Impacted (A Policy Research Partnership to Reduce Youth Substance use), as part of Bridging the Gap: Research Informing Practice for Healthy Youth Behavior. We thank Steve Chicken and DRP. Laura Kahn at the Division of Adolescent and School Health, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, for their assistance with the YEARS data. We also thank DRP. Jamie Christi and Joanna King at the Matched Corporation for their assistance with the smoke-free air index and verification of the smoke-free air data.
ABSTRACT: OBJECTIVES: To determine the relationship between state-specific estimates of youth and adult cigarette smoking prevalence, overall, and after adjusting for cigarette prices and strength of smoke-free air laws. METHODS: Crude relationships were determined using state-specific adolescent and adult smoking estimates from three national surveillance systems conducted during 1997, 1 999, or 2000. Weighted legislatures regression analyses were conducted to assess crude and adjusted relationships between state-specific estimates of adolescent and adult smoking.
RESULTS: In each crude analysis inducted, adolescent smoking prevalence was significantly and positively related to adult smoking prevalence. These relationships were eaten dated, but generally persisted, after controlling for cigarette prices and strength of smoke-free air laws. CONCLUSIONS: Results support the premise that adult smoking influences adolescent smoking behavior. Fenders and policy makers need to consider that an effective youth prevention strategy may be to curb smoking among adults. 3 Introduction Cigarettes are the most common form of tobacco used in the United States, among both youths and adults (1, 2).
Interest in preventing adolescent uptake f tobacco use increased substantially during the early and mid-1 ass’s (3), as adolescent smoking initiation and prevalence increased (1, 4-9). This prompted considerable debate in the public health community about the relative merits of a youth or adult-centered tobacco control approach (10-14). A focus on youth has often been viewed by policy-makers as more politically palatable to the communities they serve; however, many researchers have argued that since the problem of tobacco affects people of all ages, effective solutions must do so as well, thereby favoring a more balanced strategy (10-14).
An effective approach would target audiences in every age group, encouraging adults to quit without ignoring the reality that virtually all new tobacco users are children or adolescents. A considerable number of studies have noted relationships be;en parental and adolescent smoking (15-21 Banana and colleagues noted that a key distinction in studies of parental and adolescent smoking was to distinguish whether the parents were current, former, or never smokers. When they made such distinctions, they found that the relationship between parental smoking status and adolescent smoking was as strong as that for peer mocking (16-17).
Chains and colleagues found that parental smoking cessation may help to lower the risk for adolescent smoking when the other parent was not a current smoker (20). Parkas and colleagues noted that the earlier parents quit, the less likely their children were to become smokers (21 4 To test the hypothesis that state-specific smoking prevalence for adolescents and adults would be directly related, we initially studied the relationship using data from the 1 997 Youth Risk Behavior Surveillance System and Behavioral Risk Factor Surveillance System (22).
We documented a direct relationship, a ending also noted by Males (23). To assess this phenomenon more fully, we conducted similar analyses using data from additional years and another surveillance system (the National Household Survey on Drug Abuse). Furthermore, because we recognized that cigarette prices and the strength of smoke-free air laws could influence both adolescent and adult smoking prevalence, we also studied the relationship after controlling for these important policy variables (2426).
We hypothesized that the relationship between adolescent and adult smoking would be attenuated, but not eliminated, after controlling for these potential covariates. Methods Data Youth and adult smoking data for this study were taken from three inconsiderateness’s surveillance systems: 1) the Youth Risk Behavior Surveillance System (YEARS); 2) the Behavioral Risk Factor Surveillance System (BRASS); and 3) the National Household Survey on Drug Abuse (NASDAQ). The YEARS provides state-specific adolescent data on public high school students between the approximate ages of 14 to 18 years.
For this study, we used the following measures of adolescent smoking from YEARS: current smoking prevalence, frequent cigarette use, youth ever smoking, and youth ever-daily smoking. The 1997 and 1 999 YEARS define current smoking prevalence (current cigarette use) as having smoked on at least 1 Of the 30 days preceding the survey, and frequent cigarette use as having smoked 5 on at least 20 of the 30 days preceding the survey. The 1 997 and 1999 YEARS define youth ever smoking (I . E. Lifetime cigarette use) as having ever tried cigarette smoking, even one or TV’0 puffs (6, 27).
The 1999 YEARS defines youth ever-daily smoking as having ever smoked at least 1 cigarette every day for 30 days (27). Weighted YEARS data were published for 24 states In 1997, and for 22 states in 1999. The Centers for Disease Control and Prevention (CDC) weighted these stratospheric estimates to adjust for moroseness and varying probabilities of selection. The data are considered to be representative of all public high school students (grades 9-12), in the respective states. In our analyses, we only included data from states with weighted YEARS data.
State-specific sample sizes ranged from 1,325 to 8,636 participants in 1997, and from 1 , 248 to 25 participants in 1999 (6, 27). Standard errors for these weighted 1 997 and 1999 YEARS data were provided by the Centers for Disease Control and Prevention, and were used to estimate arranges for analyses. The BRASS provides state-specific estimates of major risk behaviors among adults aged 18 years and older. Adult current smoking and adult ever smoking measures were included as independent predictor variables from 1997 and 1999 BREWS data.
In the 1997 and 1999 BRASS, current smokers were those who had ever smoked at least 100 lifetime cigarettes and who currently smoked every day or some days. Adult ever smoking was defined by the 1 997 and 1 999 YEARS as having ever smoked 1 00 lifetime cigarettes. We used adult BRASS data from all states for which we also ad YEARS data, which were 24 states in 1997 and 22 states in 1999. State- specific sample sizes ranged from 1,595 to 3,596 participants in 1997, and from 1,633 to 5,011 participants in 1999 (28-29). The NASDAQ provides state-specific adolescent and adult data on substance abuse for adolescents between the ages of 12 to 17 years, adults between the ages of 18 to 25 years (referred to below as young adults), and adults greater than or equal to 26 years (referred to below as adults). In the 1999-2000 NASDAQ, current smokers were those who smoked all or part of a cigarette on at least one of the 30 days preceding the survey. Representative samples were drawn from all 50 states and the District of Columbia, with sample sizes ranging from 900 to 1,030 in 42 states and the District of Columbia, and from 3,600 to 4,630 in 8 states.
About one-third of each sample represented each age category: 12 to 17 years; 18 to 25 years; and 26 years (30). State- specific estimates for price, as Of November 1st of each year, were taken from The Tax Burden on Tobacco (31 ). The average price of a pack of cigarettes was constructed by using weighted averages for a pack of 20 cigarettes based on the prices of single packs, cartons, and vending machine sales, where the eights are the national proportions of each type of sale.
These prices are inclusive of state level sales taxes applied to cigarettes, but are exclusive of local cigarette taxes. Because the price published is as of November 1st, and because the surveys are conducted throughout the year, we created a weighted average annual cigarette price measure by subtracting state and federal excise taxes from the current years price and the previous/following year’s price and weighting the pre-tax prices accordingly.
Average federal and state excise taxes for the whole year were calculated and added to the weighted average pre-tax price. Data on state-specific smoke-free air legislation were compiled to construct a smoke-free air (SFA) legislation index, using a multi-step process. Initially, these legislative data were taken from the American Lung Association’s ‘State Legislated 7 Actions on Tobacco Issues’ (SLATS) System, and the Centers for Disease Control and Prevention’s ‘State Tobacco Activities Tracking and Evaluation’ (STATE) system.
We then contracted with the Matched Corporation to validate initial coding, and expand upon our initial categorization scheme by incorporating legislative information on additional locations, such as schools, correctional facilities, and cultural facilities. The state-specific SFA index values were constructed from ratings given to each state, based upon the levels of restriction provided for the following 10 locations in 1997, 1999, and 2000: private worksheet, health facilities, restaurants, recreational facilities, cultural facilities, retail/grocery stores, shopping centers, public transit, public schools, and private schools.
SFA ratings were summed for each of these 10 locations, and additional weighting Was given to 6 designated youth-oriented locations (restaurants, recreational facilities, cultural facilities, shopping enters, public schools, private schools), which were multiplied by 2 prior to summation. After the ratings were summed, 20% of this total SEA score was then subtracted for the existence of any state preemption clauses. The calculation of the subtracted preemption percentage was based upon the average estimated percentage of states with SFA preemption in relevant youth-oriented categories, as described in a paper by Christi et al (2002) (32).
Preemption clauses prevent a local area, within a state, from enacting smoke- free ordinances that are stronger or more protective than state smoke-free air laws. Statistical Analysis Weighted least-squares regression analyses were conducted using SPAS software. Regression analyses Of adult smoking measures, as the independent predictor variables, on adolescent smoking measures, as the dependent outcome variables, were conducted 8 for BRASS, YEARS, and NASDAQ data. Analyses with YEARS data were conducted overall and by gender (male, female).
All regression analyses were weighted by the reciprocal of the variance of the dependent variables. Average price of a pack of cigarettes and strength of smoke-free air legislation were included as potential covariates in adjusted weighted least squares regression analyses. Crude and adjusted beta coefficients were calculated and reported, along with standard errors, r-squared values, and statistical probabilities (p-values). Additional weighted least-squares regression analyses were conducted to further adjust for income disparity.
These analyses did not produce noticeably different results for youth-adult data; therefore, income disparity was not considered relevant for adjustment. Rest Its Table 1 presents crude and adjusted results from the weighted least-squares regression analyses of youth and adult smoking measures. In each crude analysis conducted, adolescent smoking prevalence was significantly and costively related to adult smoking prevalence. These relationships were attenuated, but generally persisted, after controlling for cigarette prices and strength of smoke-free air laws.
Adjusted overall relationships for 1997 YEARS and BREWS data, between youth-adult current smoking prevalence and frequent use, were attenuated; but remained significant. This attenuated, but significant, relationship persisted among males for current smoking prevalence (with borderline significance among females), and among both males and females for frequent use. Crude relationships between youth-adult current smoking prevalence and frequent use were significant for 1999 YEARS and BRASS data, and adjusted relationships remained significant among females for current prevalence and frequent use.
Crude relationships for NASDAQ data from all states and the District of Columbia were also highly significant for youth, young adult, and adult smoking in 1999-2000 (See also: Figure 1). Adjusted relationships for 1999-2000 NASDAQ data also remained significant for all youth, young adult, and adult smoking data. Table 2 presents results from additional weighted least-squares regression analyses that were conducted to explore a possible legislation between youth and adults with respect to measures Of smoking initiation.
These analyses, using 1997 YEARS and BRASS data, showed a significant adjusted relationship between youth ever-smoking and adults ever-smoking at least 100 cigarettes. Analyses using 1999 YEARS and BREWS data showed significant crude and adjusted relationships between youth ever-daily smoking and adults ever smoking at least 1 00 cigarettes. Discussion These analyses were conducted to determine the relationship between stratospheric estimates of youth and adult cigarette smoking prevalence, verbal, and after adjusting for important policy covariates.
In each crude positively related to adult smoking prevalence. After adjustment, the adolescent-adult relationship was attenuated, but remained significant, for: 1997 overall and male current prevalence; 1997 overall, male, and female frequent use; 1999 female current prevalence and frequent use; and all age groups tested using 1999/2000 NASDAQ data. Therefore, the relationships generally persisted after controlling for two important policy variables, price and strength of smoke-free air 10 legislation.
Adjusted analyses, using 1997 and 1999 YEARS and BRASS data, also showed a significant relationship between the following measures of smoking initiation: youth ever smoking and adults ever smoking at least 1 00 cigarettes; and youth ever-daily smoking and adults ever smoking at least 100 cigarettes. There are several limitations regarding these analyses. Results for the YEARS/ BREWS data may be influenced by the relatively small number of states with weighted data used in analyses. There were 24 states with weighted YEARS data in 1997, and 22 States with weighted YEARS data in 1999.
BREWS data from 1997 and 1 999 were only used for the same number of corresponding dates with weighted YEARS data in both respective years. The ecological fallacy may also be involved, since smoking behavior data were drawn and analyzed from state-specific population data. Other variables, such as relationship quality between adolescents and parents, may mediate the relationship between adolescent and adult smoking prevalence. Further research is needed to explore additional variables, which cannot be ruled out by these analyses, and may affect the state-specific relationship between adolescent and adult cigarette smoking.
Results are consistent with the notion that adult smoking influences adolescent smoking. Findings are also consistent with parental literature, suggesting that youth behavior models adult behavior, and other research, suggesting that if adults quit youth may be less likely to smoke (16, 17, 19-21 These data support the belief that efforts to prevent initiation and promote quitting, among both adolescents and adults, would be included as key components of an optimal tobacco control strategy and an effective public health effort to reduce tobacco- related mortality and morbidity.
An optimal tobacco control strategy would also include a component to protect non-smokers from 11 environmental tobacco smoke. Gallant and Jameson have proposed that tobacco control efforts directed at adolescents and young adults need to also emphasize smoke-free air policies, which encourage smoking cessation anonymous, as well as adults (26). Research suggests that population tobacco control strategies that influence adult smoking, like price and smoke- free air, also influence youth smoking (33-38). Therefore, these strategies have a two-for-one effect.