Factors influencing tobacco smoking among men aged 15–49 years in Nepal

Tobacco smoking is one of the leading preventable causes of death worldwide, responsible for millions of deaths each year¹. The ongoing tobacco epidemic is one of the most serious public health challenges², causing over 8 million deaths annually, and of these, more than 7 million are directly attributable to tobacco use, while approximately 1.3 million result from second-hand smoke exposure¹. As of 2020, approximately 1.18 billion people globally smoke tobacco, with an overall smoking prevalence of 32.6%³.

Tobacco smoking is a critical public health issue, especially in low- and middle-income countries (LMICs), which account for 80% of the world’s 1.3 billion users⁴. Globally, cigarette smoking is the most prevalent form of tobacco use, alongside other products such as waterpipe tobacco, cigars, cigarillos, heated tobacco, roll-your-own tobacco, pipe tobacco, bidis, kreteks, and smokeless tobacco¹. Tobacco chewing is popular in South Asia (particularly in India, Bangladesh, and Sri Lanka)⁵ and in some parts of the United States, chewing tobacco is used either as loose leaf, plug tobacco, or twist tobacco⁶. Also, it can be either sniffed through the nose or placed in the mouth and is particularly popular in Northern Europe, especially in Sweden and the United Kingdom⁷. Dipping tobacco which is similar to chewing tobacco also involves placing a pinch of loose tobacco leaves between the lip and gum⁸.

Data from 82 LMICs revealed an average smoking prevalence of 16.5%, with significant variation between countries⁹. Projections indicate that by 2030, about 70% of the estimated 10 million smoking-related deaths will occur in these countries if current trends continue¹⁰. Tobacco is linked to numerous chronic diseases such as cancer, cardiovascular diseases, and respiratory issues^11,12. Approximately, 7.8% of total Disability-adjusted life years (DALYs) are linked to tobacco consumption¹³. Despite the significant tobacco burden, comprehensive cessation services remain scarce in low-income countries, with none offering best-practice services as of 2020¹⁴. This issue is evident in Nepal, where 28.3% of adults used tobacco in 2019¹⁵. The trend persisted in the 2022 Nepal Demographic Health Survey, which found that 28% of men still smoked some form of tobacco¹⁶.

Tobacco smoking is often influenced by role models such as parents, peers, and celebrities. This behavior can be explained by the Social Learning Theory, as formulated by Albert Bandura¹⁷. This theory posits a particular behavior is acquired through observational learning, modeling, imitation, attitudes, and emotional reactions of others. Younger people who see family members using tobacco may perceive it as normal behavior, significantly increasing the likelihood of them starting to use tobacco themselves^18,19. Peer influence is particularly potent during adolescence. Teens are more likely to begin smoking if they believe their friends approve of smoking and they become driven by a desire to fit in^20,21. Additionally, the portrayal of smoking by celebrities and in various media forms or advertisements adds a layer of glamour to smoking, further influencing attitudes toward tobacco use²².

Several studies have identified predictors of tobacco smoking among Nepalese men, including factors such as age²³, educational status^13,15,23, economic status²⁴, and peer pressure²⁵. Additionally, the 2016 Nepal DHS found associations between age, marital status, wealth index, religion, ecological region, and frequency of listening to the radio with tobacco use²⁴. However, since 2016, shifts in social, economic, and public health landscapes may have influenced the predictors of tobacco smoking among Nepalese men, potentially altering their impact on smoking behaviors. Changing educational levels, urbanization, increased exposure to global media, and evolving cultural attitudes could play a role in either reinforcing or mitigating tobacco use patterns. Thus, this study aims to investigate which factors still significantly influence the persistent trend in tobacco smoking among men in Nepal. Identifying comprehensive risk factors can inform targeted interventions, improve health outcomes for Nepalese men, and contribute to global efforts in achieving Sustainable Development Goal 3 (SDG 3.a), which strengthens the implementation of the WHO Framework Convention on Tobacco Control²⁶.

Data source and design

This study utilized data from the 2022 Nepal Demographic and Health Survey (NDHS). The 2022 NDHS was implemented by New ERA under the aegis of the Ministry of Health and Population¹⁶. The 2022 NDHS employed a cross-sectional study design to collect information on fertility, marriage, family planning, breastfeeding practices, nutrition, food insecurity, maternal and child health, childhood mortality, awareness and behavior regarding Human Immune Virus (HIV), Acquired immunodeficiency syndrome (AIDS) and other sexually transmitted infections (STIs), women’s empowerment, domestic violence, fistula, mental health, accident and injury, disability, and other health-related issues such as smoking, knowledge of tuberculosis, and prevalence of hypertension¹⁶. This survey, as part of the global demographic health survey series, achieves national representativeness for developing countries. The survey employed a stratified two-stage cluster sampling approach to ensure that the dataset accurately reflects populations at both national and regional levels¹⁶. Approval for access and use of the dataset was obtained from ICF International via the DHS website (http://www.dhsprogram.com). In drafting this paper, the study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines²⁷.

Definition of study variables

Outcome variable

The outcome variable for this study was tobacco smoking, assessed using the variable "frequency currently smokes tobacco," which included three response options: "do not smoke," "every day," and "some day." To create a binary response for analysis, individuals who responded "do not smoke" were classified as “no” for tobacco smoking, while those who responded either “every day” or “some day” were combined into a single “yes” category for tobacco smoking.

Covariates

Based on the literature^13,15,23,24 and the availability of variables in the dataset, we considered twelve¹² independent variables as covariates. These variables included individual-level factors (age, marital status, educational level, occupation, internet use, mobile phone use, alcohol consumption, depression “have you been told by the doctor or health worker you have depression” and self-reported health status), community-level factors (wealth index, residence, and ecological zone).

Statistical analyses

Data analysis was conducted using STATA version 17. The analysis included 4,913 men from the male recode (MR) file of the 2022 NDHS¹⁶. To examine factors of tobacco smoking, a multilevel logistic regression was performed across four models. The first model, the Null Model, accounted for random effects and assessed model fitness. Model 1 evaluated the association between tobacco smoking and individual-level factors, while Model 2 focused on community-level factors. The final model (Model 3) examined the combined association of both individual and community-level factors with tobacco smoking. All estimates were adjusted for any potential sampling bias by applying individual sampling weight variable (mv005). Due to the complex nature of the DHS survey, we used the survey command (svyset) in STATA to account for clustering effects and sample weight effects. Test of statistical significance was carried out at a 5% level of significance in all the statistical analyses. Results were presented as adjusted odds ratios (AORs) with 95% confidence intervals (CIs). Model fit was assessed using the Akaike Information Criterion (AIC) to assess how well the model’s predictions match the observed data. The lowest AIC values indicate the best fit²⁸; Model 3 (AIC = 5182.024) was identified as the best-fitting model in this study. Multicollinearity was assessed using the variance inflation factor (VIF), with no evidence detected (VIF = 1.37). The possibility of multicollinearity is ruled out since the VIF score is below 3²⁹.

Socio-demographic characteristics of respondents

Table 1 below shows the socio-demographic characteristics of respondents in the study. The majority 1,842 (37.5%) of the respondents were aged 15–24 and 3,462 (70.5%) resided in urban areas. Out of the 4,913 men sampled, 2,244 (45.7%) of them had secondary-level education. Most 3101 (63.1%) of the respondents are currently in a union, and 1156 (23.5%) are in agriculture or self-employment. The majority 1135 (23.1%) and 1137 (23.1%) of them belonged to the richer and richest categories respectively. A higher proportion 4498 (91.6%) of them possessed a mobile phone and 3829 (77.9%) had access to the internet. In terms of health, 2347 (47.8%) of the respondents reported moderate health status. Interestingly, only a small percentage 76 (1.6%) of them reported depression. Alcohol use was prevalent among 2077 (42.3%) of the men, while most of them resided in hilly 1973 (40.2%) or terai 2685 (54.7%) regions.

Prevalence and distribution of tobacco smoking across the variables

The prevalence of tobacco smoking was found to be 28.0% among respondents. The prevalence was high among men aged 25–34 (33.4%), with basic level education (35.6%), previously in a union (47.0%), and skilled manual laborers (37.4%). Moreover, tobacco smoking was higher among men who did not use the Internet (30.9%) and consumed alcohol (43.6%) (Table 1).

Factors influencing tobacco smoking among respondents

Our findings revealed that men aged 35–44 (aOR = 0.80, 95% CI [0.82–1.29]) and 45–49 (aOR = 0.70, 95% CI [0.53–0.99]) had reduced odds of using tobacco compared to younger men aged 15–24. Men with higher educational levels had decreased odds of smoking tobacco (aOR = 0.40; 95% CI [0.24–0.61]) as compared to men with no formal education. Men who lived in hilly (aOR = 0.70; 95% CI [0.50–0.92]) and terai regions (aOR = 0.70; 95% CI [0.50–0.96]) also had reduced odds of smoking tobacco as compared to those living in mountainous areas (Table 2).

However, the odds of smoking tobacco were highest among men who were involved in skilled manual labor (aOR = 2.60; 95% CI [1.89–3.61]) relative to those who were not working. Men who consumed alcohol had higher odds of smoking tobacco (aOR = 3.80; 95% CI [3.28–4.40]) than those who did not consume alcohol. Men in the richest wealth category also showed increased odds of smoking tobacco (aOR = 1.40; 95% CI [1.02–1.89]) compared to those in the poorest category. Moreover, men who reported good (aOR = 1.20, 95% CI [1.02–1.38]) or moderate (aOR = 1.60, 95% CI [1.13–2.14]) health status had higher odds of smoking tobacco compared to those who reported bad health status (Table 2).

The present study reported a prevalence of tobacco use to be 28.0% among men in Nepal. This finding is lower than the prevalence reported in the Maldives (41.2%) but higher than the prevalence reported in Pakistan (20.1%), Afghanistan (22.0%), and India (23.0%)³⁰. The evaluation of the existing tobacco control policies in Nepal reveals significant gaps that suggest a need for improvement in their effectiveness. According to the Tobacco Products (Control and Regulatory) Act, 2011, various measures including smoke-free public places, pictorial health warnings, bans on advertising, and taxation have been instituted to curb tobacco consumption³¹. However, the current high prevalence of tobacco use reported in this study indicates that these policies are not fully effective in discouraging use or promoting cessation, especially in comparison with neighboring countries^30,32. This implies that although the governmental framework appears robust on paper, challenges in implementation and enforcement significantly undermine its potential impact. In particular, while the act aims to create smoke-free environments, compliance remains low as many individuals and establishments continue to allow smoking in public places³³. In terms of taxation, raising the prices of tobacco products has been documented as a successful method to reduce consumption internationally, particularly among price-sensitive demographics including youth³⁴. Nevertheless, the taxation levels in Nepal are not sufficiently high to exert the desired effect on consumer behavior limiting the potential impact of the Tobacco Control Act³⁵. Nepal’s tax currently covers only approximately 38% of the retail price, far below the WHO-recommended rate of 75%³⁶. Although advertising and promotion bans exist, enforcement is inconsistent, with violations observed at points of sale and through surrogate advertisements³⁷. Furthermore, studies indicate that pictorial health warnings tend to be more effective in attracting consumer attention compared to text-only warnings^38,39. However, the implementation of these warnings in Nepal has not yet reached optimal effectiveness, possibly due to poor dissemination and a lack of public awareness surrounding their existence³³.

Our study found that older men were less likely to smoke tobacco compared to younger men. This aligns with findings from a previous study conducted in Ghana, which also reported higher odds of tobacco smoking among younger men than older men⁴⁰. However, our result contrasts with the findings of Islam et al. who observed that in parts of Asia, tobacco smoking was more prevalent among older men than younger men³⁰. These regional differences may be explained by the Social Learning Theory, which posits that behaviors are learned through observing and imitating others. Specifically, older men may have been exposed to influential role models such as family heads and celebrities during adolescence and early adulthood, encouraging early initiation of tobacco smoking among the younger men, as observed in our study. This finding suggests a shift in social norms or increased health risk awareness with age consistent with research showing that older individuals are more likely to quit smoking⁴¹.

Higher education emerged as a significant factor influencing tobacco smoking in our study. Constant with extant literature, this finding aligns with a study in India and Japan^2,42. It is reasonable to hypothesize that men with higher levels of education typically exhibit greater self-efficacy, healthier lifestyles, and better access to information⁴³ than those with lower levels of education.

We also found high odds of tobacco smoking among respondents who consumed alcohol. Our finding is consistent with previous studies in England and China^44,45. The interaction between alcohol and brain nicotinic receptors, followed by the activation of brain reward pathways and the release of dopamine, offers a theoretical explanation for this effect^46,47. This association may further be influenced by the social and cultural acceptance of tobacco and alcohol products in Nepalese society^48,49.

Also, residing in hilly and terai regions decreased the odds of smoking tobacco than living in mountainous regions. This finding resonates with prior studies in Nepal^50,51. The authors argued that mountainous regions often experience harsh environmental conditions, such as cold temperatures, which might encourage smoking as a perceived way to stay warm or to deal with stress and isolation. However, agricultural practices in hilly and terai areas may prioritize subsistence farming or other economic activities over tobacco cultivation, potentially limiting the local availability of tobacco products.

Being employed increased the odds of tobacco smoking among respondents. Men who were involved in some form of work were more likely to smoke tobacco. Similar findings have been reported in India³⁰. It can be explained that work-related stress, particularly among men may drive them to use tobacco as a coping mechanism to manage stress. Work environments where tobacco smoking is normalized, such as in certain industries or among colleagues who smoke, can further encourage this behavior. Additionally, men with disposable income from employment might find tobacco products more affordable and accessible, contributing to higher usage. The link between job strain and increased smoking intensity in adult workers is well-established^52,53.

Our study also found that men in the richest category had higher odds of smoking tobacco than their counterparts. This association is consistent with that of Terefe et al.³⁶ in East Africa. Conversely, it contradicts the findings of Islam et al.³⁰ in South Asia, Nketiah-Amponsah et al.⁴⁰ in Ghana, Gutema et al.⁵⁴ in Ethiopia, and Pénzes et al.³⁷ in Hungary. The possible explanation could be that wealthier men have greater disposable income, allowing them to purchase tobacco products without financial strain, even as prices increase due to taxation or other economic deterrents. However, the contradictory findings may reflect contexts where economic constraints limit tobacco smoking among lower-income individuals, or where the financial burden of purchasing tobacco outweighs its perceived benefits⁵⁵.

Findings from the study also revealed that those who consume alcohol were more likely to smoke tobacco. This finding aligns with studies from China⁴⁵, England⁵⁶ and India⁵⁷. This supports the theory of cross-tolerance, which suggests that the consumption of one substance can affect tolerance to another through shared neurological pathways, especially in the brain’s reward system^58,59. This phenomenon is observed with substances such as alcohol and tobacco; using one can increase tolerance to and the effects of the other, potentially leading to heightened consumption of both substances.

Furthermore, we found that men reporting good or moderate health status had increased odds of smoking tobacco than those who reported having poor health status. This finding is consistent with that of Tonnesen⁶⁰ in Copenhagen. The plausible explanation for this finding could imply that individuals who perceive themselves as healthy may not prioritize the risks associated with smoking or may engage in smoking as a form of social behavior.

To strengthen tobacco control in Nepal, the government and relevant stakeholders must rigorously enforce existing regulations and adopt new policy measures that are proven effective elsewhere. These may include plain packaging laws, comprehensive bans on indirect advertising, and regulation of emerging tobacco products such as e-cigarettes and smokeless tobacco which are currently underregulated and increasingly used, especially among the youth. Subsequently, a periodic review of Nepal’s adherence to the Tobacco Products Act, 2011 provisions should be instituted to ensure that commitments are being effectively implemented.

Additionally, while the country is doing well in mandatory graphic warnings covering 90% of the front and back of tobacco product packages and its ban on the 2011 act which prohibits all forms of tobacco advertising and promotion including in print, electronic media, and at points of sale⁶¹, challenges remain in enforcement and compliance. Continued efforts must be put in place to strengthen implementation, increase public awareness, and address industry interference to achieve the desired public health outcomes.

For tobacco smokers, public health strategies must address both pull and push factors driving tobacco use. Increasing awareness of health risks through media campaigns as well as increasing tobacco prices through taxation are deterrents that need greater reinforcement. Additionally, the government of Nepal must expand cessation support services such as counseling and nicotine replacement therapy to aid those seeking to quit. Furthermore, preventive strategies must prioritize community-based education programs to emphasize the health consequences of tobacco use, especially among non-smokers.

The strength of this study lies in its use of a sizable, nationally representative sample drawn from a high-quality survey with excellent response rates. The study benefitted from thorough interviewer training and the use of validated questionnaires, providing sufficient statistical power to analyze the predictors of tobacco smoking among men in Nepal. Standardized procedures employed in the DHS contribute to the robustness of the study, ensuring both internal and external validity. Moreover, the data were weighted for analysis to enhance the generalizability of the results.

Despite these strengths, the study has some notable limitations which must be acknowledged. As the data were derived from secondary sources and the surveys were cross-sectional, no conclusions about temporality can be drawn. Data on factors associated with smoking were based on self-reports and could not be verified through records, given that this was a cross-sectional study. As a result, the data may be subject to recall and interviewer biases. Additionally, cultural and perception-related factors that could influence tobacco smoking were not accounted for in this study. Furthermore, the prevalence of tobacco smoking may be underreported due to societal stigma and individuals’ tendency to be more reserved.

The study found that the prevalence of tobacco smoking among men in Nepal was quite significant. Key factors influencing tobacco smoking included age, education, occupation, alcohol consumption, self-reported health status, ecological region, and wealth index. These findings highlight the need for the government of Nepal to enforce existing laws, adopt new policies, and regulate emerging products. Special attention should be directed toward addressing the demographic and socio-economic disparities among men to ensure tailored and impactful public health strategies.