Race and the Incidence of Cigarette Smoking Among Adolescents in the United States

Abstract

After almost two decades of decline, cigarette smoking is on the rise among adolescents in the United States. Although in recent years both AfricanAmerican and white adolescents have experienced increases in cigarette smoking, the prevalence of smoking among African-American adolescents has been consistently lower than that of whites (DPrevious studies (2,3) have not been able to explain why the race differential in adolescent cigarette smoking occurs. However, these studies have been crosssectional in nature and primarily examined only the contribution of sociodemographic and behavioral factors. In response, a prospective cohort study was conducted among AfricanAmerican and white adolescents who participated in both the first (1989) and the second (1993) waves of the Teenage Attitudes and Practices Survey (TAPS). The purpose of this study was to determine whether the race differential in the incidence of cigarette smoking is attributed to differences in sociodemographic, environmental, personal, or behavioral factors. TAPS was designed to provide national household-based smoking data on teenagers who were 11-18 years of age. The first TAPS (TAPS I) was conducted in 1989. The primary method of data collection was computer-assisted telephone interviewing. Teenagers who were not available by telephone were sent a mailed questionnaire. The eligible population included 12 097 adolescents. Of that number, 9965 interviews were completed either by telephone or by mail, with a resulting survey response rate of 82% (4,5). For TAPS II, interviews were conducted in 1993. Of the 9965 TAPS I respondents, 9135 were eligible (those completing a telephone interview) and 7960 (87%) teenagers participated in TAPS II (6). Given the 7960 participants, analyses were restricted to African-American and white adolescents (ages 11-17 years) who had never tried cigarettes at base line (n = 3531). Forty-seven (1.3%) respondents were excluded because of missing data on at least one study variable. Thus, the final study population consisted of 3484 adolescents for whom information was complete. Race was the exposure variable, and smoking status was the outcome variable. Based on two questions concerning main racial background and Hispanic origin, respondents described themselves as either non-Hispanic AfricanAmerican or non-Hispanic white. With respect to smoking status, respondents were asked the following questions: 1) Have you ever smoked a cigarette? 2) Have you ever tried or experimented with cigarette smoking, even a few puffs? Those who answered no to both questions were considered never smokers. Those who answered yes to either question were considered trial smokers. Control variables were categorized as sociodemographic (sex, age, and parental education), environmental (household smoking and number of same-sex friends who smoke), personal (five items on perceived benefits of smoking), or behavioral (intention to smoke, organized physical activity, and school performance). Weighted percentages were used to estimate the cumulative incidence of trial smoking for each population subgroup. Then, logistic regression analysis (7) was used to estimate the odds ratio (OR) of trial smoking for whites (compared with African-Americans) before and after adjustment for confounding factors. Logistic regression-model estimates were calculated with the use of SUDAAN, a procedure for analyzing complex sample survey data (8). After 4 years of follow-up, 30.0% of African-American adolescents became trial smokers compared with 43.1% of whites. The crude OR was 1.6 (95% confidence interval [CI] = 1.3-2.0). In Table 1, the crude OR was adjusted for multiple confounding factors. The adjusted OR of 1.8 (95% CI = 1.5-2.3) was virtually the same as the crude OR. Other predictors of trial smoking included household smoking, the number of same-sex friends who smoke, intention to smoke, and school performance. These data suggest that selected sociodemographic, environmental, personal, and behavioral factors do not account for the race differential in the incidence of adolescent cigarette smoking. The present findings are consistent with those of previous studies (2,3). One limitation of this analysis is a possible differential misclassification. African-American adolescents may be more likely to underreport their smoking habits than white adolescents (9,10), resulting in an overestimation of effect. Differential misclassification, however, is not likely to explain the race-smoking association. Investigators (9) have found that the race differential in adolescent smoking persists, even when biochemical measures of cigarette smoking are used. These findings underscore the need for more research on contributors to the race differential in teen smoking. African-American adolescents may be less likely to smoke than whites because of stronger actions against teen smoking by African-American parents (11-13), the belief among African-Americans that tobacco products are being marketed specifically to them (75), or a lower likelihood among African-