Age at Initiation of Cigarette Use in a Nationally Representative Sample of US Youth, 2013-2017

This cohort study uses data from the Population Assessment of Tobacco and Health (PATH) study waves 1 to 4 to assess at what age young people aged 12 to 17 years initiate cigarette use.


Introduction
Identifying the age of initiation of tobacco use has been pivotal to achieving reductions in tobacco use nationwide, as tobacco use remains one of the leading causes of preventable diseases and death in the US. 1,2 According to the 2012 Surgeon General's report, 88.2% of adult daily smokers aged 30 to 39 years recall initiating cigarette use at or before age 18 years. 3 Despite laws in place enforcing the minimum age of tobacco sales (18 years until December 2019; 21 years since December 2019), it is evident that youth are able to access tobacco products, including cigarettes.
Susceptibility to cigarette smoking has been identified as a risk factor for smoking initiation before youth start using cigarettes. 4,5 A previous publication of the Population Assessment of Tobacco and Health (PATH) study, conducted from September 12, 2013, to December 14, 2014, reported the prevalence of susceptibility to cigarette use among youth (aged 12-17 years) overall (28.6%) 6 and by race/ethnicity (for Hispanic youth, 31.3%; for non-Hispanic Black youth, 30.5%; and for non-Hispanic White youth, 26.9%). 6 The 2019 National Youth Tobacco Survey (NYTS) reported cigarette susceptibility by sex (boys, 46.4%; girls, 45.5%). 7 In addition, previous reports of PATH among youth have reported initiation of ever (3.8%) 8 and past 30-day (1.6%) 8 cigarette use after 1 or 2 years of follow-up, 9 as well as differences by sex and race/ethnicity. Middle and high school students from the 2014-2016 NYTS reported their median recalled age of cigarette initiation as 13 years, but this finding is prone to recall bias. [10][11][12] As the popularity of other tobacco products has been increasing among youth in recent years, 13 nearly all cigarette users initiate before age 18 years, 3 and earlier ages of cigarette use are associated with increased nicotine dependence 3 and higher risk for chronic diseases. 14,15 Therefore, it is necessary to examine the age of cigarette initiation for tobacco control.
Information that has been missing from the previously mentioned studies is the age of cigarette initiation overall, by sex, and by race/ethnicity, estimated prospectively using survival analyses with 4 years of follow-up. Therefore, we conducted a secondary analysis of PATH 16 prospectively estimating the age of initiation of cigarettes in youth (aged 12-17 years) for (1) susceptibility to cigarette use, (2) ever use, (3) past 30-day use, and (4) fairly regular use among nonsusceptible (1) and never users of cigarettes at their first wave of PATH participation. Fairly regular use is a subjective measure used to identify consistent, committed cigarette use. 17

Study Design and Participants
PATH is a nationally representative, longitudinal cohort study of US youth and adults that studies tobacco use behaviors, attitudes and beliefs, and tobacco-related health outcomes. 16 The target population of PATH consisted of individuals 12 years and older across the US, and 13 651 youth (aged 12-17 years) completed wave 1 (September 12, 2013, to December 14, 2014). Our study includes 2 subpopulations: (1) youth aged 12 to 17 years who were nonsusceptible to cigarette use at their first wave of PATH participation (waves 1-3) and (2) youth aged 12 to 17 years who were never cigarette users at their first wave (waves 1-3) of PATH participation. In addition, youth aged 9 to 11 years at wave 1 were considered "aged-up youth" when they turned 12 years old and were eligible to participate in the study at waves 2 to 3; 2091 and 2045 aged-up youth were included in our study at waves 2 (October 23, 2014, to October 30, 2015) and 3 (October 19, 2015, to October 23, 2016), respectively. Youth who turned 18 years of age were invited to complete the adult measurements; 1915, 1907, and 1900 aged-up adults completed the adult questionnaire from waves 2 to 4 (October 23, 2014, to January 3, 2018), respectively. Data from waves 2 to 4 were used to track outcomes for all participants. Informed oral consent was obtained from the parents of the youth participants, and the youth provided oral assent. 16 The University of Texas Health Center at Houston granted institutional review board approval (HSC-SPH-17-0368). This report followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies.

Measures Cigarette Outcomes
The following questions were used to measure susceptibility to cigarette use among nonusers of cigarettes across all 4 waves (2013-2017): (1) "Have you ever been curious about smoking a cigarette?" 18 (2) "Do you think that you will try a cigarette soon?" 18 and (3) "If one of your best friends were to offer you a cigarette, would you smoke it?" 18 Response options for the first question were "very curious," "somewhat curious," "a little curious," and "not at all curious." 18 Response options for the next 2 questions were "definitely yes," "probably yes," "probably not," and "definitely not." 18 Respondents who answered "not at all curious" to the first question and "definitely not" to the next 2 questions were considered nonsusceptible to cigarette use, and any other combination of responses was considered susceptible to cigarette use.
The PATH study used a derived variable to represent ever cigarette use (yes or no) at each wave. 18 Past 30-day cigarette use was measured with the question "In the past 30 days, on how many days did you smoke cigarettes?" 18 Numeric response options included 0 to 30 days, and participants were considered past 30-day cigarette users if they reported 1 or more days. Fairly regular cigarette use (yes or no) was measured with the question "Have you ever smoked cigarettes fairly regularly?" 18 For all questions, participants who responded "don't know" or "refused" were excluded from the analysis.

Sex and Race/Ethnicity as Exposures
The PATH study imputed the self-reported participant sex variable by using the household information at wave 1 but not at waves 2 and 3 18 and categorized participants as boys and girls. 18 The following categories measured self-reported participant race: White alone, Black alone, Asian alone, other (including multiracial). Participants' ethnicity was categorized as either Hispanic or non-Hispanic. 18 In order to be comparable to the Surgeon General's report 3 and previous publications, 9,10,19 investigators combined race and ethnicity to create 4 categories: non-Hispanic White, Hispanic, non-Hispanic Black, and non-Hispanic other (including multiracial participants).

Age of Initiation
The PATH study provided a derived variable for participant age in years at each wave, because participant date of birth is not included in the restricted-use data. 18 The PATH study provided another derived variable to represent the number of weeks between waves that youth participated in. 18 Participant age was converted from years to weeks and added to the number of weeks between survey dates to give us a more precise estimate of participant age. Lower and upper age bounds were created to identify the age at the last wave when the participant reported nonsusceptibility to cigarette use and the age at the first wave when they reported susceptibility to cigarette use. For participants who did not report initiation of susceptibility to cigarette use over the follow-up period (2014-2017), their upper bound was considered censored. For ever use, past 30-day use, and fairly regular use outcomes, the age of initiation was estimated in the same way as susceptibility based on when the outcome was first reported (2014-2017) for those who become users and the last report of nonuse for those who do not.

JAMA Network Open | Public Health
Age at Initiation of Cigarette Use in a Nationally Representative Sample of US Youth, 2013-2017

Statistical Analysis
Secondary analyses of the PATH study restricted-use data sets were conducted 18 from October 7, 2019, to May 1, 2020, using SAS, version 9.4 (SAS Institute). 20 A type I error level of 0.05 was used, and all P values were 2 sided. All data analyses incorporated sampling weights, 100 balance repeated replicate weights, and a Fay correction factor of 0.3 to account for the PATH complex study design. 18 Weighted summary statistics (means for continuous variables and proportions for categorical variables) are provided. Weighted interval-censoring survival analysis [21][22][23][24][25] was implemented to estimate the probability of the age of initiation of susceptibility to cigarette use, ever use, past 30-day use, and fairly regular cigarette use. Hazard functions for each cigarette use outcome overall are reported as cumulative percentages, and 95% CIs are reported using the Turnbull method. [21][22][23][24][25] Differences in the age of cigarette initiation for each outcome by sex and by race/ethnicity were estimated by fitting weighted Cox proportional hazards regression models to interval-censored data with a piecewise constant function as the baseline hazard function. [21][22][23][24][25] Hazard functions and 95% CIs are only reported by sex and by race/ethnicity for the outcomes that exhibited statistically significant differences. There was very little missingness in PATH, and the number of participants with missing sex or race/ethnicity are reported in Table 1.

Results
A total of 15 776 youth who were never cigarette users and, among them, 11 022 youth who were nonsusceptible to cigarette use, were included in the study. Among the 11 022 youth who were nonsusceptible to cigarette use, the weighted mean (SE) age was 13.5 (0.01) years, 69.6% (SE, 0.31%) entered the study at wave 1, 58.6% (SE, 0.46%) identified as non-Hispanic White, and 51.0% (SE, 0.32%) were boys (Table 1). Among the 15 776 who were never users of cigarettes at their first wave of PATH participation, the weighted mean (SE) age was 13.7 (0.01) years, 55.0% (SE, 0.29%) were non-Hispanic White, and 51.0% (SE, 0.15%) were boys (Table 1).

JAMA Network Open | Public Health
Age at Initiation of Cigarette Use in a Nationally Representative Sample of US Youth, 2013-2017

Discussion
To our knowledge, this study is the first to provide prospective estimates for the distribution of the  By age 18, 24.4% of youth initiated ever cigarette use in our study. This percentage was slightly higher than in the NYTS surveys, which indicated that among middle and high school students, 21% in the 2014 to 2016 period had ever used cigarettes. 10 These differences could be due to the longitudinal nature of our study, whereas the NYTS was cross-sectional across multiple years, or due to reporting incidence in our study vs prevalence in NYTS. The 2019 NYTS survey also reported finding significant differences between the prevalence of ever using cigarettes by sex, with 18.3% of boys reporting ever using cigarette compared with 14.2% of girls. 7 Our study goes beyond the NYTS    That study found that the proportion of ever cigarette initiation occurring in early adulthood more than doubled from 20.6% in 2002 to 42.6% in 2018. 29 Additionally, a recent study 30  years, which could be explained by the shifting of cigarette initiation to later ages. Although the age of cigarette initiation may be shifting toward early adulthood rather than adolescence, suggesting that cigarette interventions should be implemented among young adults, our study suggests that a high amount of cigarette initiation still occurs among youth. The age of initiation shifting toward later ages could be due to the introduction of e-cigarettes, 7 successful interventions targeted toward youth, [31][32][33] or other possible explanations. However, a recent article 34 shows that the introduction of e-cigarettes was followed by a slowing decline in past 30-day cigarette use and acceleration in the ever cigarette use trend between 2004 and 2018. In addition, another article 35 indicates that there was actually no decline in cigarette use because of the introduction of e-cigarettes. More research is needed to determine which of these explanations holds.
There is considerable evidence that earlier ages of cigarette initiation are associated with greater nicotine dependence and greater exposure to nicotine, 36 which can result in adverse health outcomes, including altering brain development among youth 37 and respiratory illnesses. 38 Previous research has established that nicotine-dependent youth are less likely to want to quit using cigarettes, and youth are uniquely vulnerable to developing nicotine addiction, even at low levels of nicotine use. 39 Interventionists and the public can use our results to identify the particular ages at which the campaigns may be most effective to implement to prevent youth from becoming susceptible to cigarette use and engaging in harmful cigarette use behaviors, such as past 30-day and fairly regular cigarette use.

Strengths and Limitations
Strengths of this study include using a nationally representative data set and the longitudinal prospective analysis of cigarette initiation outcomes across PATH waves 1 to 4. One of the limitations of our study is that we depended on self-reported data for the cigarette use outcomes to estimate the age of initiation because asking participants the exact date they initiated cigarettes is unrealistic.

Conclusions
To our knowledge, this cohort study is the first to provide prospective estimation of the age of initiation of susceptibility, ever, past 30-day, and fairly regular cigarette use among youth in the US between 2013 and 2017. This study provides windows of opportunity by identifying the ages of 4 cigarette use outcomes in which never users are most likely to initiate cigarette use, including by sex and by race/ethnicity. The results from our study suggest that among those who initiated cigarette use, the onset of the 4 cigarette outcomes occurred for the majority of youth before they were aged 18 years (between 16 and 18 years). Because the age of initiation of cigarette use was different by sex and by race/ethnicity, it is important to tailor preventive interventions considering those differences. Boys and non-Hispanic White youth had an increased risk of initiating cigarette use at earlier ages, suggesting that tobacco control interventions should focus on these vulnerable subgroups at earlier ages. In December 2019, the government changed the legal age to purchase tobacco products from 18 to 21 years. 40 Though the law did not affect our results, this study provides additional evidence to support the new regulation and can serve as a reference point for age of initiation of cigarette use in future studies after implementation of the Tobacco 21 campaign. 40