|
to cover the administration of an
<br />enforcement program and regular
<br />compliance checks in each store. An
<br />A grade also required (1) an annual
<br />renewal of this local license; (2) a
<br />provision that any violation of local,
<br />state, or federal law is a violation
<br />of the license; and (3) a graduated
<br />penalty system for violators,
<br />including financial deterrents such
<br />as fines or other penalties, including
<br />license revocation or suspension."
<br />The remaining study jurisdictions
<br />were assigned an F grade (8) or a
<br />D grade (1). An F grade indicated
<br />either (1) no local ordinance
<br />mandating a license fee or (2) a fee
<br />insufficient to fund administrative
<br />and compliance checks as well as
<br />none of the 3 other provisions for an
<br />A grade. The jurisdiction with the D
<br />grade had a licensing fee that was
<br />insufficient to cover administration
<br />and compliance checks, but ithad
<br />at least 1 of the other 3 provisions
<br />listed above that were needed for an
<br />A grade. The D and F communities
<br />were collapsed for data analysis,
<br />because the insufficient annual fee
<br />is a central feature of regulation to
<br />reduce youth access.7,11 No study
<br />jurisdiction in this sample had B or C
<br />grades corresponding to TRL policies
<br />of intermediate quality."
<br />ALA assigned grades to other
<br />categories of tobacco policy (smoke -
<br />free housing policy, smoke -free
<br />outdoor policy, and overall tobacco
<br />policy)." These policies, which are
<br />not specific to youth tobacco product
<br />access, were not associated with
<br />tobacco product use in this study, and
<br />results are not presented.
<br />Covariates
<br />Self-administered questionnaires
<br />completed by parents of
<br />participants were used to assess
<br />sociodemographic characteristics,
<br />including sex, ethnicity (Hispanic,
<br />non -Hispanic white, other), age at
<br />baseline, and parental education
<br />(completed high school or less, some
<br />college, or completed college or
<br />more).
<br />Statistical Analysis
<br />Unconditional logistic regression
<br />models were used to evaluate the
<br />associations of living in a jurisdiction
<br />with an ALA grade A versus D or F
<br />TRL ordinance with baseline ever
<br />and past 30-day use of cigarettes,
<br />e-cigarettes, hookah, cigars, or use
<br />of any of these tobacco products in
<br />separate models. Models were also fit
<br />to evaluate associations of ALA grade
<br />with the initiation of each product,
<br />with or without past 30-day use. In
<br />models used to evaluate the initiation
<br />of use of each tobacco product
<br />between baseline and follow-up, the
<br />sample was restricted to baseline
<br />never users of that product Odds
<br />ratios (ORs) and 95% confidence
<br />intervals (CIs) were used to estimate
<br />the association of each tobacco
<br />product use with an ALA grade.
<br />All models were adjusted for sex,
<br />ethnicity, highest parental education,
<br />and baseline age, factors that have
<br />been associated both with e-cigarette
<br />use and cigarette use in previous
<br />studies.13.14 Each tobacco product —
<br />specific model was also adjusted for
<br />a baseline history of use of any other
<br />tobacco product, because there was
<br />clustering of the tobacco product
<br />outcomes.13 A missing indicator
<br />category for covariates and any other
<br />tobacco product use was included
<br />where appropriate. Additionally, all
<br />models included a random effect for
<br />community to account for similarities
<br />among subjects within jurisdictions.
<br />In a sensitivity analysis, models were
<br />further adjusted for time between
<br />baseline and follow-up questionnaire
<br />completion. Statistical analyses were
<br />based on 2-sided hypotheses tested
<br />at a O.OS level of significance, using
<br />SAS 9.4 (SAS Institute, Inc, Cary, NC).
<br />RESULTS
<br />Of the 2097 participants, 31.1%
<br />(652) lived in a jurisdiction with an
<br />ALA 2014 TRL A grade, and 68.9%
<br />(1445) students lived in jurisdictions
<br />with D or F grades. Sex and ethnic
<br />distributions were similar in A and
<br />D or F jurisdictions, but students in
<br />A jurisdictions were more likely to
<br />come from less -educated households
<br />(Table 1). Unadjusted prevalence
<br />and initiation rates for each tobacco
<br />product were lower in jurisdictions
<br />with A than with D or F grades,
<br />with the exception of new initiation
<br />of hookah with past 30-day use.
<br />Initiation rates were substantial
<br />among never tobacco product
<br />users at baseline, in particular for
<br />e-cigarette use. Both prevalence and
<br />initiation rates of past 30-day tobacco
<br />product use generally did not exceed
<br />10% for any product
<br />For baseline prevalence of ever and
<br />past 30-day use of cigarette and
<br />e-cigarette ever use, and to a lesser
<br />degree for prevalence of cigar use,
<br />jurisdictions with A grades had
<br />generally lower use rates than D or
<br />F jurisdictions (Supplemental Fig 3).
<br />However, within both grade groups,
<br />there was considerable variability in
<br />prevalence rates across jurisdictions
<br />for all tobacco products. Rates in
<br />individual jurisdictions had wide Cls
<br />(results not shown) because of small
<br />sample size. Rates of tobacco product
<br />initiation at follow-up were also
<br />generally quite variable across the
<br />jurisdictions within both A and D or F
<br />grades (Supplemental Fig 4).
<br />At baseline, participants living in the
<br />4 jurisdictions with A grades had
<br />lower odds of ever using a cigarette
<br />(OR 0.61; 95% Cl 0.41-0.90) and
<br />of past 30-day use (OR 0.51; 95%
<br />Cl 0.29-0.89) than participants in
<br />10 D- to F-grade jurisdictions, after
<br />adjusting for sociodemographic
<br />covariates and other tobacco product
<br />use at baseline (Fig 1).
<br />Living in A -grade jurisdictions
<br />was associated with lower odds
<br />of initiation of cigarette use
<br />between baseline and the follow-up
<br />questionnaire (OR 0.67; 95% Cl
<br />0.45-0.99 [Fig 2]). The risks of
<br />Downloaded from www.aappublications.org/news by guest on May 7, 2019
<br />PEDIATRICS Volume 143, number 2, February 2019
<br />
|