Using data from the 1998 Quebec Social and Health Survey, 4149 male and 3116 female workers meeting inclusion criteria were identified. MSD was defined as the presence of significant musculoskeletal pain interfering with activities frequently or all the time in the previous 12 months and identified using a self-administered questionnaire. Perception of whether symptoms were partly or entirely related to work was measured. Three main outcomes were studied: work-related MSD, non-work-related MSD and MSD without distinction of work-relatedness. SOC occupational codes were categorized into 5 occupational classes from senior managers and professionals to unskilled manual workers. Work exposure measures included a 5-item index of physical work demands, general work posture, Karasek decision latitude and psychological demand 9-item scales, intimidation at work, difficult situations with the public and work hours. For each MSD outcome, the Adjusted Risk Ratio (ARR) for each occupational class was compared in 4 multinomial logistic regression models: Model 1) occupational class alone; Model 2) model 1+ physical work demands and work posture; Model 3) model 1 + psychosocial work demands; Model 4) model 1 + all work exposures. All models were adjusted for age, BMI, smoking, leisure time physical activity, social support & work hours. Separate analyses were carried out for men and women.