Tobacco smoking is a leading preventable cause of death globally (1), claiming over eight million lives each year, with an increasing burden particularly in low- and middle-income countries (LMICs) (2). In Myanmar, an LMIC, smoking prevalence remains high, especially among men (36.5% in 2022) (3), significantly contributing to the country’s burden of both communicable and noncommunicable diseases (NCDs), including cardiovascular conditions, cancers, and chronic respiratory illnesses, and infectious diseases such as tuberculosis (4). Vascular diseases such as coronary artery disease (5), stroke (6), and peripheral artery disease (7) are notably vulnerable to smoking-related harm. Industrial workers represent a high-risk group for tobacco use due to occupational environments characterized by physical and psychosocial stressors (8, 9), peer influence (10), and limited access to health promotion and cessation programs (11). This increased prevalence heightens their risk for lung diseases and cardiovascular issues (12), exacerbating the overall NCD burden, while also leading to productivity losses, workplace accidents, and rising healthcare costs.

Although behavioral models like the Theory of Planned Behavior (TPB) (13), the Transtheoretical Model (TTM) (14, 15), and Social Cognitive Theory (SCT) (16) have been employed to understand health behaviors, the Health Belief Model (HBM) (17–19) remains one of the most widely used in health education (HE), disease prevention, and community interventions. The HBM’s focus on health motivation and preventive behavior makes it particularly suitable for exploring smoking behaviors among at-risk populations such as industrial workers (20). It comprises six key constructs (18): perceived susceptibility, perceived severity, perceived barriers, perceived benefits, self-efficacy, and cues to action. Definitions of these constructs are provided in Supplementary material 1 (18). However, research applying the HBM in Myanmar’s industrial settings is scarce. Addressing this gap is essential for developing targeted interventions in occupational contexts where health promotion resources are often limited.

The adapted conceptual framework guiding our study (Figure 1) categorizes predictors into two domains: modifying factors and individual beliefs, reflecting their theorized roles in shaping the likelihood of taking health-related action. While the framework outlines potential intermediary pathways, our hypotheses focused on the direct effects of these domains on smoking behavior outcomes. Additionally, quit smoking intention, though not a core HBM construct, was included as a motivational modifying factor due to its demonstrated role in behavior change (21). Thus, our study aimed to examine the associations between modifying factors, individual beliefs, and smoking behaviors, including quit attempts and smoking intensity, among industrial workers in Myanmar.

Table 1 presents the participant characteristics (modifying factors) of the HBM constructs. All respondents were males (N = 292) with a median age of 28.00 years. More than half were married (54.1%), and the majority (78.1%) had completed high school or higher education. The median monthly income was 18.00 per 10,000 Myanmar Kyats (MMK), equivalent to approximately 180,000 MMK (around 86 USD; based on the Central Bank of Myanmar exchange rate of 1 USD = 2,100 MMK on January 30, 2025) (35). Respondents had a median smoking initiation age of 18.00 and a duration of 10.00 years. Over half of the respondents (57.5%) reported no intention to quit smoking. The health knowledge score ranged from 0 to 3, with a median score of 2.00.

Table 2 displays individual beliefs based on HBM constructs. The median scores were 2.64 [41.0% of the Percentage of Maximum Possible Score (POMP)] for perceived susceptibility, 2.90 (47.5%) for perceived severity, 3.00 (50.0%) for perceived barriers, and 2.80 (45.0%) for perceived benefits, all measured on scale of 1–5. The median self-efficacy score was 7.00 (29.2% POMP) on a scale of 0–24.

Table 3 presents the smoking behavior outcomes (likelihood of action) related to the HBM constructs. The results indicate that a significant majority of participants (90.4%) have not attempted to quit smoking. Additionally, the respondents are nearly evenly split in terms of smoking intensity, with 52.4% classified as low-intensity smokers and 47.6% as high-intensity smokers.

Table 4 interprets association between HBM constructs and smoking behavior outcomes. Higher self-efficacy was significantly associated with increased odds of quit attempts in the crude model (OR = 1.11, 95% CI: 1.01, 1.23, p = 0.029) but lost significance in the adjusted model (p = 0.163). For smoking intensity, higher health knowledge showed a significant reduction in the likelihood of being a high-intensity smoker in both the crude (OR = 0.65, 95% CI: 0.49, 0.88, p = 0.005) and adjusted models (AOR = 0.53, 95% CI: 0.38, 0.75, p < 0.001). Higher self-efficacy was also significantly associated with lower odds of being a high-intensity smoker in the adjusted model (AOR = 0.93, 95% CI: 0.86, 0.99, p = 0.044).

This study examined the associations of modifying factors and individual beliefs with smoking behaviors, specifically quit attempts and smoking intensity, among 292 industrial workers in Myanmar, using cross-sectional baseline data from our longitudinal quasi-experimental study (17). The findings provide contextual insights into smoking behaviors in a low-resource occupational setting, independent of any intervention effects.

Compared to similar studies conducted in community or clinical settings, our descriptive analysis revealed low levels of health knowledge (26, 36, 37), perceptions (26, 38, 39), self-efficacy (40), as well as a low prevalence of quit attempts (41). These deficits may be attributed to contextual barriers, such as limited access to health information, time constraints inherent to industrial workers, and a workplace culture that does not prioritize HE. The predominance of physical labor and the social acceptability of tobacco use in such environments likely contribute to a reduced perception of smoking-related risks and diminished motivation to quit. These findings point to a critical need for HE tailored to industrial contexts, where the HBM-based approaches may be more impactful when grounded in real-world behavioral baselines. Although our broader study (17) demonstrated the effectiveness of an HBM-guided intervention combing HE and mobile phone SMS in improving smoking-related health knowledge, perceptions, and self-efficacy, this cross-sectional analysis provides distinct, pre-intervention insights into behavioral predictors. To enhance future intervention design, studies should consider incorporating environmental and social predictors, such as peer influence, media exposure, tobacco accessibility, and workplace culture, into the behavioral framework.

Regarding factors associated with quit attempts, the intention to quit was not significantly associated (p = 0.656), which contrasts with theoretical expectations derived from models like the TPB. This may reflect the influence of contextual elements such as economic insecurity, insufficient institutional support, and sociocultural norms, all of which may undermine the translation of intention into action. In this occupational context, individual beliefs may be insufficient drivers of behavior in the absence of enabling conditions. This points to the need for multi-level approaches in future interventions, those that combine individual behavior change strategies with supportive structural changes, including employer-endorsed cessation programs and occupational health policies that discourage smoking.

Notably, none of the HBM perception-based constructs were significantly associated with quit attempts in either the bivariate or multivariable models. This finding contrasts with studies conducted during acute public health threats, such as the COVID-19 pandemic (42, 43), where stronger perceptions of risk prompted more cessation efforts. In contrast, among industrial workers in a relatively stable setting, perceived severity and susceptibility related to smoking may not seem important enough to trigger immediate behavior change. This suggests that in such settings, interventions must actively cultivate a stronger sense of urgency, using approaches like graphic health warnings, narrative storytelling, testimonials, and integration of risk communication into occupational health programs. Future studies should consider longitudinal designs to better assess how these perception shifts evolve and how they influence long-term quit behavior.

Self-efficacy emerged as a significant factor associated with quit attempts (p = 0.029) in the bivariate analysis, consistent with the HBM framework and prior studies (44, 45), which affirm the central role of confidence in behavior change. Notably, the effect size (COR = 1.11, 95% CI: 1.01–1.23) indicates a modest increase in the likelihood of quit attempts per unit increase in self-efficacy, suggesting that even small improvements in self-efficacy may be meaningful. However, in the multivariable model, this association was not statistically significant (p = 0.163), indicating that self-efficacy may not be independently associated with quit attempts when broader contextual factors are accounted for. This contrasts with Spek et al.’s (31) validation of SASEQ, and our null finding could be partly explained by the lack of structured workplace cessation programs in the Myanmar context. Even so, the bivariable effect size highlights the potential practical relevance of self-efficacy interventions. Without structural and institutional support, high self-efficacy alone may be insufficient to facilitate quit behavior, particularly in environments that do not actively support smoking cessation. In this regard, workplace-based policies that enable and reinforce cessation, such as smoke-free zones, cessation support groups, and peer champion models, may amplify the effectiveness of individual-level interventions.

With respect to smoking intensity, a different pattern emerged. The borderline significance of intention to quit (COR = 1.48, 95% CI: 0.93–2.36, p = 0.09) in the bivariable analysis suggests a potentially meaningful effect size, indicating that motivation could still influence cigarette consumption, even if it does not translate into full cessation attempts. While perception constructs did not show a significant relationship with smoking intensity, this may reflect the normalization of tobacco use in the cultural and social fabric of adult industrial workers, a contrast to findings from a U.S. study (46) among adolescents or populations in settings with more prominent anti-smoking messaging. These differences highlight the need for locally tailored HE that addresses cultural acceptance of smoking and highlights immediate occupational health risks, especially in countries with limited anti-smoking enforcement.

Multivariable analysis showed that both health knowledge (AOR = 0.53, 95% CI: 0.38–0.75, p < 0.001) and self-efficacy (AOR = 0.93, 95% CI: 0.86–0.99, p = 0.044) were significantly associated with lower smoking intensity. These effect sizes indicate that even modest improvements in health knowledge or self-efficacy are linked with meaningful reductions in smoking, reinforcing their practical and clinical significance. This corroborates evidence from previous studies indicating that greater health knowledge of smoking risks (47) and higher self-efficacy (45) are linked with reduced tobacco consumption. These findings reinforce the value of HE programs that not only provide accurate risk information but also empower individuals to act. Even in the absence of complete cessation, reductions in smoking intensity are meaningful public health outcomes, particularly for populations at high risk for NCDs, highlighting the real-world importance of these interventions. Evaluating the long-term impact of reduced smoking on NCD outcomes should be a priority for future research, aligning with health system planning and resource allocation in LMICs. Furthermore, the successful integration of mobile phone-based interventions like SMS, demonstrated in our broader study (17), offers a promising tool to sustain behavior change, especially in settings with limited face-to-face access.

Our baseline study highlights the intricate interactions between modifying factors and individual beliefs influencing smoking behaviors in industrial workers in Myanmar. While health knowledge and self-efficacy are pivotal in reducing smoking intensity, their effect on quit attempts is limited without broader contextual support. These findings highlight the need for theory-informed interventions that integrate behavioral strategies with structural and environmental support to effectively promote smoking cessation in occupational settings.