Risk Factors for E-cigarette Ban Relaxation in Homes with Adolescents

Abstract

Purpose:

To identify risk factors for relaxing a strict e-cigarette ban in households with adolescents.

Methods:

Youth (ages 12–17) in the Population Assessment of Tobacco and Health (PATH) Study whose parents held a strict vaping ban in 2016 (n=6,514; 51.5% male) and their parents provided follow-up data on up to 4 occasions through 2020 on whether the ban was relaxed.

Results:

13.5% of households with strict vaping bans relaxed them in a subsequent wave. Results from a logistic regression model showed that the odds of relaxing strict bans were higher if, at baseline, parents vaped (OR=2.20; 95% CI: 1.22–3.97, p<.01), parents smoked tobacco (OR=2.55; CI: 2.00–3.26, p<.001), youth smoked tobacco (OR=2.27; CI: 1.29–4.00, p<.01), parents reported no smoking ban (OR=2.68; CI: 1.88–3.81, p<.001), youth did not know that their household had a vaping ban (OR=1.95; CI: 1.50–2.54, p<.001), and parents perceived low harm from vaping (OR=1.60; CI: 1.16–2.19, p<.01). Although most sociodemographic controls were not independently associated, parents were less likely to relax bans if they had a college degree (OR=0.71; CI: 0.51–0.998, p<.05), graduate degree (OR=0.50; CI: 0.43–0.72, p<.001), or children who were non-Hispanic Black (OR=0.69, CI: 0.49–0.96, p<.05) or Hispanic (OR=0.62, CI: 0.47–0.81, p<.001).

Discussion:

While most households with adolescents prohibited e-cigarette use indoors, nearly 1 in 7 relaxed prior strict vaping bans. Parents need support to maintain clear, consistent, and continuous restrictions that communicate that vaping is not safe or permissible for youth.

Implications and Contribution:

Risk factors for relaxing strict bans on e-cigarette use in households with adolescents include parent and youth nicotine use, youth being unaware of the ban, and parents perceiving vaping as less harmful. Although most US households hold vaping bans, parents need strategies and support to maintain them consistently.

Introduction

Electronic cigarette (e-cigarette) use has surpassed use of smoked and smokeless tobacco among youth (1). In 2022, 9% of 8th graders, 17% of 10th graders, and 26% of 12th graders reported past 30-day e-cigarette use (i.e., vaping) (2). These trends threaten dramatic public health gains toward less nicotine dependence (3). Notwithstanding the potential value of e-cigarettes for cessation or reduction of adult tobacco use (4), myriad concerns remain about risks of e-cigarettes for young people, including nicotine dependence, transitions to cigarette smoking, carcinogen exposure, changes in brain development, respiratory symptoms, and second- and third-hand vapor exposure (5, 6). Given the public health impact of youth e-cigarette use, increasing attention focuses on how parents can reduce children’s nicotine exposure via e-cigarettes (7).

Developmental theories and research on family influences suggest that parents are important influences of children’s values and lifestyles (8), including adolescents’ attitudes and behavior concerning substances like tobacco. Mechanisms by which parents shape teen tobacco use range from interpersonal relationship characteristics and parenting strategies to domain-specific modeling and antismoking socialization (9). Specifically, parents can reduce children’s risk of tobacco initiation by not smoking (10) and through smoking-specific parenting strategies to discourage smoking, such as restricting tobacco access, setting rules, and establishing clear tobacco norms (11).

Household tobacco smoking bans represent a promising deterrent of youth tobacco use (11, 12). Research on smoking bans can inform emerging work on similar household e-cigarette bans. Preventing smoking within indoor spaces reduces smoke exposure for smokers and non-smokers alike, directly benefitting health (13, 14). Banning smoking at home can encourage adult smokers to reduce or quit smoking (15). Further, smoking bans reflect and communicate family anti-smoking values and norms to children more tangibly than smoking-specific communication alone (11, 16). While more common in homes without smokers and homes with young children, the prevalence of household smoking bans has continued to increase since the early 1990s (17–20). In fact, 9 in 10 U.S. adults reported household smoking bans in 2018/2019 (19).

With the advent of e-cigarettes, many household bans now extend to cover vaping. In 2015/2016, almost 4 in 5 U.S. households endorsed having a ban prohibiting e-cigarette use by any person inside the home (21). While vaping bans remain less common than smoking bans (17, 21), household vaping bans may be particularly important for preventing or delaying adolescent e-cigarette use, as parents often have less awareness of their children’s e-cigarette versus smoked tobacco use (16). Household vaping bans communicate parents’ disapproval of e-cigarette use and increase youths’ perceptions of vaping harmfulness (22). Empirically, recent studies have identified that strict household vaping bans may be effective deterrents of regular and past-month e-cigarette use among youth (22, 23). Given that nicotine use often begins during adolescence and can rapidly progress to dependence (3, 24), household vaping bans may be a promising parental prevention strategy (17, 22, 23).

Most homes prohibit smoking or vaping (21), and awareness among US adults is growing about the harms of vaping (25). In this context, it is puzzling why some parents would relax or eliminate a strict ban of e-cigarette use in the home. Yet in the national Population Assessment of Tobacco and Health (PATH) Study, almost 1 in 11 households who previously prohibited all tobacco use in the home reported allowing some tobacco or nicotine use (smoked, smokeless, or e-cigarette use) a year later (13). The present study extends such research by (a) focusing specifically on e-cigarette bans, and (b) identifying parent and adolescent risk factors for relaxing home vaping bans.

Taking a developmental epidemiological approach, we focus on proximal risk factors that may be observable to practitioners and potentially malleable by parents during the critical period of adolescence (26). Empirical studies have identified various factors related to whether a household implements a home vaping ban. These include parental tobacco use and vaping (13, 23, 27, 28), parental perceptions of vaping harmfulness (22), youth already smoking or vaping (23), and socioeconomic status (13, 27). Yet, it is unknown whether these predictors of not implementing strict household vaping bans are also related to relaxing home vaping bans over time. Although most households report a ban at any given time (21), ban effectiveness likely requires consistent implementation (12). As such, the current study sought to identify risk factors for families relaxing strict household vaping bans. We hypothesized that parental smoking or vaping (27, 28), child smoking or vaping (23), not having a smoking ban (13), youth not concurring with the vaping ban (16), less structure as indexed by child curfews (29), and parents not seeing vaping as harmful (22) would predict increased risk of relaxing home vaping bans. Sociodemographic factors were also included. If health providers and public health messaging can provide better targeted support to families at the greatest risk of dropping bans, parents may be more likely to maintain a potentially protective household norm and thereby delay or deter youth nicotine vaping. Thus, our primary research question is: What factors are associated with parental relaxation of their prior strict household ban on e-cigarettes?

Method

Data

In the PATH study’s stratified probability sample of U.S. households, adolescents aged 12 to 17 and their parents completed computer-assisted interviews in their primary residences in 2013–14 plus longitudinal follow-ups through 2020 (30, 31). Parents provided informed consent and youth provided assent. Children turning 12 in participating households were added annually; youth data through age 17 are used here. The present secondary data analyses were deemed non-human-subjects research by the Pennsylvania State University Institution Review Board. Deidentified public-use data were utilized (31).

Parents were first asked about household e-cigarette bans in Wave 3 (2016), and then again in Waves 4 (2017), 4.5 (2018), 5 (2019), and 5.5 (2020). To identify risk factors for relaxing strict household e-cigarette bans, of the 11,547 families with youth aged 12–17 participating in Wave 3, we excluded (a) 2,023 families in which parents did not report a strict vaping ban in Wave 3 (defined below); (b) 2,072 families lacking follow-up data from parents on household vaping bans in Waves 4–5.5 (due to youth turning 18, parent item-missing data, or attrition); and (c) 938 families missing covariate data. The analytic sample was 6,514 families.

Measures

Outcome (Waves 4 to 5.5)

Strict Vaping Ban Relaxed.

Starting in Wave 3, parents reported household rules for “everyone who might be in your home, including children, adults, visitors, guests, or workers”. Response options for rules about using e-cigarettes or other electronic nicotine products inside the home were a) “not allowed anywhere or at any time,” b) “allowed in some places or at some times,” or c) “allowed anywhere and at any time.” To index ban relaxation, we identified parents who reported a strict vaping ban at Wave 3 (a) but more relaxed rules (b or c) in any subsequent wave (Wave 4, 4.5, 5, 5.5). A binary indicator contrasted households that maintained a strict vape ban (0) with households that relaxed or dropped a prior ban (1).

Predictors (Wave 3)

Parent Nicotine Use, Rules, and Attitudes.

Parents were asked, “In the past 30 days, have you used any of the following products, even once or twice?”. Parental vaping reflected any use of e-cigarettes or other electronic nicotine products (1). Parental smoking reflected any use of cigarettes, traditional cigars, cigarillos, filtered cigars, or pipes (1). Parents reported their household rules for smoking “tobacco products that are burned” inside their home. A binary measure indicated no strict ban on tobacco smoking if participants endorsed combustible tobacco products were “allowed in some places or at some times” or “allowed anywhere and at any time” in the home (1) versus “not allowed anywhere or at any time” (0). Parents’ perceived harm of vaping was assessed by whether parents “think people harm themselves” when they use electronic nicotine products. A binary measure contrasted “no harm” or “little harm” (1) versus “some harm” or “a lot of harm” (0).

Youth Nicotine Use and Rules.

Youth indicated the last time they either “used an electronic nicotine product, even one or two times” or “smoked a cigarette, even if only one or two puffs”. Vaping and smoking were defined as use in the past month (1) versus all other options (0). Youth agreement about e-cigarette rules was based on a question about, “…the rules about using e-cigarettes and other electronic nicotine products inside your home?” Response options were a) “not allowed anywhere or at any time,” (i.e., strict rules); b) “allowed in some places or at some times,” (i.e., inconsistent rules); or c) “allowed anywhere or at any time” (i.e., no rules). Inconsistent or no rules (1) were contrasted with clear, strict rules (0). Non-nicotine-specific parental strictness was assessed via parent report of whether their child had “a curfew or set time that [he/she] needs to be home” on weekday and on weekend nights. A binary measure of youth curfew contrasted youth with curfews on both (0) versus youth with only one or no curfew (1).

Sociodemographic Indicators.

Parents reported past 12-month total household income, coded to Less than $10,000, $10,000-$24,999, $25,000-$49,999, $50,000- $99,999, and $100,000 or more (reference). The highest level was contrasted with other levels and with “Don’t Know” and “Refused to Answer” to retain the 4.4% not providing income data.

Indexing education, parents indicated the “highest grade or year of school completed” for themselves and a spouse or guardian in the household. Eleven response categories were collapsed for the parent with greater qualifications into Less than high school, High school diploma or equivalent (reference), Some college, College degree, and Graduate degree.

Marital status contrasted currently married (1) versus all others (0).

Renting (coded 1) versus owning the home (0) was included to account for external restrictions on smoking or vaping set or enforced by landlords or other tenants (32).

Shared custody was reported by parents. A binary indicator contrasted youth staying with another parent who lives somewhere else “less than half the time” or more (1) with never staying elsewhere (0).

Parent role. PATH Study data on the parent’s relationship with the child were used to create two dummy variables contrasting the biological mother (reference) versus the biological father and any other guardian or relation.

Youth reports of age, sex, and racial/ethnic identity in the public use files are coarsened to protect deductive identification. Age contrasts youth aged 15 to 17 (1) versus 12 to 14 (0) at Wave 3. Youth sex was coded as female (0) versus male (1). Youth race/ethnicity was coded as: “Non-Hispanic White,” (used as reference), “Non-Hispanic Black,” “Hispanic,” and “Non-Hispanic Other.”

Plan of Analysis

A multivariable logistic regression model, inclusive of all factors, estimated the likelihood of relaxing a strict vaping ban after Wave 3. All analyses were adjusted with Wave 3 PATH Study single-wave and replicate weights in Stata 18. The combination of these weights is recommended by the PATH Study to account for selection, differential nonresponse rates, and sampling frame. The “svy, brr” option with Fay’s adjustment method is used, as recommended for additional stability in variance estimation. Predictors were the parent and youth risk factors along with the sociodemographic indicators. Supplemental sensitivity analyses examined whether results were similar when (a) data collected in 2020 were excluded, (b) data after Wave 4 were excluded, to compare predictors of dropping bans quickly versus later, and (c) multiply imputed data rather than the PATH Study replicate weights were used.

Results

Table 1 presents the weighted prevalence of Wave 3 risk factors and results of a logistic regression predicting parents reporting relaxing their household vaping ban. Of the initial sample reporting a strict household vaping ban at Wave 3 (N=6,514), a total of 13.5% later relaxed that ban (7.4% in Wave 4, 3.2% in Wave 4.5, 2.3% in Wave 5, and 0.6% in Wave 5.5). Given sample selection criteria, parent and youth nicotine use was relatively low, with less than 1 in 8 parents reporting smoking tobacco and small percentages (<3%) of youth smoking and parents or youth vaping. More than 9 in 10 parents also reported a smoking ban and saw vaping as harmful. Finally, almost 1 in 10 youth did not know or report a home vaping ban.

Table 1.

Risk Factors for Relaxing or Dropping a Prior Strict Household Vaping Ban: Weighted Descriptive Statistics, Odds Ratios, and 95% Confidence Intervals from Multivariable Logistic Regression

	Weighted	Weighted Logistic Regression
	Descriptives	Predicting Relaxed Ban
	%	OR	95% CI
Outcome
Parent relaxed strict vaping ban, W4–W5.5	13.5	--	--
Parent and youth behavior and attitudes, W3
Nicotine use (past month)
Parent vapes e-cigarettes	1.4	2.20 **	[1.22, 3.97]
Parent smokes tobacco	12.2	2.55 ***	[2.00, 3.26]
Youth vapes e-cigarettes	2.4	1.02	[0.54, 1.94]
Youth smokes tobacco	1.6	2.27 **	[1.29, 4.00]
Household rules and norms
Parent: No household smoking ban	4.6	2.68 ***	[1.88, 3.81]
Youth: No household vaping ban	9.6	1.95 ***	[1.50, 2.54]
Parent: Youth has no curfew	17.5	1.21	[0.94, 1.56]
Parent: Little or no harm from vaping	6.8	1.60 **	[1.16, 2.19]
Sociodemographic predictors, W3
Parent role
Biological mother (ref)	68.6	--	--
Biological father	19.4	0.98	[0.77, 1.24]
Parent/guardian, other	12.1	0.99	[0.77, 1.26]
Parent is married (vs. not)	66.8	1.04	[0.81, 1.34]
Parent education
Less than high school	7.2	0.75	[0.53, 1.07]
High school diploma (ref)	14.5	--	--
Some college	29.1	0.93	[0.73, 1.19]
College degree	24.5	0.71 *	[0.51, 0.998]
Graduate degree	24.8	0.50 ***	[0.34, 0.72]
Household income
$100K or more (ref)	30.5	--	--
$50K–100K	26.3	0.97	[0.69, 1.35]
$25K–49K	19.6	1.47	[0.96, 2.26]
$10K–25K	13.1	1.33	[0.88, 2.00]
Less than $10K	6.1	1.50	[0.87, 2.56]
Don’t know	1.8	0.95	[0.43, 2.09]
Refused to answer	2.6	1.30	[0.68, 2.50]
Home is rented (vs. owned)	34.7	1.06	[0.83, 1.34]
Youth is male (vs. female)	51.5	1.09	[0.91, 1.31]
Youth lives part-time in another home	20.4	1.21	[0.92, 1.58]
Youth aged 15-17 (vs 12-14)	39.3	0.47 ***	[0.39, 0.57]
Youth race/ethnicity
Non-Hispanic white (ref)	53.1	--	--
Non-Hispanic Black	13.1	0.69 *	[0.49, 0.96]
Hispanic	23.1	0.62 ***	[0.47, 0.81]
Non-Hispanic other	10.7	0.83	[0.60, 1.14]

N=6,514.

^*P<.05.

^**P<.01.

^***P<.001.

Odds ratios are estimated in one multivariable logistic regression model.

W=Path Study Wave

Number of bans relaxed in each wave (out of 6,514 at W3: 7.4% (n=479) in W4; 3.2% (n=212) in W4.5; 2.3% (n=159) in W5; 0.6% (n=45) in W5.5

All values weighted using PATH Study Wave 3 single-wave and replicate weights to account for sample selection and differential nonresponse rates at Wave 3, PATH Study Public Use Files User Guide (Section 5.4).

Odds of relaxing household bans were 2.20 times higher if parents vaped e-cigarettes (p<.01), 2.55 times higher if parents smoked tobacco (p<.001), and 2.27 times higher if youth smoked tobacco (p<.01). Odds were 2.68 times higher if parents reported no tobacco smoking ban (p<.001), 1.95 times higher if youth did not report a strict vaping ban (p<.001), and 1.60 times higher if parents perceived little to no harm from vaping (p<.01). Neither prior youth vaping nor youth having curfews predicted ban relaxation. These associations were net of each other and of the links between sociodemographic indicators and relaxing vaping bans.

Many of the sociodemographic predictors were not independently associated with the outcome, but vaping bans were 29% and 50% less likely to be relaxed in households with parents with a college (p<.05) or graduate (p<.001) degree, respectively, compared to households with parents with a high school diploma; 53% less likely if adolescents were older at baseline (p<.001); and 31% less likely among non-Hispanic Black youth (p<.05) and 38% less likely among Hispanic youth (p<.001) versus non-Hispanic white youth.

Predicted values are presented in Figure 1 to illustrate the estimated percentage of relaxed household vaping bans for families with and without each of the six significant nicotine-related risk factors. All other predictors were set to the sample mean. In households where parents vaped, 44% relaxed their strict vaping ban compared to 13% where parents did not. Similarly, 35% of homes with parents who smoked relaxed their ban versus 11% of non-smoking parent homes. A similar ratio was observed distinguishing homes without (42%) versus with (12%) a smoking ban. If the youth reported smoking at Wave 3, 29% of households relaxed their ban compared to 13% of homes where the youth did not. Finally, 23% of parents who viewed vaping as causing little or no harm relaxed their ban, compared to 13% of parents who viewed vaping as risking some or a lot of harm.

Figure 1.

Predicted percent of households who drop household vaping ban by six nicotine-related risk factors

Alternative Specifications

The findings were similar in models that: a) excluded Waves 4.5 to 5.5 (to assess whether households that relaxed bans in Wave 4 were different from those who relaxed rules later); b) excluded just Wave 5.5 (to account for COVID-related changes to home life); and c) used imputed data (to retain missing covariate data but did not permit the use of PATH replicate weights). Findings also supported the coding of a binary outcome since the focal predictor variables did not distinguish households where the ban was dropped and reinstated (6.4%) versus dropped in all observed waves (7.1%). These supplementary analyses thus support the main findings.

Discussion

Reductions in nicotine dependence over the past several decades are threatened by growing adolescent e-cigarette use (1), facilitated by aggressive marketing of e-cigarettes as healthy, trendy, and concealable (33). Parents, however, can delay and possibly deter adolescent nicotine uptake with consistent norms and home rules (11). Notably, most parents with strict vaping bans maintained them over time. However, almost 1 in 7 households relaxed their strict policy against e-cigarette use in the home, with most of these households doing so within one to two years.

Six behavioral and attitudinal markers of risk for relaxing bans were identified, beyond associations with parent and youth sociodemographic factors. Relaxing vaping restrictions was more likely among parents who vaped, parents who smoked, and households without smoking bans, consistent with research on the presence and effectiveness of household tobacco smoking bans (9, 11, 12, 18, 34). In addition, the few households in which parents viewed vaping as unharmful, youth smoked, or youth reported no consistent vaping ban were more likely to report no strict vaping ban at a later wave. An exception was that families of the 2.4% of youth who vaped at Wave 3 were not more or less likely to relax their vaping ban over time. We speculate that some of these parents may be unaware of their youth’s vaping (16), creating no conflict or pressure within the household to relax the family vaping rule.

Apart from this exception, nicotine behaviors and norms among parents and youth that are known to predict not having a vaping ban can also be seen as risk factors for relaxing a ban once established (13, 22, 23, 27, 28). We suggest it is harder to sustain a practice of not vaping in the home in families with members, friends, or visitors who vape elsewhere. Conversely, it is no great feat to maintain a ban if family members or guests do not even tentatively ask if vaping would be acceptable. Thus, no critique is implied of families with people who vape, as these households often face difficult challenges in reducing nicotine exposure in the home (35). However, as illustrated in Figure 1, families with established vaping bans who have these nicotine-specific behavioral and attitudinal risk factors are at greater risk of relaxing them. Drehmer et al. (2019) found that a minority of parents who smoke and/or vape report being asked about their household practices by pediatric medical providers. As such, we concur that well-child check-ups with medical providers are a significant and underutilized opportunity to support families by reinforcing the value of a vape-free home (27).

Overall, relatively few of the demographics were independently predictive of relaxing bans. Parental education at the college degree level or higher was protective against relaxing strict vaping bans, consistent with household smoking ban research (19, 27, 34, 36) and well-established education gradients in health behaviors (37). Similarly, households of non-Hispanic white youth were more likely to relax bans than households of non-Hispanic Black and Hispanic youth, aligning with non-Hispanic White youth being less likely to have a vaping ban in the first place (23). Numerous other sociodemographic factors did not uniquely predict household ban relaxation.

Study strengths include the large nationally representative, longitudinal sample of parents and youth and the unique focus on understanding changes in household vaping bans as a possible malleable risk factor for lifelong nicotine dependence. A limitation is the focus on static predictors of subsequent ban relaxation. Future work using multiple methods should examine dynamic predictors of ban changes as well as households consistently lacking e-cigarette bans to understand emerging drivers of ban establishment, maintenance, and relaxation. We note that broader public opinion and social norms around e-cigarette use, not studied here, will likely also be important contributors to the adoption of vaping bans, as they were for indoor tobacco use (18). Whether sociodemographic factors may moderate observed links is also an important direction for future research. Finally, readers are reminded that the period covered was one of increasing public knowledge about e-cigarettes. In the U.S., this included the EVALI scare, Tobacco 21 laws, and restrictions on flavors (38).

Important questions and future directions following from the risk factors studied here include how to correct societal misinformation about vaping’s harmlessness (28, 39); socially acceptable strategies parents can use to maintain nicotine-free rules and norms for children, family, and friends; and effective strategies to support cessation of nicotine use broadly.

Conclusions

As innovative and highly marketed nicotine-delivery modes proliferate, parents need effective tactics to protect their children’s health. Similar to tobacco, household vaping bans may help prevent nicotine exposure and uptake among youth. Medical providers, teachers, religious and community leaders, and public health messaging can support parents to maintain clear, consistent, and continuous home restrictions to communicate that vaping is not safe or permissible for youth.