Bangladesh is one of the leading exporters of ready-made garments (RMG) worldwide producing at very low cost almost exclusively for Western markets. Empirical evidence on psychologically adverse working conditions and their association with health in the RMG setting remains sparse. Drawing on insights from previous ethnographic research, we conducted a cross-sectional epidemiological study among 332 RMG workers in Dhaka, Bangladesh. High work-related demands and poor interpersonal resources represented key components of work stress and were important determinants of poor health. The key work stress components observed in this study partly differed from those identified in Western work place settings.

Evidence on the association of work stress with cortisol levels is inconsistent and mostly stems from Western countries, with limited generalizability to other regions of the world. These inconsistencies may partly be due to methodological limitations associated with the measurement of cortisol secretion in saliva, serum or urine. The present study set out to explore associations of work stress with long-term integrated cortisol levels in hair among 175 workers of an export oriented ready-made garment (RMG) factory in Dhaka, Bangladesh. Work-related demands (WD), interpersonal resources (IR) and work-related values (WV) were assessed using a psychometrically evaluated interview. WD consisted of four items on physical demands, time pressure, worries about mistakes and exposure to abusive language. IR comprised five items addressing support, recognition, adequate payment, workers' trust in the management, and the management's trust in workers, as perceived by the workers. WV captured job security, promotion prospects and job latitude by three items. Hair cortisol concentrations (HCC) were analyzed by liquid chromatography-mass spectrometry. Stepwise multivariable linear regression models (backward elimination of predictors) were used to estimate associations of HCC with the three work stress components. For significant work stress component(s), further multivariable linear regression analyses were conducted to explore whether, and if so, which individual item(s) contributed most. The mean HCC equaled 3.27 (SD 2.58) pg/mg. HCC were found to be significantly associated with WV (beta=0.209, p=0.021). Additional analyses of the three WV items revealed that this association was largely driven the item on "promotion prospects" (beta=0.230, p=0.007) implying that the perception of good promotion prospects was associated with higher HCC. The finding of elevated HCC with good promotion prospects may initially seem counter-intuitive, but is supported by research documenting that job promotion may result in poorer mental well-being. Moreover, being promoted in the Bangladeshi RMG industry may represent a stressful experience: job promotions are rare in this setting and are associated with the need to meet exceptional job-related demands. Further research from ethnic and culturally diverse occupational settings is needed to test this hypothesis, to shed light on the reproducibility of our findings and to improve our understanding of the psychobiological implications of psychosocial working conditions across cultures and contexts.