Parental Awareness of Youth Tobacco Use and the Role of Household Tobacco Rules in Use Prevention

The PATH Study is an ongoing household cohort study of tobacco use and health among US youth (aged 12–17) and adults.²¹  An area-probability, 4-stage stratified sampling design was implemented at wave 1 (September 2013 to December 2014) to represent the US noninstitutionalized civilian population. Parents (or guardians) were asked about their participating children separately from questionnaires administered to youth directly. The weighted wave 1 response rate for the youth survey was 78.4% among households screened for participation.²¹  Respondents were followed annually in waves (wave 2: October 2014 to October 2015; wave 3: October 2015 to October 2016; wave 4: December 2016 to January 2018). Respondents who reached age 18 before follow-up were invited to join the adult component. “Shadow youth,” aged 9 to 11 at wave 1, were invited to join the study at the wave at which they reached age 12. An additional “replenishment” youth sample was enrolled at wave 4.

PATH Study investigators obtained a National Institutes of Health certificate of confidentiality and ethical approval from the Westat Institutional Review Board. Parents or guardians provided informed consent for assenting youth. Youth received $25 for participation and parents or guardians received $10. In the present analysis, we used fully deidentified public use files.²² 

Each wave, youth responded separately about lifetime use (ever, even once or twice) and past 30-day use (≥1 day in the past 30 days) of multiple tobacco products. For this analysis, tobacco use was categorized as never use, former use (ever used ≥1 product but none in the past 30 days), and past 30-day use of only cigarettes, only e-cigarettes, only noncigarette combustible products (cigars, pipes, hookah, bidis, and/or kreteks), only smokeless tobacco (snuff, chewing tobacco, snus, and/or dissolvable tobacco), and polytobacco (use of products from ≥2 of these groups). The wave 2 questionnaire introduced the term “electronic nicotine products,” of which “e-cigarettes (including vape pens and personal vaporizers)” was a subset. In this analysis, we considered the most inclusive electronic nicotine product wording at each wave as e-cigarette use.

The first research question was used to assess parent or guardian knowledge or suspicion of their child’s tobacco use. Each wave, parents or guardians were asked, “As far as you know, has [child’s first name] ever smoked a cigarette or used other tobacco products, such as e-cigarettes, cigars, a pipe, a hookah, smokeless tobacco, dissolvable tobacco, bidis, or kreteks?” Parents were categorized as knowing or suspecting tobacco use if endorsing “you know that she/he has” or “you strongly suspect she/he has.” The responses “you don’t think she/he has” and “you are confident that she/he has not” were categorized as not aware or suspicious. “Don’t know” responses were uncommon (<0.2% each wave) and were coded as not aware or suspicious.

The second research question involved longitudinal analyses of youth tobacco use initiation. Parents or guardians and youth were independently asked to consider “rules about using tobacco inside your home” as applied to “everyone who might be in your home, including children, adults, visitors, guests, or workers.” Separate items referred to “tobacco products that are burned, such as cigarettes, cigars, pipes, or hookah” and “tobacco products that are not burned, like smokeless tobacco, dissolvable tobacco, and e-cigarettes.” Endorsing that product use “is not allowed anywhere or at any time inside my home” was classified as strict household rules, whereas endorsing “in some places or at some times,” “anywhere and at any time,” or “don’t know” was considered more permissive. Additionally, youth were asked, “In the past 12 months, have your parents or guardians talked with you, even once, about not using any type of tobacco product?” which we categorized as “yes” versus “no” or “don’t know.”

For both research questions, covariables included in multivariable models were parent or guardian educational attainment (4 levels: less than high school through college graduate); the child’s age (12–14 vs 15–17 years), sex, and race and/or ethnicity; whether anyone who now lives with the child uses tobacco (any product versus none, by child-report); and whether the child lives somewhere else with another parent (at least part-time versus never). Models of parental knowledge or suspicion additionally included whether the parent or guardian was the child’s biological mother. Models of tobacco initiation additionally included child ever use of alcohol and cannabis, whether the child has a curfew (both school nights and weekend nights), and the sensation-seeking score.²³ 

Of the total number of youth respondents in wave 1 (n = 13 651), wave 2 (n = 12 172), wave 3 (n = 11 814), and wave 4 (n = 14 798), the pseudo cross-sectional time-series analysis was limited to respondents with nonmissing data for parent knowledge or suspicion of youth tobacco use (wave 1: n = 13 600; wave 2: n = 12 129; wave 3: n = 11 807; wave 4: n = 14 701). In this pseudo time-series, each cross-sectional wave is weighted to be nationally representative, but some participants appear in multiple waves, resulting in n = 52 237 observations from n = 23 170 individuals. Longitudinal analyses were limited to youth who had never used any tobacco product as of wave 1. We assessed the outcome, initiation of tobacco use, at waves 2, 3, and 4, defined as reporting ever use of ≥1 tobacco product as of the time point of interest, including youth who “aged-up” into the adult survey by reaching age 18 at follow-up (sample sizes: wave 1 to 2, n = 8994; wave 1 to 3, n = 8340; wave 1 to 4: n = 7691).

For the pseudo time-series analysis, parental knowledge or suspicion of their child’s tobacco use was modeled as the dependent variable by using logistic regression with wave-specific, cross-sectional, balanced repeated replicate weights, allowing each wave to be nationally representative despite participant overlap. Youth tobacco use category was the main predictor variable, with tobacco use × survey wave interaction terms added to assess wave-specific differences in parental knowledge or suspicion according to youth tobacco use status.

For longitudinal analyses, youth tobacco initiation at waves 2, 3, and 4 were the dependent variables in separate logistic regression models with longitudinal weights. Wave 1 household tobacco rules and talking about not using tobacco were the main predictor variables. Household rules was specified as a 5-level categorical variable: both parent or guardian and child endorse more permissive rules on both product types, both endorse strict rules on burned tobacco (one or both more permissive for not-burned tobacco), both endorse strict rules on not-burned tobacco (one or both more permissive for burned tobacco), parent and youth discordant on both product types, and both endorse strict rules on both products types.

As exploratory analyses, we examined several alternative model specifications. We hypothesized that changes in social environments as youth age could reduce any impact of household rules on tobacco initiation over time. Therefore, we explored interactions of household rules with baseline age. Additionally, we explored whether household rules may differentially impact initiation of different types of tobacco by specifying multinomial logistic regression models for a 4-level dependent variable: no initiation, initiation of combustible tobacco only, initiation of noncombustible tobacco only, initiation of both product types. Finally, given that exposures and covariables are potentially time varying, we specified a repeated-measures model using generalized estimating equations for 1-year initiation outcomes, taking any observation of a youth tobacco never user in waves 1, 2, and 3.

For all models, missing covariable values (2.6% of covariable data) were multiply imputed (15 replications). Although missingness was uncommon for any one tobacco use variable (1.4% of tobacco data), missing tobacco responses were also imputed when examining parental knowledge or suspicion because missingness compounded when deriving categories: 7.7% of participant observations were missing ≥1 tobacco variable. Analyses were conducted by using Stata 16.0 (Stata Corp, College Station, TX). Results were considered statistically significant at P < .05.

Tobacco poly use was the most common behavior among current youth tobacco users in waves 1 to 3; in wave 4, e-cigarette only use was most common (Table 1). Among polytobacco users, 77% to 80% reported smoking cigarettes, depending on the survey wave. Social and demographic variables were similar in distribution over time (Table 1).

In all waves, parents or guardians were substantially less likely to report knowledge or suspicion that their children had used tobacco if their children reported use of only e-cigarettes, noncigarette combustible products, or smokeless tobacco compared with use of cigarettes or multiple tobacco products (Fig 1, Supplemental Table 3). In covariable-adjusted models, other factors associated with greater parent knowledge or suspicion included lower parent educational attainment; the child being older, being male, identifying as non-Hispanic white, living with a tobacco user, or residing in another home; and the parent respondent being the child’s mother (Supplemental Table 3).

Among wave 1 youth who reported never using tobacco, most parent-child pairs mutually endorsed having strict household rules that prohibited use of any burned tobacco (78%) and not-burned tobacco (70%). Half (50%) of youth reported that a parent or guardian had talked with them about not using tobacco within the past 12 months. There was high percentage agreement between parent and youth responses regarding household rules (Supplemental Table 4), although interrater reliability was constrained under a high marginal prevalence of strict rules (Cohen’s κ: burned, 0.52; not burned, 0.40, unweighted data). Household rules and talking about tobacco were uncorrelated (Supplemental Table 4).

Strict household tobacco rules were associated with less tobacco initiation (Table 2). Among wave 1 tobacco never users, 15% initiated use of ≥1 tobacco product by wave 2, 24% by wave 3, and 33% by wave 4. At all time points, children in households with the strictest rules prohibiting tobacco use had 20% to 26% lower odds of tobacco initiation compared with children in the most permissive households (adjusted for known risk factors) (Table 2). Households with strict rules only for burned or not-burned tobacco were also numerically associated with less initiation compared with the most permissive households, albeit not statistically significantly in these groups of smaller sample size. In contrast, youth who reported that their parent or guardian had talked with them about not using tobacco did not demonstrate lower odds of tobacco initiation; in fact, tobacco initiation was higher at waves 3 and 4 (Table 2). Other factors positively associated with tobacco initiation in all waves were the child being older, the child living with another tobacco user, the child residing at least part-time in another home, the child having used alcohol or cannabis, and greater sensation seeking (Table 2).

In exploratory analyses, although interaction by child age was not statistically significant overall, numerically, strict household rules were associated with lower odds of tobacco initiation among children who were younger at wave 1 (Supplemental Table 5). In multinomial models, strict household rules were associated with lower odds of noncombustible tobacco initiation at all 3 time points but not necessarily with lower odds of combustible tobacco initiation (Supplemental Table 6). However, initiation of only combustible tobacco was uncommon (eg, 1-year initiation: combustible only, 2.9%; noncombustible only, 7.1%; both, 5.0%), which yielded imprecise estimates. In the repeated-measures analysis, strict household rules remained associated with lower odds of tobacco initiation within a year (Supplemental Table 7).

In this assessment, we identified substantial lapses in parents’ awareness of their children’s tobacco use. Most parents or guardians registered suspicion when their children smoked cigarettes or reported polytobacco use (typically including cigarettes). Only approximately half as many knew or suspected when their children used only e-cigarettes or noncigarette combustible products. Of parental strategies to prevent future tobacco use by their children, setting strict household rules that prohibit all forms of tobacco use by anyone within the home was associated with less youth tobacco initiation, whereas talking with children about tobacco was not.

The percentage of parents aware or suspecting their children’s cigarette smoking (∼70%) was higher than in previous findings suggesting poor parental awareness of youth smoking (∼40%).^19,20  Greater awareness may be due to increasing social concern around youth smoking or survey measurement differences. However, low parental awareness of e-cigarette use belies rising public and media attention surrounding youth vaping. Constantly changing e-cigarette device designs and terminology pose an increasing challenge for parents to recognize, whereas lack of smoke and odor enhance concealability.¹⁸  Notably, PATH Study data were collected before a 2019 outbreak of vaping-associated lung injury,²⁴  which could heighten parental awareness going forward. Cigarette-smoking youth smoke more frequently than e-cigarette users use e-cigarettes, potentially increasing parental awareness opportunities. Lower awareness for cigars and hookah, which do produce smoke and odors, suggests a wider need for parents to monitor for all tobacco products, including those they may perceive as less common or concerning.

Findings related to tobacco-free households align with previous research revealing that home antismoking attitudes and rules contribute significantly to youth smoking prevention.^9,25,26  The present work suggests that this benefit extends beyond cigarettes to include initiation of any tobacco product use. Creating home tobacco-free environments offers the additional advantage of protecting children from harmful secondhand smoke exposure and may also benefit household adults by aiding smoking cessation.²⁷ 

Our results align with longitudinal findings revealing a benefit of household smoking bans, whether or not youth lived with smokers.²⁸  Our finding that the benefits of strict household rules appear greatest at younger ages suggests a need for additional focused prevention when adolescents transition to young adulthood and potentially enter new social environments. Unexpectedly, strict household tobacco rules were more strongly associated with prevention of noncombustible tobacco use than combustible tobacco use. This result must be interpreted cautiously because many youth initiate use of both product types, and strong concordance between burned and not-burned household tobacco rules makes it difficult to isolate independent effects. Nonetheless, setting household tobacco use rules may be a promising tool against the rise in youth e-cigarette use.

Contrary to rules governing tobacco use in the home, youth who reported that that their parent or guardian had talked with them about not using tobacco were at higher odds of initiating tobacco use after 2 or 3 years. An implication of this result is that telling children not to use tobacco does not benefit youth compared with setting norms and examples via tobacco-free rules that apply to everyone. Alternatively, it is possible that parents were more inclined to talk about tobacco with youth already at elevated risk of tobacco use on the basis of personality aspects not captured by study variables.¹⁵  In the current study, we did not measure the quality or frequency of parents’ antitobacco communication: likely key elements of effectiveness.¹³  Therefore, although strong household rules appears to be a much more promising approach, it should not necessarily be concluded that all parental communication is unhelpful in youth tobacco use prevention.

Advantageously, the current study features a large, prospective, nationally representative sample. To our knowledge, this analysis is the first to assess prospective outcomes of home tobacco use policies on youth initiation of cigarette and noncigarette tobacco use and the first national study to assess parental awareness of their children’s use of multiple noncigarette tobacco products. Numeric findings were robust to the length of follow-up and adjusted for an extensive suite of established youth tobacco use risk factors.

Among methodologic considerations and limitations, the main outcome measure in this study, initiation of any tobacco use, is a critical milestone, particularly among youth for whom daily smoking may develop over several years.²⁹  However, long-term established tobacco use deserves attention in future research. Speculatively, youth might report tobacco use differently with in-home (near parents) questionnaire administration versus in-school (near peers) questionnaire administration, with unclear impact on the present results. In our 2- and 3-year models, we did not account for the timing of tobacco initiation or time-varying exposures or covariables. However, in an 11-year longitudinal study in Italy, the authors reported that living with smoking family members in adolescence and absence of a household smoking ban in young adulthood were both associated with established smoking among young adults.³⁰  In the current study, findings were similar in a 1-year repeated-measures analysis that allowed for intraperson variation.

Results strongly support tobacco-free home environments as part of comprehensive tobacco control strategies. From a regulatory standpoint, clearly labeling emerging and novel products as tobacco products, along with greater public communication targeting parents about the appearance, forms, and possible harms of noncigarette tobacco, would position parents for better surveillance of their children’s tobacco use. In pediatric practice, raising parental awareness of noncigarette tobacco products should be part of tobacco anticipatory guidance and prevention support.³¹ 

e-cigarette

electronic cigarette

PATH

Population Assessment of Tobacco and Health