and alternative tobacco products among Romanian adolescents. Int J Public Health Res. 2016;61(2):199–207. doi: 10.1007/s00038-015-0774-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Huh J, Leventhal AM. Progression of poly-tobacco product use patterns in adolescents. Am J Prev Med. 2016;51(4):513–517. doi: 10.1016/j.amepre.2016.04.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Mathur C, Stigler MH, Erickson DJ, et al. Socioeconomic differences in the risk profiles of susceptibility and ever use of tobacco among Indian urban youth: a latent class approach. Nicotine Tob Res. 2014;16(4):454–460. doi: 10.1093/ntr/ntt184. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Chassin L, Presson CC, Sherman SJ, et al. Parent educational attainment and adolescent cigarette smoking. J Subst Abuse. 1992;4(3):219–234. doi: 10.1016/0899-3289(92)90031-r. [DOI] [PubMed] [Google Scholar]
  • 51.Madarasová Gecková A, Stewart R, van Dijk JP, et al. Influence of socio-economic status, parents and peers on smoking behaviour of adolescents. Eur Addict Res. 2005;11(4):204–209. doi: 10.1159/000086403. [DOI] [PubMed] [Google Scholar]
  • 52.Barbeau EM, Leavy-Sperounis A, Balbach ED. Smoking, social class, and gender: what can public health learn from the tobacco industry about disparities in smoking? Tob Control. 2004;13(2):115–120. doi: 10.1136/tc.2003.006098. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Scales MB, Monahan JL, Rhodes N, et al. Adolescents' perceptions of smoking and stress reduction. Health Educ Behav. 2009;36(4):746–758. doi: 10.1177/1090198108317628. [DOI] [PubMed] [Google Scholar]
  • 54.Evans W, Wolfe B, Adler N. The SES and Health Gradient: A Brief Review of the Literature. New York, NY: Russell Sage; 2012. [Google Scholar]
  • 55.Creamer MR, Perry CL, Harrell MB, et al. Trends in multiple tobacco product use among high school students. Tob Regul Sci. 2015;1(3):204–214. doi: 10.18001/trs.1.3.2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Thiri Aung A, Pickworth WB, Moolchan ET. History of marijuana use and tobacco smoking topography in tobacco-dependent adolescents. Addict Behav. 2004;29(4):699–706. doi: 10.1016/j.addbeh.2004.02.012. [DOI] [PubMed] [Google Scholar]
  • 57.Guerra LM, Romano PS, Samuels SJ, et al. Ethnic differences in adolescent substance initiation sequences. Arch Pediatr Adolesc Med. 2000;154(11):1089–1095. doi: 10.1001/archpedi.154.11.1089. [DOI] [PubMed] [Google Scholar]
  • 58.Vaughn M, Wallace J, Perron B, et al. Does marijuana use serve as a gateway to cigarette use for high-risk African-American youth? Am J Drug Alcohol Abuse. 2008;34(6):782–791. doi: 10.1080/00952990802455477. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59.White HR, Jarrett N, Valencia EY, et al. Stages and sequences of initiation and regular substance use in a longitudinal cohort of black and white male adolescents. J Stud Alcohol Drugs. 2007;68(2):173–181. doi: 10.15288/jsad.2007.68.173. [DOI] [PubMed] [Google Scholar]
  • 60.US Depatment of Health and Human Services (US-DHHS) Preventing Tobacco Use among Youth and Young Adults: A Report of the Surgeon General. Atlanta, GA: US-DHHS, US Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion Office on Smoking and Health; [Accessed October 28, 2016]. p. 2012. Available at: http://www.surgeongeneral.gov/library/reports/preventing-youth-tobacco-use/ [Google Scholar]

RESOURCES