Tobacco use is a leading cause of morbidity and mortality globally. Tobacco's contribution to ill health is projected to rise if current smoking trends are not reduced across the world 1. Although much of the long-term effects of smoking are observed in adulthood, the majority of smokers initiate the habit of smoking as adolescents.

There is a growing body of literature on adolescent tobacco use. Much of the research on the estimate of the prevalence and associated factors of adolescent smoking has been driven by the Global Youth Tobacco Survey (GYTS); an initiative spearheaded by the United States' Centers for Diseases Control and Prevention (CDC), the World Health Organization (WHO) and the Canadian Public Health Association 2, and a group of authors who have been conducting secondary analysis of GYTS data 3. Traditional reporting of results from the GYTS has concentrated on ages 13 to 15 years, and estimates of prevalence and the general socio-environment that may be associated with teenage smoking.

Odenwald et al 4 have reported on the more than usual use of khat in Somalia as an indication of heightened prevalence of substance abuse in an environment of civil strife as is characteristic of most of Somalia. The prevalence of any tobacco use was 32.8% (26.8% among males, and 33.1% among females) and 15.8% currently smoked cigarettes (10.2% among males, and 11.1% among females) in the GYTS 2004 among adolescents 13 to 15 years of age 5.

While the prevalence of cigarette smoking in Somaliland has been reported, we are unaware of efforts to associate antismoking messages with current smoking status, and attempt to explain the prevalence of cigarette smoking within a conceptual or theoretical framework of health behaviors. We therefore carried out this study to estimate the prevalence of current cigarette smoking among all study participants in the Somaliland GYTS, identify positive antismoking messages, and explain their associations with current smoking status within theory-based outlook.

Altogether 1563 adolescents participated in the survey. For 1285 study participants who reported their sex, 63.3% and 36.7% were males and females, respectively. However, 1122 responded to the question on whether they smoked cigarettes in the previous 30 days. Altogether 15.8% of the respondents reported having smoked cigarettes in the previous 30 days. There was no significant difference in the rate of reported smoking between sexes (10.3% in males versus 11.1% in females).

Males tended to be older than females. Female respondents were more exposed than males to antismoking media messages; antismoking messages at sports events, fairs, concerts, community events or social gathering; teaching on the dangers of smoking; and antismoking messages in newspapers or magazines. No associations were observed between being exposed to religious organization that discouraged young people from smoking and gender. Further description of the study population is presented in Table 1.

Table 2 shows results of multivariate analysis after adjusting for age and sex. The following factors were not associated with reported smoking: being taught dangers of smoking, frequency of antismoking posters seen, frequency of antismoking messages seen in newspapers or magazines, youth groups discouraged young people from smoking, and health professionals explained why smoking is dangerous to health. Inconsistency results were obtained for the factor: frequency of antismoking messages seen on billboards. While there was no difference in the odds of reported smoking between those who had not seen messages on billboards and those who had seen a lot of messages, respondents who had seen a few messages were more likely to report smoking compared to those who had not seen any message.

The prevalence of current smoking was 15.8% with no sex difference in reported smoking prevalence between males and females. The effect of sex on susceptibility to smoking is unclear. Some studies found that boys were at increased risk, some studies found that it is girls who were at increased risk, while others found no significant difference males and females 141516171819.

In multivariate analysis, some knowledge or information related factors were associated with reported non-smoking while others were associated with reported smoking. Factors associated with knowledge of the harmful effects of smoking (having discussed with family or taught in school about harms of smoking) were associated with reported non-smoking. According to the HBM, teaching adolescents on the dangers of smoking may be important to enhance their perception of the severity of the condition. Individuals who perceived smoking as harmful to health would be less likely to smoke. In addition, the KAP model's aim would be to improve the knowledge levels of the adolescents thereby leading to behavioral change. Previous reports on school based smoking intervention programs have shown mixed results 202122. We again believe that the content and mode of delivery of these programs do matter as to their effectiveness.

It is interesting to note that exposure to anti-smoking messages through health workers, and print media (posters, billboards, newspapers, and magazines) was not associated with current reported smoking status. However, having been exposed to anti-smoking messages through sports or community events seem to have achieved a negative result of exposure to these messages being associated with reported smoking. This finding highlights the fact that the content and the delivery mode of anti-smoking messages may result in possibly unintended consequences.

Cohen et al 23 in their review of 399 televised anti-smoking advertisements that were designed to discourage adult and teen smoking found that most advertisements had emphasized the consequences of smoking more than the viewer's self-efficacy to quit smoking. Within the HBM and KAP models, these advertisements may have improved knowledge of harmful effects of smoking (perceived severity) but were limited in promoting self-efficacy that may spur the smoker to quit. Furthermore, advertisements were more likely to use informational and humor appeals, rather than sadness, fear, or anger appeals. Biener et al 24 have reported that anti-smoking advertisements featuring messages about serious health consequences that had been independently rated as high in negative emotion were more likely to be recalled and were perceived as more effective by youth survey respondents than ads featuring messages about normative behavior for teens or ads relying on humor. Advertising intensity, while contributing to recall, was negatively related to perceived effectiveness.

Flay and Sobel 25, and Rogers and Storey 26 have reported that mass media anti-smoking messages that promote inter-personal communication are more effective than those that do not. We found that adolescents who were exposed to interpersonal communication were less likely to report smoking than if they had not. We also believe that many of the billboards, posters, and newspaper anti-smoking messages may have possibly been a health warning on a pro-tobacco advert. Such advertisements, promoting glamour are less likely to result in reduced smoking among adolescents. Thrasher et al 27 have reported that cigarette packages whose warning labels contain prominent graphic imagery are more likely than text-only warning labels to promote smoking-related knowledge and smoking cessation.

Mass media campaigns are performed to promote health knowledge, change attitudes and may change behavior 2829303132. The KAP model, if effective, would be expected to lead to reduced smoking rates among adolescents. Staten and Ridner 33 have reported that inappropriate messaging may be less effective if no targeting is used.

We found that adolescents who reported that their religion discouraged smoking were less likely to report smoking. The lower rates of reported smoking could have been mediated through social norms within society where if society perceives a practice as unbecoming, adolescents are less likely to smoke. However, results on the positive role of religion are not consistent. Knowledge of pronouncement that smoking was sin was not associated with lower level of smoking in Egypt 34.

Adolescents who were taught that people of their age do not smoke may have reported less smoking as a result of improved efficacy that they are able not to smoke, as well as in conformity to normative beliefs (social cognitive theory).

In a study of school-going adolescents in Somaliland, we found that adolescents who were exposed to anti-tobacco messages or discussions in school and at home were less likely to report smoking than those who were not exposed to such experiences. A combination of school and home based antismoking interventions may be effective in the control of adolescent smoking in Somaliland. Exposure to anti-smoking print messages was not associated with reported non-smoking in the current study, and we propose that studies be done on the effectiveness of anti-smoking messages (in terms of size, color and content) on adolescent smoking in Somaliland.

The authors declare that they have no competing interests.

SS led data analysis, participated in the interpretation of findings and drafting of manuscript. ASM participated in the interpretation of the findings and drafting of the manuscript. ER participated in the interpretation of the findings and drafting of the manuscript. All authors read and approved the final manuscript