Introduction

The construction sector is globally recognized as one of the most hazardous and high-risk work environments, where workers are routinely exposed to diverse physical, chemical, ergonomic, and psychosocial hazards. The International Labour Organization (ILO) highlights the substantial risks faced by construction workers, including falls from height, being struck by objects, and exposure to harmful substances all of which contribute to a significant incidence of occupational injuries and fatalities1. Furthermore, research indicates that construction tasks are physically demanding, often involving repetitive motions, awkward postures, and heavy liftin factors that signicantly increase the likelihood of musculoskeletal disorders (MSDs)2,3.

Occupational stress among construction workers is intensified by factors such as demanding work schedules, job insecurity and challenging environmental conditions4. Prolonged exposure to such stressors has been shown to adversely affect mental health, contributing to elevated rates of anxiety, depression and burnout5. A Malaysian study reported that more than of half of construction professionals experienced stress (54.9%), high prevalence rates of anxiety (48.5%) and depression (37%), and with gender identified as a significant factor influencing these outcomes6. Similar trends are evident in other high-risk occupations, such as firefighting, where post-traumatic stress disorder (PTSD) is prevalent7.

Demographic and work-pattern factors including shift rotations, job tenure and socioeconomic background have been shown to significantly influence the prevalence and severity of occupational stress8 .Irregular work hours can disrupt circadian rhythms, impair physiological recovery and increases vulnerability to metabolic stress and mental health disorders9.These findings reinforce growing evidence that occupational well-being is shaped not only by immediate workplace hazards but also by broader lifestyle, health and social factors10. The Job Demand–Resource (JD-R) model provides a theoretical framework for understanding these relationships, positing that excessive job demands, when not counterbalanced by adequate resources such as rest breaks, job control, and social support, increase the risk of burnout and compromised health outcomes11.

In addition to psychosocial stressors, construction workers are frequently exposed to hazardous substances such as silica dust and asbestos, which are associated with chronic respiratory diseases and other severe health effects1,12. Addressing these risks requires effective safety management systems, enhanced hazard awareness, and targeted training programmes as critical strategies to protect worker health13.

Despite increasing recognition of these occupational hazards, empirical research in Malaysia has rarely examined the intersection of occupational stress, quality of life, and sociodemographic determinants in construction settings. This study addresses this gap by assessing occupational stress and quality of life among Malaysian construction workers, with particular attention to the influence of sociodemographic variables. Using validated instruments, the Occupational Stress Index (OSI) and the World Health Organization Quality of Life-BREF (WHOQOL-BREF) and situating the analysis within the JD-R framework, this research aims to provide robust, evidence-based insights to inform occupational health policies and targeted interventions in the construction sector.

Literature review

The construction industry is widely regarded as one of the most hazardous sectors, exposing workers to diverse occupational risks, including physical injuries, chemical agents, ergonomic stressors, and psychosocial pressures14. Studies conducted in affluent nations such as United States, Australia and the United Kingdom consistently identify high job demands, irregular work hours, job insecurity and physically strenuous tasks as primary contributors to adverse mental health. These stressors are linked to higher incidence of anxiety, depression, burnout, and occupational musculoskeletal disorders (OMSDs)15.

The Job Demand–Resource (JD-R) model provides a comprehensive framework for examining these dynamics. It posits that high job demands, when coupled with insufficient resources, exacerbate burnout and diminish quality of life16.Conversely, job control, social support, and opportunities for professional growth can mitigate these negative outcomes and enhance worker well-being11,17,18. This paradigm has been extensively applied in occupational stress research within developed contexts, where psychosocial work conditions have been correlated with various adverse health consequences.

In developing countries, poor enforcement of occupational safety regulations and limited access to protective measures contribute to a precarious work environment. Research in India, Pakistan and Nigeria has revealed that construction workers, particularly those in informal or subcontracted roles, face compounded vulnerabilities, including physical hazards, low wages, lack of social protection, and extended working hours19,20,21.Among migrant labourers, additional stressors including wage delays, overcrowded housing, social isolation, and language barriers further exacerbate occupational stress and erode quality of life2223. Despite well-documented evidence of these risks, empirical research integrating psychosocial and quality-of-life dimensions in such contexts remains scarce.

Malaysia’s construction sector is a major economic driver, but it continues to record high injury and fatality rates24. Local research has largely focused on safety compliance, accident causation, and physical hazards, with comparatively fewer studies addressing psychosocial stressors. Evidence from Tharumalay et al. (2020) and Muhammad Badruddin et al. (2025) indicates that irregular working hours, income instability, and insufficient rest are significant determinants of occupational stress and poor mental health. However, these studies predominantly examine skilled or supervisory personnel, leaving underexplored the experiences of general labourers, who face the highest exposure to both physical and psychosocial risks25,26. A defining characteristic of Malaysia’s construction workforce is its heavy reliance on migrant labour from countries such as Indonesia, Bangladesh, Nepal, and Myanmar. Predominantly employed in low-skilled positions, these workers frequently face substantial challenges, including job insecurity, inadequate healthcare access, and substandard living conditions27. Many experience fluctuating wages and extended working hours, conditions that contribute to poor nutrition and heightened occupational stress, ultimately undermining their overall quality of life28,29,30.

Although global research has advanced understanding of the pathways through which occupational stressors affect worker well-being, there is a paucity of empirical evidence from Malaysia, particularly within the residential construction subsector and among migrant labourers. Existing studies rarely employ validated instruments such as the Occupational Stress Index (OSI) and WHO Quality of Life-BREF (WHOQOL-BREF) in conjunction with theoretical models like the JD-R framework. This methodological gap constrains the development of a nuanced understanding of the multidimensional nature of occupational stress and its implications for quality of life among the most vulnerable groups in the construction workforce.

Materials and methods

Study design

A cross-sectional study was conducted across three residential construction sites in Johor and Kuala Lumpur, purposively selected to represent variation in project scale (high-rise, medium rise, landed housing), workforce composition and management practices. This diversity allowed for meaningful comparisons of occupational stress and quality of life across distinct operational contexts while maintaining industry-specific relevance.

Sampling

A two-stage sampling process was applied. In Stage 1, purposive sampling ensured inclusion of workers meeting the following eligibility criteria: aged ≥ 18 years, employed for at least three months at the current site, able to understand Bahasa Malaysia or English, and willing to provide written informed consent. In Stage 2, convenience sampling was employed to recruit participants available during scheduled site visits, due to restricted access windows and fluctuating workforce attendance.

The final sample comprised 85 workers across multiple job categories, reflecting varied occupational exposures. Although modest, this sample size aligns with previous construction sector studies facing similar site-specific recruitment challenges9,10. A post-hoc power analysis using G*Power 3.1 confirmed adequacy for detecting medium effect sizes (f² = 0.15) at α = 0.05 and power = 0.80, with a minimum required sample of 77 participants.

Instrumentation and psychometric properties

Data were collected using a structured questionnaire comprising four Sect. (94 items). Sections A and D were author-developed and did not require permission. Sections B and C employed validated instruments (OSI and WHOQOL-BREF). The OSI is widely used for academic research with proper citation and the WHOQOL-BREF is freely available from the World Health Organization for research and publication.

  • Section A gathered demographic, occupational, lifestyle, and self-reported health information.

  • Section B included 46 items from the validated Occupational Stress Index (OSI) which measures work-related stress using a 5-point Likert scale (1 = strongly disagree to 5 = strongly agree)31.

  • Section C consisted of 26 items from the World Health Organization Quality of Life–BREF (WHOQOL-BREF)32 also scored on a 5-point Likert scale (1 = not at all/very dissatisfied to 5 = an extreme amount/very satisfied) to assess participants perceived quality of life.

  • Section D included additional items related to occupational exposures and psychosocial conditions.

The OSI has demonstrated high reliability in construction worker populations (Cronbach’s α = 0.78–0.86). In this study, α = 0.831. The WHOQOL-BREF has been validated internationally with domain α values ranging from 0.66 to 0.84 33.

Translation and expert validation

The English questionnaire was translated into Bahasa Melayu through a forward–backward translation process was conducted by bilingual experts to ensure conceptual and linguistic accuracy. A linguistic specialist reviewed the final version for semantic clarity. Pre-testing with 10 non-sample construction workers confirmed cultural relevance and ease of comprehension, resulting in minor vocabulary adjustments in the bilingual (English–Bahasa Melayu) version.

Expert validation was undertaken to strengthen content validity and ensure alignment with the study’s conceptual framework. Three subject matter experts, two with extensive experience in occupational health research and one specialising in ergonomics were purposively selected based on professional qualifications, peer-reviewed publications, and practical expertise in construction workplace risk assessments. The review covered four domains:

  • Content Relevance: Items were systematically examined to confirm alignment with the intended constructs of occupational stress, quality of life, and occupational exposures. All but one item achieved perfect relevance scores (I-CVI = 1.00) and the remaining item scored 0.80 indicating moderate but acceptable relevance.

  • Construct Coverage: The questionnaire was determined to thoroughly encompass all pertinent dimensions. The OSI component effectively assessed psychosocial stresses, the WHOQOL-BREF evaluated physical, psychological, social, and environmental dimensions of quality of life, and Section D provided context by incorporating occupational exposure variables such as heat, noise, and ergonomic risks. No content deficiencies were detected.

  • Clarity and Cultural Appropriateness: The language was clear and understandable for both Malaysian and expatriate staff. Minor vocabulary modifications were proposed for three items in the bilingual (English–Bahasa Melayu) version to enhance clarity and maintain semantic equivalence across languages.

  • Scaling Appropriateness: The 5-point Likert scale was deemed suitable for measuring agreement, perception, and intensity, with both positively and negatively phrased items helping to reduce response bias.

Data collection procedures

The final version of the questionnaire was administered on-site during break periods in a quiet, designated area. Trained enumerators helped respondents and clarified any questions. Each participant took approximately 20 to 30 min to complete the questionnaire. Written informed consent was obtained from all respondents prior to participation.

Data analysis

Statistical analyses were performed using Statistical Package for the Social Sciences (SPSS) version 27. Descriptive statistics (frequencies, percentages, means, and standard deviations) summarised sociodemographic characteristics, occupational stress levels, and quality of life scores. Differences in stress levels and quality of life across construction sites were examined using one-way ANOVA, with Shapiro–Wilk and Levene’s tests applied to verify normality and homogeneity of variances. Pearson’s correlation assessed the strength and direction of linear associations between continuous variables, specifically OSI and WHOQOL-BREF scores. Multiple linear regression identified significant predictors of occupational stress and quality of life, adjusting for potential confounders including age, income, body mass index (BMI), and migrant status. Effect sizes (η² for ANOVA; Cohen’s f² for regression) were computed, and regression diagnostics assessed multicollinearity, normality, and homoscedasticity to ensure model validity. Statistical significance was set at p < 0.05.

Ethical approval

This study was approved by the Secretariat Ethics Committee UKM (Human) at Universiti Kebangsaan Malaysia (approval no. PPI/111/8/JEP-2024-844). All procedures complied with relevant guidelines and regulations, and informed consent was obtained from all participants. The study adhered to the ethical principles outlined in the Declaration of Helsinki.

Results

Risk factors based on socio-demographic

Table 1 shows that the majority of respondents were non-Malaysian (70.6%), while Malaysian citizens comprised 29.4%. Most respondents were male (95.3%), with females representing only 4.7%. In terms of body mass index (BMI), more than half (52.9%) were classified as overweight or obese (BMI > 25 kg/m²), whereas only 23.5% had a normal BMI. Regarding physical activity, 36.5% reported being moderately active, and 45.9% engaged in regular exercise. Notably, there was a high prevalence of smoking (87.1%) and alcohol consumption (56.5%). More than half of the respondents (58.8%) reported experiencing sleep difficulties, which may exacerbate occupational stress levels.

Table 1 Risk factors based on socio-demographic.
Full size table

As presented in Table 2, the majority of respondents (70.6%) were general labourers, with a mean household income of RM 2,302.94 (SD = 1,212.85) and an average of three dependents (SD = 1.46). Long working hours were common, with 80.0% exceeding eight-hour shifts, and 96.5% reporting frequent overtime (mean = 14 days/month, SD = 7.75). Fatigue levels were high, with 55.3% reporting severe fatigue. Additionally, 42.4% of respondents reported heavy energy expenditure, and 20.0% reported very heavy workloads.

Table 2 Risk factors based on working information.
Full size table

Occupational stress level and quality of life

Occupational stress, measured using the Occupational Stress Index (OSI), was categorized as low (29.4%), moderate (22.4%), and high (48.2%) (Table 3). Quality of life, assessed using the WHOQOL-BREF, was classified as low (55.3%) and high (44.7%) (Table 4). A substantial proportion of workers experiencing high occupational stress also reported lower quality of life.

Table 3 Occupational stress level.
Full size table
Table 4 Quality of life.
Full size table

Comparison across construction sites

The present study involved three residential construction sites located in Johor and Kuala Lumpur, each selected based on their active project status, accessibility, and management approval for participation. Although all three sites shared a common construction type residential housing key differences were noted in their scale and workforce composition. Site A was a high-rise residential development employing approximately 90 workers. Site B comprised medium-rise housing units with around 65 workers while Site C focused on landed residential units and had approximately 85 workers. These site-specific variations in project scale and workforce dynamics provided a suitable context for comparing occupational stress and quality of life across different working environments within the same sector.

A one-way ANOVA revealed significant differences in OSI scores across sites, F (2, 82) = 76.81, p = 0.001, η² = 0.652. Site C recorded the highest mean OSI score (165.83 ± 30.38), followed by Site A (147.50 ± 18.07), and Site B (85.70 ± 19.35). Significant differences were also observed in WHOQOL-BREF scores, F (2, 82) = 19.80, p = 0.001, η² = 0.326. Site B reported the highest mean quality-of-life score (76.52 ± 17.76), followed by Site C (51.23 ± 19.63) and Site A (48.00 ± 11.17) (Table 5).

Table 5 Occupational stress level and quality of life among three construction sites.
Full size table

Association between sociodemographic, lifestyle, health, and occupational factors with occupational stress and quality of life

Table 6 showed the Pearson’s correlation analysis that revealed a significant negative correlation between body mass index (BMI) and occupational stress levels (r = -0.306, p = 0.006), suggesting that overweight and obese individuals may experience lower stress compared to those with a normal BMI. Conversely, a weak but significant positive correlation was observed between BMI and quality of life (r = 0.272, p = 0.015), indicating that workers with higher BMI reported slightly better well-being. Household income exhibited a weak negative correlation with occupational stress (r = -0.354, p = 0.001) and a positive correlation with quality of life (r = 0.546, p < 0.001) reinforcing that financial stability plays a crucial role in reducing workplace stress and improving well-being. Additionally, a weak positive correlation was identified between the number of dependents and occupational stress (r = 0.233, p = 0.037), suggesting that financial and familial responsibilities may contribute to heightened stress levels. However, length of service demonstrated a weak positive correlation with quality of life (r = 0.237, p = 0.034), indicating that workers with longer job tenure tend to report better well-being, possibly due to increased job familiarity, stability, and coping mechanisms. Further analysis revealed that break duration was inversely correlated with occupational stress (r = -0.289, p = 0.009) and positively correlated with quality of life (r = 0.282, p = 0.011), suggesting that sufficient rest periods may serve as a protective factor against workplace stress while enhancing overall well-being. A moderate positive correlation was observed between the number of working days and occupational stress (r = 0.524, p < 0.001), highlighting the adverse effects of excessive work hours on mental health. Conversely, an inverse correlation was found between working days and quality of life (r = -0.372, p < 0.001) emphasizing that prolonged work schedules contribute to deteriorating well-being.

Table 6 Occupational stress level and quality of life between sociodemographic data, lifestyle, health status and working information.
Full size table

Association between total score of OSI and total score of WHOQOL-BREF

A strong negative correlation was observed between total OSI scores and WHOQOL-BREF scores (r (85) = -0.622, p < 0.001) (Table 7), indicating that higher occupational stress is significantly associated with lower quality of life. Figure 1 visually depicts this relationship between occupational stress and well-being among construction workers.

Table 7 Relationship between total score OSI and total score WHOQOL-BREF.
Full size table

Multiple linear regression results

A multiple linear regression analysis was conducted to identify significant predictors of occupational stress (OSI) and quality of life (WHOQOL-BREF) with age, household income, BMI and migrant status entered simultaneously. The results for Model 1 (OSI) and Model 2 (WHOQOL-BREF) are presented in Table 8.

Fig. 1
figure 1

Correlation WHOOQOL-BREF With OSI score.

Full size image

Model 1-occupational stress (OSI)

The model significantly predicted OSI scores, F(4, 80) = 15.92, p < 0.001, with R² = 0.443 (adjusted R² = 0.417), indicating that 44.3% of the variance in occupational stress was explained by the predictors. Household income was a significant negative predictor (β = –0.352, p < 0.001), indicating that higher income was associated with lower stress levels. Working days per month was a strong positive predictor (β = 0.495, p < 0.001), while BMI was a weaker but significant negative predictor (β = –0.210, p = 0.036). Migrant status was not statistically significant (β = 0.098, p = 0.183). The overall effect size, Cohen’s f² = 0.80, indicated a large effect.

Model 2-quality of life (WHOQOL-BREF)

The second model also reached statistical significance, F(4, 80) = 18.71, p < 0.001, with R² = 0.483 (Adj. R² = 0.459). Household income was the strongest positive predictor (β = 0.511, p < 0.001), followed by break duration (β = 0.235, p = 0.019). Working days per month was negatively associated with quality of life (β = –0.327, p< 0.001), and BMI showed a weaker but significant positive association (β = 0.176, p = 0.048). The model’s effect size, Cohen’s f² = 0.93, indicated a large effect.

Discussion

This study generated three principal findings among Malaysian construction workers: (1) a strong negative association between occupational stress and quality of life, (2) significant predictive effects of socioeconomic factors particularly income level and number of working days—on both stress and quality of life, and (3) notable inter-site differences in stress levels despite similar industry contexts. Together, these findings provide empirical support for the Job Demand–Resource (JD–R) model34 while also highlighting context-specific determinants within Malaysia’s construction labour force.

The inverse relationship between occupational stress and quality of life aligns with the JD–R framework, wherein elevated job demands coupled with inadequate recovery resources contribute to diminished well-being28,29. Although previous research has documented the deleterious effects of excessive workload on psychological and physical health in construction settings35,36, only a limited number of studies have quantified this relationship using validated measures such as the Occupational Stress Index (OSI) and WHOQOL-BREF. The strength of the observed correlation indicates that the combined effects of physical hazards, psychosocial pressures and restricted rest opportunities may impose a greater burden in the construction sector compared with less physically demanding occupational groups37,38.

Socioeconomic position emerged as a critical factor in shaping stress and quality-of-life outcomes. Higher household income was associated with lower stress and improved quality of life, likely reflecting enhanced access to healthcare, better nutrition, and greater financial security, all of which facilitate physical recovery and psychological resilience35,39. Financial stability may facilitate more measured decision-making on-site, reducing unsafe behaviours and thereby lowering accident risk and improving productivity40 [Ray, 2021]. These findings corroborate earlier studies in both construction and other labour-intensive industries, which have linked income disparities to variations in occupational health outcomes41.

The analysis revealed an inverse association between body mass index (BMI) and stress levels which warrants further examination. Rather than indicating better health, This pattern is likely attributable to occupational stratification, wherein individuals in supervisory or administrative roles often associated with higher BMI due to reduced physical demands experience lower stress owing to increased job security, autonomy, and access to organisational resources38,42. This interpretation is consistent with prior work suggesting that job role and employment conditions can moderate the health effects of physical workload28.

The number of working days per month also demonstrated a significant positive association with stress, in line with cumulative fatigue theory43 and previous evidence from construction workforces41. Longer rest intervals particularly those allowing for recovery in hot and humid climates were positively associated with improved quality of life. This finding reinforces the importance of contextually adapted work–rest cycles in tropical environments, where climatic stress can amplify physiological strain37.

The significant variability in stress levels between sites underscores the influence of organisational culture and managerial practices as potential moderating factors. Workplaces characterised by punitive supervision, poor communication, and heavy reliance on migrant labour exhibited higher stress compared to sites with supportive management, enforced safety protocols, and culturally sensitive communication28,44. This observation aligns with the JD–R model’s proposition that organisational resources such as leadership style, participatory decision-making, and safety climate buffer the negative effects of high job demands34.

These findings carry important practical implications for workplace policy and management strategies. At the organisational level, construction companies should implement structured welfare programmes that integrate fatigue management into occupational safety systems [Yusof et al., 2019], provide shaded rest areas39 and offer multilingual psychological support to accommodate the diverse linguistic backgrounds of workers44. Regular stress monitoring using validated tools should be embedded into safety audits43 to enable early detection of psychosocial risks. At the policy level, regulatory bodies should enforce maximum weekly working hours40 ensure transparent overtime compensation37 and mandate employer-provided mental health resources35. In addition, national guidelines should explicitly address climate-adapted work–rest scheduling41 to mitigate heat-related stress.

Table 8 Multiple linear regression predicting occupational stress (OSI) and quality of life (WHOQOL-BREF).
Full size table

In summary, this study advances understanding of occupational stress and quality of life in Malaysia’s construction sector by demonstrating the interplay between workload intensity, socioeconomic conditions, organisational culture, and environmental factors. The integration of empirical findings with the JD–R model offers a robust conceptual framework for both research and practice. Implementing targeted practical method grounded in these findings could substantially improve worker well-being, reduce preventable strain, and enhance long-term productivity. Future research should build upon this work by adopting longitudinal designs to capture stress trajectories across project phases28 employing mixed-method approaches to explore culturally specific stressors29 and examining organisational-level moderators such as leadership style and work culture to guide evidence-based and context-sensitive interventions34.

Limitation and future research

The limitations identified in the study of occupational well-being among Malaysian construction workers provide a basis for future research directions while guiding practical implications. These limitations influence the interpretation of findings and warrant caution in drawing generalizable conclusions. The use of a relatively small, geographically constrained sample obtained through non-probability sampling reduces the representativeness of the results, thereby limiting their applicability to the wider population of construction workers across different regions and sectors. Reliance on self-reported questionnaires introduces potential social desirability and response biases as participants may have provided responses they deemed socially acceptable rather than accurately reflecting their experiences. Moreover, the complexity of the instruments used to assess occupational stress and quality of life may have led to misinterpretation by some respondents, potentially affecting the reliability of the data. The cross-sectional design further limits the ability to establish causal relationships between occupational stressors and quality of life, restricting the analysis to observed associations rather than causal inferences.

Future research should address these methodological constraints by employing larger and more diverse samples, incorporating workers from multiple geographic regions and a wider range of construction sectors, to improve external validity. Longitudinal study designs are particularly recommended to monitor changes in stress and quality of life across different project phases, thereby capturing both acute and cumulative effects over time. The adoption of mixed-method approaches integrating quantitative surveys with qualitative interviews or focus groups would yield richer, contextually grounded insights into worker experiences, particularly those shaped by cultural and organisational factors. Comparative analyses across residential, commercial, and infrastructure projects could help identify sector-specific stressors and resilience factors. Furthermore, investigating organisational-level moderators such as leadership style, safety climate, and workplace culture would deepen the explanatory framework and facilitate the development of more targeted interventions. Such evidence would be instrumental in informing regulatory reforms, guiding worker welfare programmes, and integrating on-site stress monitoring systems, ultimately contributing to improved occupational health, safety, and productivity in the construction industry.

Conclusion

This study provides robust empirical evidence of a significant negative correlation between occupational stress and quality of life (r = − 0.622, p < 0.001) among Malaysian construction workers, underscoring the critical role of psychosocial well-being in this high-risk sector. The findings further indicate that key determinants including working days per month, household income, break duration, and site-specific conditions substantially influence both stress levels and overall quality of life. These results extend the application of the Job Demand–Resource (JD–R) model by affirming its relevance in physically demanding, hazard-intensive environments, where both job demands and the adequacy of recovery resources play decisive roles in shaping worker well-being. From a theoretical standpoint, this research reinforces the JD–R framework’s predictive capacity in construction settings, highlighting the compounded effects of physical workload, socio-economic pressures, and organisational climate on occupational health outcomes. From a practical perspective, it offers an evidence-based roadmap for intervention. Policy measures should include revising regulations to enforce maximum working hours, ensuring transparent and equitable overtime compensation, and institutionalising climate-adapted rest periods. Organisations are encouraged to invest in comprehensive worker welfare programmes incorporating multilingual psychological support, financial literacy training, and structured fatigue management systems. Furthermore, integrating routine stress monitoring protocols into occupational safety audits could enable the early detection of psychosocial risks and promote proactive risk mitigation. By explicitly linking empirical findings to actionable strategies, this study advances both scholarly understanding and policy discourse on construction worker well-being. Implementing these recommendations has the potential to reduce preventable occupational strain, enhance long-term productivity, and contribute to the sustainable development of Malaysia’s construction industry.