Project FLAVOR: 1-Year Outcomes of a Multicultural, School-Based Smoking Prevention Curriculum for Adolescents

Abstract

To evaluate a multicultural smoking prevention curriculum, 16 schools were randomized to receive the multicultural curriculum or a standard curriculum and program effects on 1-year smoking initiation among 1430 never smokers were assessed. Hispanic boys who received the multicultural curriculum were less likely to initiate smoking than were those who received the standard curriculum; effects were insignificant among other groups. The prevention effect among Hispanic boys is encouraging, but additional research is needed to improve prevention effects among other groups.

Adolescent smoking prevalence in the United States varies by ethnicity, nationality, and acculturation.1–5 Despite the cultural diversity of many schools, few prevention programs have targeted multiple ethnic minority groups or compared culturally targeted curricula with untargeted curricula.6–15

This study evaluated the 1-year outcomes of a multicultural smoking prevention curriculum that addressed psychosocial risk factors for smoking16–18 with multiculturally relevant activities and lessons. The multicultural curriculum was compared with a curriculum that addressed the same psychosocial risk factors without cultural references.

METHODS

Respondents were sixth-grade students attending ethnically diverse middle schools in Southern California in 2000. Details of school selection and student recruitment procedures are described elsewhere.19 Schools were randomized to receive the multicultural curriculum or the standard curriculum (8 schools per experimental condition). Students participated if they and their parents provided active written consent.

Students completed surveys in sixth grade, before the intervention, and 1 year later in seventh grade. The outcome variable was lifetime smoking in seventh grade among sixth-grade never smokers (i.e., trying smoking between the sixth- and seventh-grade surveys). Covariates included ethnicity, generation in the United States, age, and gender.

Two smoking prevention curricula were created for this study, each including 8 weekly classroom sessions conducted by health educators (Ritt-Olson et al., unpublished data). Both curricula addressed smoking-related psychosocial concepts (e.g., social norms, refusal skills). The multicultural curriculum, Project FLAVOR (Fun Learning About Vitality, Origins, and Respect) related these issues to the values of multiple cultures. For example, in 1 activity students assemble a Wheel of Life collage about health using the Asian yin-yang concept of health as a balanced body, mind, and spirit. In another activity, students act out a telenovela (soap opera) about the effects of a son’s smoking on a Mexican-American family. The standard curriculum addressed the same psychosocial risk factors with similar activities (e.g., role-plays, trivia games, art projects), but without cultural references. For example, the Wheel of Life was replaced with a summary wheel of key concepts. Details about the curricula are available from the authors.

Chi-squared and ANOVA analyses assessed the preintervention equivalence of the 2 groups. To evaluate differential attrition, a logistic regression model regressed attrition (coded as 0/1) on experimental condition, baseline smoking, and the experimental condition × smoking interaction.20 Multilevel logistic regression models with school as a random effect to control for intraclass correlation21 assessed the effect of the intervention on smoking initiation between sixth and seventh grades.

RESULTS

Of 2775 invited students, 2131 (77%) provided active parental consent, 211 (8%) active parental refusal, and 433 (15%) no parental response. Of the 2131 students with parental consent, 1970 (92%) completed the sixth-grade survey, and of these respondents, 1571 (80%) completed the seventh-grade survey. Results are presented for 1430 never smokers at baseline.

Table 1 ▶ shows the demographic characteristics of the 1970 sixth-grade respondents. There were no significant demographic differences between groups. One-year attrition rates were 16.83% in the multicultural intervention group and 16.77% in the standard intervention group. Attrition did not differ by experimental condition (Wald χ² = 0.24, P = .62). The interaction of experimental condition × baseline smoking was not significant (Wald χ² = 0.44, P = .51), indicating no differential attrition of smokers versus nonsmokers between conditions. Attrition was higher among baseline ever smokers (22%) than among never smokers (16%) (Wald χ² = 5.05, P < .05) .

TABLE 1—

Demographic Characteristics and Lifetime Smoking Prevalence of Baseline Sample (n = 1970): Southern California, 2000

	Multicultural Intervention	Standard Intervention	Percentage Who Had Smoked by Seventh Grade, Multicultural Intervention	Percentage Who Had Smoked by Seventh Grade, Standard Intervention
n	1040	930
Mean age (SD)	11.3 (0.5)	11.3 (0.5)
Gender
Male	387 (47%)	335 (45%)	9	16
Female	442 (53%)	407 (55%)	7	8
Race/ethnicity
Hispanic	472 (57%)	449 (61%)	10	15
Asian American	221 (27%)	169 (23%)	5	4
Othera	136 (16%)	124 (17%)	9	8
Generation in United States
Student and/or parent(s) born outside United States	568 (69%)	569 (77%)	7	12
Student and parent(s) US-born	261 (31%)	173 (23%)	12	11
Lifetime smoking prevalence—6th grade
No	752 (91%)	678 (91%)	8	11
Yes	77 (9%)	64 (9%)	. . .	. . .

Note. There were no significant demographic differences between groups.

^aThe “other” category was 37% White only, 24% Asian/Hispanic, 12% African American/Hispanic, 9% Asian/Hispanic/White, 6% African American only, 6% ethnicity not specified, and 6% other combinations of these categories.

Among the sixth-grade never smokers, 8% of those who received the multicultural intervention and 11% of those who received the standard intervention tried smoking by seventh grade. Table 2 ▶ shows odds ratios of smoking initiation by seventh grade among the 1430 sixth-grade never smokers. Overall, the program effect was not significant (odds ratio [OR] = 0.75, 95% confidence interval [CI] = 0.48, 1.18). However, the multicultural curriculum prevented smoking initiation among Hispanic boys (OR = 0.49, 95% CI = 0.27, 0.88).

TABLE 2—

Odds Ratios for Smoking Initiation Among Baseline Never Smokers (n = 1430), Multicultural Curriculum vs Standard Curriculum, by Gender and Race/Ethnicity: Southern California, 2000

	Male		Female
Group	OR	95% CI	OR	95% CI
Hispanic	0.49	(0.27, 0.88)	0.71	(0.38, 1.32)
Asian American	0.35	(0.04, 2.91)	1.89	(0.35, 10.07)
Other	0.68	(0.20, 2.37)	2.10	(0.30, 14.82)
Overall sample, OR (95% CI)		0.75 (0.48, 1.18)

Note. OR = odds ratio; CI = confidence interval. All odds ratios are adjusted for age, generation in the United States, and the random effect of school.

DISCUSSION

In this evaluation of a multicultural smoking prevention curriculum relative to a similar program without cultural references, the multicultural curriculum prevented smoking initiation among Hispanic boys but not among other groups. Because Hispanic boys had the highest risk for smoking initiation in this study, preventing smoking among this high-risk group is an important public health goal.

Because only 9% of the respondents had already smoked by sixth grade, this study did not have sufficient power to assess program effects on smoking escalation among baseline ever smokers. The low prevalence of smoking initiation between sixth and seventh grade also limited the power to detect significant program effects on initiation. Longer-term follow-up is needed to determine which students in this sample will try smoking later in adolescence. Both 2- and 3-year follow-ups are planned.

All California public schools are required to provide antitobacco education in grades 4 through 8.22 Therefore, a true control group does not exist. This study had limited power because it compared culturally enhanced prevention strategies with other effective, evidence-based strategies. These results are limited by self-report data23,24 and might not generalize to the students who were lost to attrition or did not provide initial parental consent.20

Research is needed to understand how adolescents in diverse settings integrate the values from their families’ cultures of origin with their day-to-day experiences in the urban US culture to form their ethnic identities, and how those ethnic identities influence their decisions about smoking. Additional research is needed to determine the best strategies for smoking prevention among diverse groups of adolescents.