This study investigated the impact of school-based smoking prevention programs on adolescents’ smoking knowledge, attitude, intention and behaviors in Aceh, Indonesia. The primary objective of the study was to identify appropriate strategies for smoking prevention programs among Indonesians adolescents, especially in Aceh. It was anticipated that findings of this study would help provide practical recommendations for professionals within Indonesia and perhaps other Muslim populations to challenge the high rates of smoking prevalence, morbidity, and mortality across the country.
Findings from this study suggest that the school-based programs were effective in increasing students’ knowledge about both health-related aspects of cigarette smoking and Islamic teaching and rulings on tobacco smoking in Aceh. Knowledge about health-related aspects of tobacco smoking increased among participants in all of the intervention programs. However, knowledge about Islamic teaching and rulings on smoking increased only in the Islamic-based and combined programs. One possible explanation for this is that participants in the Islamic-based program received a large amount of information about Islamic teaching and the health-related effects of tobacco smoking. Participants in the combined program received components of both the health- and Islamic-based programs, while participants in the health-based program received no information about tobacco smoking from an Islamic viewpoint.
Our results extend findings from previous studies and reviews of school-based smoking prevention programs among adolescents in non-Muslim countries that found school-based programs had a positive impact on adolescents’ smoking-related knowledge [9, 10, 39, 40]. A previous meta-analysis of 47 school-based smoking programs conducted by Rundall and Bruvold  showed that 98% of the reviewed programs successfully increased participants’ knowledge in treatment groups. Similarly, a more recent meta-analysis  showed that 73% of 11 school-based smoking prevention programs resulted in significant improvement in participants’ knowledge about smoking.
Attitude towards smoking of participants in the Islamic-based intervention group improved significantly following the intervention. This finding is consistent with previous work in school-based smoking prevention programs reported in other areas of the world. In Greece, a school-based peer-led smoking prevention program improved anti-smoking attitude , and in Yilan County, Taiwan  a significant change in anti-smoking attitude was observed among junior high school students after a one-week program intervention. In their review of 11 school-based smoking prevention programs in South Korea, Park  reported small, medium, and large effects for four, three, and two studies respectively.
Our study demonstrates that Islamic teaching has a significant role to play in increasing anti-smoking attitude among students in Indonesia. Our analysis suggests that the attitude changes were larger among participants in the Islamic-based program compared with non-Islamic programs. These findings support the recommendation of the educators in our preliminary feasibility test which suggested school-based smoking prevention programs in Aceh would be most effective if they included Islamic teaching and rulings concerning tobacco smoking.
Having an intention to smoke is considered an important predictor of smoking behaviors [33, 37]. In a review of previously published reviews of school-based tobacco use prevention programs by Dobbins et al.  one of two reviews reported positive effects of the programs in reducing smoking intentions while the other reported a promising effect. In our study we also found a positive effect of the programs on the smoking intention, although the effects were insignificant and similar for both the health- and Islamic-based program.
In a previous review , only six of 12 reviews found positive effects of tobacco use prevention programs on smoking behaviors, with two reporting promising effects, and three no effects. Park  found no significant effects of school-based programs on smoking behaviors. Our findings were consistent with these studies showing only insignificant reductions in smoking behaviors for both the health- and Islamic-based programs. This was not however unexpected given the relatively low prevalence of smoking amongst the participants.
The ICCs for outcome variables were generally small to moderate, with the highest being for health knowledge (0.10) and the lowest for the weekly and monthly smoking variables (0.02). Overall, the ICCs for knowledge and attitude (continuous variables) were slightly higher than the ICCS for the smoking behaviors and intentions (categorical variables). This finding suggests that the effects of the program on participants within the same class were more highly correlated with respect to their knowledge and attitude when compared to the effects on behavior and intentions.
There are some potential limitations in this study. One of the limitations of this study is the differences between groups at baseline in knowledge, attitude, and behaviors after randomization due to the cluster-randomized nature of the study with randomization of schools rather than individuals. Although we adjusted for the differences in these variables at baseline in our models, there may still have been residual confounding. For example, students with lower levels of knowledge may also have been different in other characteristics which could have led to the intervention being less effective in this group. In addition, the low smoking rates in certain behavior and intention categories imposed constraints on estimating each program’s effectiveness for behaviors. Additionally, outcome measures were assessed using a self-reported questionnaire and participants might be inclined to underestimate their tobacco use. Although it would have been possible to validate students’ responses about their smoking behaviors using biochemical tests, we were unable to do so because the study was limited by the number of personnel, and technical and financial constraints. The questionnaire also relied on students’ recall ability over long time periods and the limitations of these were evident in their responses between baseline and follow-up questionnaires regarding their lifetime use of tobacco. Validation of our results with other instruments and biochemical tests would therefore be useful. Finally, this study only reported on the short-term impacts of the program; longer-term evaluation of the program is required to determine if the effectiveness of the intervention is sustainable in the longer term.
Notwithstanding these limitations, this study is the first RCT to assess the effectiveness of school-based smoking prevention programs in Indonesia, using specifically developed interventions that included health- and Islamic-based concepts.