Background
The rapidly growing use of nicotine electronic vaping products, commonly referred to as electronic cigarettes, e-cigarettes, and vaping, among youth is a significant concern among health agencies globally, with some public health officials referring to it as an epidemic [
1‐
4]. As an example of the harm associated with e-cigarettes, the significant increase in their use was identified as the main driver in increased tobacco use among youth in 2017–2018, based on a United States (US) nationally representative sample of grades 6 to 12 students [
5]. This essentially erased the progress made in tobacco reduction from 2011 to 2017; e-cigarettes have become the most commonly used tobacco product among middle and high school students since 2014 [
5,
6]. In Canada and the US in 2018, one study noted the highest increase in vaping prevalence documented thus far (increase of five and six percentage points, respectively) among adolescents (ages 16 to 19 years) [
7]. Increases in e-cigarette use have also been noted among young adults (ages 18 to 24 years) in the U.S. whereas the prevalence in older age groups has remained the same or even declined from 2014 to 2018 [
8]. There have been frequent calls to action to protect youth by preventing vaping initiation and helping them discontinue use [
2].
Understanding which factors are associated with youth vaping is important to inform evidence-based approaches to prevention and cessation of e-cigarette use among youth. In recent years, a growing literature shows that several risk factors, including sociodemographic variables, substance use (including tobacco use), psychological factors, household tobacco use, relationship with parents, are associated with various vaping indicators in adolescence including susceptibility to initiation (i.e., vaping intention and willingness), initiation, lifetime use, current use, and dual use with traditional tobacco products [
9‐
14]. For example, in a longitudinal study among a nationally representative sample of U.S. adolescents ages 12 to 17 years, the authors found past-year high externalizing problems, alcohol use, and older age to be associated with subsequent initiation of e-cigarette and dual (e-cigarette and combustible cigarette) use [
14]. The same study also found past-year internalizing problems associated with subsequent initiation of e-cigarette use, past year marijuana use to be significantly associated with dual (e-cigarette and combustible cigarette) use, and African American students to be less likely to initiate e-cigarette and dual use compared to White adolescents [
14]. A review of recent literature on sex and gender differences in e-cigarette use among American adolescents found limited and mixed findings; some studies reported e-cigarette use to be more prevalent among males, and other studies examined reported no sex and gender differences [
15].
The relationship between various adverse childhood experiences (ACEs, [
16]) and more traditional tobacco or nicotine use has been widely studied in the literature finding consistent associations between ACEs and tobacco use [
17‐
21], but less is known about the association between ACEs and e-cigarette use. A few studies have examined the relationship between ACEs and vaping in young adult and adult samples. In a U.S. urban community sample of young adults aged 18 to 21 years, the authors found physical abuse, sexual abuse, emotional abuse, emotional neglect, and physical neglect to be associated with greater likelihood of lifetime e-cigarette use [
22]. In a sample of adult Australian women (ages 19–26 years), the authors found psychological abuse, physical abuse, sexual abuse, household substance abuse, witnessing domestic violence, household mental illness, and parental separation or divorce (but not an incarcerated household member) were significantly associated with past-year and lifetime e-cigarette use [
23]. Another study involving adults in central Florida, found an association between a history of four or more ACEs and e-cigarette use [
24].
A few studies have examined the relationship between ACEs and vaping in adolescent samples [
25‐
27]. One study conducted with a sample of students ages 13 to 18 years in Bangkok, Thailand, found several ACEs to be associated with higher odds of lifetime e-cigarette use [
25]. A study using data from Wave 1 of the Well-Being and Experiences (WE) Study (same data used in the current study) conducted in Manitoba, Canada, examined the relationship between an expanded list of ACEs [
26] and past 30-day electronic vapor product use [
27]. The findings showed a significant association with emotional abuse, exposure to verbal intimate partner violence (IPV), parental separation or divorce, household mental illness, household substance use, parental problem gambling, and parental problems with police (but not emotional neglect, spanking, foster care placement or contact with a child protective organization (CPO), an unsafe neighbourhood, poverty in the unadjusted model, and peer victimization without other ACEs) [
27]. Furthermore, a few studies (both young adult and adolescent samples) have also found graded relationships between the number of ACEs and vaping outcomes [
23,
25,
28]. Among studies that examined emotional abuse, physical abuse, sexual abuse, and parental separation or divorce, they have all shown an association with vaping outcomes [
22,
23,
27]. For several other ACEs, there have been mixed findings [
22,
23,
25,
27].
While there is an emerging and growing literature examining risk factors associated with vaping, important gaps remain. Small community samples, along with use of different measures, various age groups, and varied vaping outcomes may, in part, explain the mixed findings noted in the literature, particularly as they relate to sex and gender [
15] and ACEs. A notable gap in the literature is the lack of longitudinal studies. One of the few longitudinal studies published examined the relationship between internalizing and externalizing problems and initiation of e-cigarette use, but not how vaping use changed over time [
14]. Little is known about what factors are associated with trajectories of vaping use over time. One longitudinal study that examined the change in frequency of e-cigarette use from spring of grade 8 (about 14 years old) to spring of grade 9 (about 15 years old) only examined the relationship to sex and gender, ethnicity, and use of other substances and found greater lifetime substance use in grade 8 and current substance use in grade 9 among accelerating users [
29].
Given the limitations noted above, the objectives of the current study were to examine what risk factors, including self-reported ACEs, were associated with a) any adolescent and young adult vaping and b) adolescents’ and young adults’ course of vaping across two time points using longitudinal data.
Results
The sample prevalence of any adolescent and young adult vaping (W1 or W2) was 45.5% (
n = 331). Using the vaping across time variable, a total of 2.7% of adolescents reported vaping at W1 only (
n = 19), 19.7% at W2 only (
n = 138), and 21.2% at both W1 and W2 (
n = 149). The prevalence of any adolescent and young adult vaping by parental and adolescent risk factors and sociodemographic variables are presented in Table
1 and by ACEs in Table
2.
Table 1
Prevalence of any Wave 1 or Wave 2 vaping by parental and adolescent risk factors and sociodemographic variables
Household income |
$49,999 or less | 53.7 (66) | 46.3 (57) |
$50,000 to $99,999 | 53.3 (136) | 46.7 (119) |
$100,000 to $149,999 | 57.7 (98) | 42.4 (72) |
$150,000 or more | 52.7 (77) | 47.3 (69) |
No response | 54.8 (17) | 45.16 (14) |
Parental marital status |
Married or common-law | 56.4 (333) | 43.7 (258) |
Separated, divorced, or widowed | 50.5 (54) | 49.5 (53) |
Never married | 25.9 (7) | 74.1 (20) |
Parental smoking or vaping |
Yes | 40.4 (40) | 59.6 (59) |
No | 56.4 (349) | 43.6 (270) |
Adolescent sex |
Male | 55.2 (186) | 44.8 (151) |
Female | 53.8 (208) | 46.3 (179) |
Adolescent age |
14 | 62.2 (140) | 37.8 (85) |
15 | 54.9 (107) | 45.1 (88) |
16 | 44.9 (83) | 55.1 (102) |
17 | 54.1 (66) | 45.9 (56) |
Adolescent mental disorder |
Yes | 48.7 (76) | 51.3 (80) |
No | 56.1 (289) | 43.9 (226) |
Table 2
Prevalence of any Wave 1 or Wave 2 vaping by adolescent adverse childhood experiences
Emotional abuse |
Yes | 41.7 (60) | 58.3 (84) |
No | 57.6 (307) | 42.4 (226) |
Emotional neglect |
Yes | 36.2 (17) | 63.8 (30) |
No | 56.0 (374) | 44.0 (294) |
Exposure to verbal IPV |
Yes | 41.1 (62) | 58.9 (89) |
No | 57.5 (293) | 42.6 (217) |
Spanking |
Yes | 54.3 (115) | 45.8 (97) |
No | 54.7 (260) | 45.3 (215) |
Household substance use |
Yes | 40.4 (40) | 59.6 (59) |
No | 58.1 (330) | 41.9 (238) |
Household mental illness |
Yes | 48.2 (106) | 51.8 (114) |
No | 61.0 (230) | 39.0 (147) |
Parental separation/ divorce |
Yes | 46.6 (82) | 53.4 (94) |
No | 57.6 (293) | 42.4 (216) |
Parental problems with police |
Yes | 39.3 (24) | 60.7 (37) |
No | 56.2 (337) | 43.8 (263) |
Parental gambling |
Yes | 33.3 (7) | 66.7 (14) |
No | 54.7 (366) | 45.3 (303) |
Foster care or contact with CPO |
Yes | 39.5 (32) | 60.5 (49) |
No | 56.6 (356) | 43.4 (273) |
Poverty |
Yes | 53.6 (59) | 46.4 (51) |
No | 53.3 (276) | 46.7 (242) |
Unsafe neighbourhood |
Yes | 33.3 (10) | 66.7 (20) |
No | 55.2 (382) | 44.8 (310) |
Peer victimization |
Yes | 44.1 (71) | 55.9 (90) |
No | 56.5 (286) | 43.5 (220) |
Table
3 presents the associations between parental and adolescent risk factors and any adolescent and young adult vaping (W1 or W2). The following risk factors were significantly associated with increased odds of any vaping: parental smoking or vaping, emotional abuse, emotional neglect, exposure to verbal IPV, household substance use, household mental illness, parental separation/divorce (unadjusted model only), parental problems with police, foster care or contact with CPO, unsafe neighbourhood, and peer victimization.
Table 3
Parental and adolescent risk factors associated with any Wave 1 or Wave 2 adolescent and young adult vaping
Parental smoking or vaping | 1.91 (1.24, 2.94)** | 1.81 (1.14, 2.85)* |
Adolescent sex (Female vs. Male) | 1.06 (0.79, 1.42) | 1.01 (0.74, 1.36) |
Adolescent mental disorders | 1.35 (0.94, 1.93) | 1.25 (0.86, 1.82) |
Emotional abuse | 1.90 (1.31, 2.76)*** | 1.92 (1.31, 2.82)*** |
Emotional neglect | 2.24 (1.21, 4.15)** | 2.15 (1.14, 4.04)* |
Exposure to Verbal IPV | 1.94 (1.43, 2.80)*** | 1.92 (1.31, 2.80)*** |
Spanking | 1.02 (0.74, 1.41) | 1.03 (0.74, 1.45) |
Household substance use | 2.05 (1.32, 3.16)*** | 1.92 (1.22, 3.01)** |
Household mental illness | 1.68 (1.20, 2.35)** | 1.56 (1.09, 2.23)* |
Parental separation/divorce | 1.55 (1.10, 2.19)* | 1.50 (0.92, 2.44) |
Parental problems with police | 1.98 (1.15, 3.38)* | 1.91 (1.09, 3.35)* |
Parental gambling | 2.42 (0.96, 6.06) | 1.89 (0.73, 4.89) |
Foster care or contact with CPO | 2.00 (1.24, 3.20)** | 1.90 (1.15, 3.14)* |
Poverty | 0.99 (0.65, 1.49) | 0.96 (0.61, 1.51) |
Unsafe neighbourhood | 2.46 (1.14, 5.34)* | 2.25 (1.01, 5.01)* |
Peer victimization | 1.65 (1.15, 2.36)** | 1.70 (1.17, 2.46)** |
Table
4 presents the associations between parental and adolescent factors and adolescent and young adult vaping across time (W1 to W2). Parental smoking or vaping (in both the unadjusted and adjusted models) was associated with significantly increased odds of W1 only and new onset (W2) vaping. Household substance use was associated with increased odds of W1 only vaping (unadjusted model only) and both W1 and W2 vaping (both unadjusted and adjusted models). Emotional neglect, foster care or contact with CPO, and unsafe neighbourhood were associated with significantly increased odds of new onset (W2) vaping in both the unadjusted and adjusted models. Emotional abuse, exposure to verbal IPV, and household mental illness were associated with significantly increased odds of new onset (W2) vaping and both W1 and W2 vaping (in both unadjusted and adjusted models). Adolescent mental disorders (unadjusted model only), parental separation or divorce (unadjusted model only), parental gambling (both unadjusted and adjusted models), and peer victimization (both unadjusted and adjusted models) were associated with significantly increased odds of both W1 and W2 vaping.
Table 4
Parental and adolescent risk factors associated with course of vaping across time (Wave 1 to Wave 2)
Parental smoking or vaping |
OR (95% CI) | 6.35 (2.41, 16.70)*** | 1.93 (1.12, 3.32)* | 1.29 (0.72, 2.30) |
AOR (95% CI) | 5.31 (1.81, 15.53)** | 1.86 (1.05, 3.30)* | 1.27 (0.68, 2.37) |
Adolescent sex (Female vs. Male) |
OR (95% CI) | 0.65 (0.26, 1.65) | 1.11 (0.75, 1.65) | 1.04 (0.71, 1.51) |
AOR (95% CI) | 0.56 (0.21, 1.48) | 1.10 (0.74, 1.65) | 0.91 (0.62, 1.36) |
Adolescent mental disorders |
OR (95% CI) | 2.22 (0.84, 5.83) | 1.06 (0.65, 1.74) | 1.63 (1.05, 2.53)* |
AOR (95% CI) | 1.69 (0.60, 4.73) | 1.11 (0.67, 1.84) | 1.40 (0.88, 2.22) |
Emotional abuse |
OR (95% CI) | 1.10 (0.31, 3.93) | 1.87 (1.16, 3.00)** | 2.09 (1.32, 3.31)** |
AOR (95% CI) | 1.11 (0.30, 4.12) | 1.92 (1.19, 3.12)** | 2.06 (1.28, 3.31)** |
Emotional neglect |
OR (95% CI) | 2.93 (0.62, 13.87) | 2.70 (1.31, 5.58)** | 1.45 (0.63, 3.32) |
AOR (95% CI) | 2.19 (0.42, 11.41) | 2.64 (1.25, 5.54)* | 1.29 (0.55, 3.07) |
Exposure to verbal IPV |
OR (95% CI) | 1.35 (0.43, 4.24) | 1.82 (1.13, 2.93)* | 2.07 (1.31, 3.26)** |
AOR (95% CI) | 1.30 (0.40, 4.19) | 1.87 (1.15, 3.04)* | 2.01 (1.25, 3.22)** |
Spanking |
OR (95% CI) | 0.45 (0.13, 1.59) | 0.98 (0.63, 1.51) | 1.09 (0.72, 1.66) |
AOR (95% CI) | 0.43 (0.12, 1.55) | 0.99 (0.63, 1.54) | 1.11 (0.72, 1.72) |
Household substance use |
OR (95% CI) | 3.17 (1.08, 9.37)* | 1.48 (0.82, 2.66) | 2.48 (1.48, 4.17)*** |
AOR (95% CI) | 2.33 (0.73, 7.42) | 1.45 (0.79, 2.64) | 2.44 (1.42, 4.19)*** |
Household mental illness |
OR (95% CI) | 2.48 (0.88, 7.02) | 1.65 (1.04, 2.59)* | 1.85 (1.21, 2.81)** |
AOR (95% CI) | 1.57 (0.50, 4.90) | 1.73 (1.07, 2.81)* | 1.63 (1.03, 2.56)* |
Parental separation/divorce |
OR (95% CI) | 1.62 (0.55, 4.81) | 1.48 (0.94, 2.32) | 1.71 (1.12, 2.63)* |
AOR (95% CI) | 0.25 (0.03, 1.92) | 1.63 (0.88, 3.01) | 1.77 (0.97, 3.25) |
Parental problems with police |
OR (95% CI) | 2.01 (0.43, 9.34) | 1.91 (0.97, 3.78) | 1.76 (0.89, 3.46) |
AOR (95% CI) | 1.45 (0.30, 7.08) | 1.87 (0.92, 3.78) | 1.78 (0.88, 3.62) |
Parental gambling |
OR (95% CI) | – | 1.23 (0.31, 4.81) | 3.87 (1.45, 10.38)** |
AOR (95% CI) | – | 0.84 (0.20, 3.62) | 3.18 (1.13, 8.90)* |
Foster care or contact with CPO |
OR (95% CI) | 1.31 (0.29, 5.92) | 2.50 (1.41, 4.42)** | 1.74 (0.96, 3.15) |
AOR (95% CI) | 0.65 (0.13, 3.28) | 2.67 (1.45, 4.93)*** | 1.62 (0.86, 3.06) |
Poverty |
OR (95% CI) | 2.13 (0.71, 6.35) | 1.02 (0.59, 1.77) | 0.83 (0.48, 1.42) |
AOR (95% CI) | 1.36 (0.41, 4.44) | 0.96 (0.53, 1.74) | 0.93 (0.51, 1.68) |
Unsafe neighbourhood |
OR (95% CI) | 4.78 (0.97, 23.61) | 2.67 (1.06, 6.70)* | 1.88 (0.70, 5.04) |
AOR (95% CI) | 2.99 (0.52, 17.21) | 2.62 (1.02, 6.75)* | 1.87 (0.67, 5.22) |
Peer victimization |
OR (95% CI) | 2.35 (0.89, 6.18) | 1.53 (0.96, 2.45) | 1.71 (1.10, 2.67)* |
AOR (95% CI) | 2.52 (0.92, 6.88) | 1.52 (0.94, 2.47) | 1.84 (1.15, 2.93)* |
Discussion
The current study identified several unique findings related to adolescent vaping and the risk factors associated with different trajectories. First, the sample prevalence of any vaping in the current study was 45.5% – higher than Canadian estimates of adolescent and young adult vaping in 2017 (23 and 29%, respectively) [
32]. Differing populations, years of data collection, and study design (i.e., cross sectional vs. longitudinal) may, in part, explain the discrepancy. The sample in the current study is not representative so the findings may not be generalizable to other populations. Second, the vast majority (88.6%) of adolescents who had vaped at W1 (
n = 168) continued to do so as older adolescents or young adults at W2. Third, the majority of ACEs and risk factors studied were associated with any adolescent or young adult vaping, including: parental smoking or vaping, emotional abuse, emotional neglect, exposure to verbal IPV, household substance use, household mental illness, parental separation/divorce (unadjusted model only), parental problems with police, foster care or contact with a CPO, an unsafe neighbourhood, and peer victimization. Fourth, the majority of these ACEs and risk factors (associated with any adolescent or young adult vaping), and the addition of adolescent mental health and parental gambling, were associated with different courses of vaping across the two time points (some in the unadjusted models only).
The findings of the current study are generally consistent with one or more published studies in the literature [
10,
14,
15,
22,
23,
25,
27,
33], with the exception of foster care or contact with a CPO, an unsafe neighbourhood, and peer victimization, which have not been found to be associated with vaping outcomes in a previous study using W1 of the WE Study data [
27]. Importantly, the current study extends knowledge by examining the associations with adolescent and young adult courses of vaping over two time points using longitudinal data. Some of the mixed findings in the literature, particularly as they relate to ACEs, along with our findings that ACEs and other risk factors are associated with different courses of vaping over time, emphasize the importance of taking a more nuanced approach with greater specificity when examining risk factors associated with vaping among adolescents and young adults. For example, in our findings, adolescent mental health problems and parental gambling were not significantly associated with the ‘any vaping’ variable; however, these risk factors were significantly associated with vaping at both W1 and W2 (mental health problems in the unadjusted model only).
The finding that the vast majority of adolescents who reported vaping at W1 also reported vaping at W2 highlights the importance of early prevention of vaping initiation. The factors associated with vaping at both waves identified in the current study may help inform prevention strategies and identify target population groups. These ACEs and other risk factors included adolescent mental disorders (unadjusted model only), emotional abuse, exposure to verbal IPV, household substance use, household mental illness, parental separation/divorce (unadjusted model only), parental gambling, and peer victimization. Although parental problems with police, foster care or contact with CPO, and an unsafe neighbourhood were not significantly associated with vaping at both waves, each association had moderate effect sizes, similar to other factors that were studied. It is possible that the lack of statistical significance is a type II error due to rare occurrences of these events and power issues due to small sample size. Future studies with larger sample sizes should examine these relationships further. The findings of the current study suggest that any of these significant risk factors could be targeted to inform prevention strategies and identify target populations at various points in adolescents and young adults’ vaping trajectories. However, these findings do not provide a clear understanding of which adolescents quit vaping after W1. More research is needed to understand protective factors that may be associated with vaping cessation as well as prevention.
This study has several limitations. First, causation cannot be inferred; the associations with the “any vaping” variable are cross sectional and, among those who reported vaping in the past 30 days in W1, it is unknown when initiation first occurred and whether it may have preceded the factors studied. Second, the adolescent and young adult vaping questions changed slightly from W1 to W2. In W1, respondents were asked about past 30-day use only and the question did not specify for nicotine use while in W2 they were asked specifically about vaping for nicotine use for lifetime, 12 months, and past 30 days. The difference in wording may result in an overestimate of the number of adolescents vaping nicotine in W1 as it may include those who used electronic vaping products for other substances (e.g., nicotine free or marijuana). However, this is likely a very small proportion of respondents. In one Canadian survey of 3034 regular e-cigarette users aged 16 to 24 years, the vast majority of respondents (91.3%) indicated using electronic vaping products containing nicotine and only 1.2 to 3.8% were currently vaping cannabis products [
34]. Furthermore, the difference in timeframes in the current study between waves may have also influenced the results. For example, the number of respondents who used an electronic vaping product in W1 may be underestimated due to the shorter timeframe (past 30 days) compared to W2 (past 12 months and/or past 30 days). When parents were asked about electronic vaping product use in W1, the question also did not specify for nicotine use. Third, some of the measures may be subject to recall bias. Fourth, due to mandatory child abuse reporting laws, we were unable to ask adolescents about their experiences of the following ACEs: physical abuse, sexual abuse, and physical neglect. Fifth, the use of single item questions to measure many of the ACEs in the current study, while common in the literature, has been critiqued as simplistic [
35]. Sixth, attrition from W1 to W2 due to the longitudinal nature of the study may have biased the results. Finally, the community sample recruited primarily by non-random recruitment methods may not be representative and therefore the findings may not be generalizable to other populations with different household incomes, racial and/or ethnic profiles, and experiences. However, it should be noted that the W1 sample closely resembled the general population regarding sex, household income, and ethnicity, using the 2017 Statistics Canada census profile [
36].
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.