The study revealed that about two-fifth (42.3%) of the young men aged 15–24 living in urban slums of Bangladesh were tobacco smokers. This prevalence is much higher when compared with youths of the population in Bangladesh (9.1%) and other neighboring countries such as, India (16.8%), Nepal (13.0%), Sri Lanka (12.4%), Maldives (8.5%), Pakistan (12.4%) and Myanmar (22.5%) [2]. It was also found that the male youths in slums are more likely to use cigarettes than bidis, which could be associated with urban culture, working status and availability of tobacco products [17]. Like the higher prevalence of TS among urban slum youths, the rate of IDU (9.1%) was also 2.6 times higher than the rate (3.4%) for the youths in Bangladesh [5], [8], [16], [36]. A higher level of risky behaviors in deprived or overcrowded areas was also reported in other studies [17], [26]–[29]. Unhealthy lifestyles in adverse socio-economic conditions, weak social norms and cultural beliefs, undesirable neighborhood characters, availability of tobacco products, and lack of preventive services in the deprived areas may have significant influence on individuals' behaviors [27], [28], [37].

Several demographic, socioeconomic and behavioral factors are identified as significant determinants of both TS and IDU among young men in urban slums. For instance, among the youths, significantly higher prevalence of TS and IDU were found among those aged 20–24 years, which are consistent with the findings of other studies in Bangladesh and elsewhere [5], [16], [17], [38]–[40]. This may be partially attributed to the traditional and cultural norms in Bangladesh and stress related issues. In the prevailing cultural norms of the Bangladeshi society, TS among youths is discouraged by elders and family members. TS by any younger members in the presence of older people are viewed as indecent and intolerable. IDU is totally prohibited in Bangladesh. In line with other studies [5], [41], [42], TS and IDU revealed a strong inverse association with level of education. The likelihoods of TS and IDU were found to be approximately five and two times higher among male youths with no formal education, respectively, compared to those with post-secondary education. The finding suggests that improvement of education could be an important strategy for reducing both TS and IDU in urban slums. The health behavior model of stress indicates that populations under stress generally engage in health detrimental behaviors, particularly in the context of low social status [43]. Based on this model, we could say that stressful conditions in slums [17] may lead to an increased risk of smoking and illicit drug use [23]–[25], [44], which is also evident from the higher prevalence of such risky behaviors among the slum youths compared to the youths in the country. Consistent with the health behavior model of stress [43] and other empirical studies [17], the likelihood of TS among youths in new settlements (like slums) is significantly (P<0.001) higher than the youths who lived in the same places since birth. While the likelihood of IDU is higher among the young male migrants who had settled in the slums for a long period of time, the likelihood among the recent migrants (duration less than 10 years) is lower. This may be partly attributed to the social network and environmental factors. Similarly, the higher rates of TS and IDU among the youths having symptoms of STIs could be outcomes of their risky lifestyles in poorly managed living conditions [5], [16], [38].

This study found a significant (P<0.001) positive association between TS and IDU which was consistent with other studies [16], [18], [45], [46]. Other empirical evidence also supported the relationship between TS and IDU. For example, regular use of tobacco was the predictor of life-time drug use [47], [48]. Our study showed that the likelihood of IDU was more than 9 times higher among slum youths who smoked regularly. In line with other studies [5], [16], [18], [49], the likelihood of IDU of any form increased with more frequent cigarette smoking. Although the likelihood of IDU increased with bidi smoking, but its relationship with the consumption rate of bidis is not as clear as that for cigarette smoking.

The strong association between TS and IDU should not be ignored in policy designs and interventions. Cross-sectional data used here are not suitable for making inference on ‘cause-effect’ relationship. However, according to the discussion of [16], generally the smoking behavior is first picked up and then followed by IDU. As both are of public health concerns, their consequences could be reduced by employing suitable intervention programs at different levels. Government and non-government organizations, community and family should be engaged in intervention programs and health promotional activities. Considering the consequences of TS, the government of Bangladesh announced several strategies for tobacco control in the National Plan of Action ‘2007–2010’. These strategies include setting appropriate price and tax policies; prohibition of advertisements and sponsorship; raising awareness through training, education and communication; restrictions on sales of tobacco products to minors; and labeling the harmful effects of tobacco products on packaging [50]. Some of these strategies are already implemented. Now, advertisements of cigarettes and bidis will have to include a warning message stating that smoking is hazardous for health. Warning messages are obligatory on the packaging of cigarettes and bidis. Regrettably, these printed messages on packaging are not totally effective for slum youths as many of them are illiterate. Moreover, as many smokers in urban slums buy cigarettes or bidis by the sticks rather than full packets, they miss the warning labels on the packaging [4], [17].

In connection with other policies [4], [17], [50], some general recommendations emerge from this study. First, given the importance of awareness, digital posters carrying warning messages on the adverse effects of TS and its relation with IDU can be displayed at places in slums where youths gather and also at points of sale. Second, community leaders along with NGO activists and law enforcing agencies should act jointly to reduce the number of available points for sale of tobacco products. They should also restrict smoking to specific places to reduce exposure of other youths to second hand smoking and prevent them from emulating the behavior. Third, since the majority of youths are Muslims, tobacco control strategies should involve religious leaders especially Imams (head of a mosque). They may deliver brief messages about the harmful effects of TS and IDU during Friday prayers (performed by the Muslims together in mosques on Fridays) as a part of tobacco and drug control policy. Fourth, as more than 50% of the youths are poor, they tend to be more price sensitive [51], [52], and increasing prices of tobacco products may limit their usage. Another recommendation is that parents, teachers, elders and other respected persons in the society should assist to prevent the youths from adopting health risky behaviors through close monitoring and mentoring.

The findings in this paper are subject to a few limitations. Self-reported data on TS and IDU from Bangladesh Urban Health Survey could suffer from recall bias and deliberate misreporting. Even though anonymity and confidentiality were ensured during the survey, respondents might have under-reported the incidence as TS and IDU in Bangladesh are not widely acceptable social norms. Such misreporting could under estimate both the prevalence of TS and IDU and lead to inaccurate measures of their relationship with other variables. Further, STIs may be under reported as the negative social stigma associated with STIs will put some pressure on the respondents in revealing the truth. Therefore, the behavioral factors explored in this study are preliminary. Furthermore, the survey did not include information on tobacco and drug initiation and cessation along with TS and IDU history of family members, and familial or environmental stress that may substantially influence smoking and drug abuse behavior. Although many available variables were analyzed, exclusion of some other important variables at the individual, family and community level such as parental background, social networking, sexual behavior and urbanization might limit our findings. A qualitative study is suggested to supplement the understanding of the determinants of TS and IDU among youths in urban slums.

Despite these limitations, the current study provides useful contributions. According to our findings, the prevalence of TS and IDU were remarkably higher among slum male youths compared to the population in general. Although our findings were not directly comparable, some studies found higher prevalence of TS among migrants than non-migrants [53], [54]. Based on these studies, we argue that higher prevalence of TS and IDU among slum youths may be related to rural-to-urban migration. Generally migrants are distinct and unprivileged group in cities, who are mainly employed in low-paying and hazardous jobs [54]. Migration may disrupt social support and network system and the migrants face a higher level of stress as they need to cope with new living conditions, social and cultural contexts and intense competitions [53]. Briefly, isolation from home and the lack of social support, pace of city life along with unstable living and employment conditions may induce a high level of stress among migrants, which ultimately increases the likelihood of smoking [53], [54], [55]. Some smokers also perceive smoking as a coping strategy to reduce stress, anxiety, sadness and anger [55]. Slums in rapidly urbanizing countries are generally featured by poor housing, overcrowding, poor environmental and healthcare services, and other risk factors related to unhealthy lifestyles [20], [56]. In Bangladesh, about 40% of the urban populations are slum dwellers, mostly migrants from rural areas. Rural-based push factors as well as urban-based pull factors lead to migration to urban areas particularly among the youths and adults [56]. The uncontrolled growth of slums put enormous strains on the urban infrastructure and environmental sustainability, thereby influencing the health of slum population in general and slum male youths in particular [20], [53], [56]. In conclusion, TS among young men living in urban slums is high compared to other youths in Bangladesh and other neighboring countries. Moreover, TS was found to be positively associated with IDU. Of all the predictors of IDU, TS revealed the strongest association with IDU. Since both tobacco and illicit drugs are perilous in all aspects and young people from poor families in slums are more likely to be vulnerable, comprehensive strategies combining the proposed and existing policies should be implemented to overcome these problems.