Introduction The aims were to explore the direct and indirect relationships between workplace risk factors, perceived stress and shoulder pain in a sample of male workers from various companies of the industrial sector and to validate the results in workers from two large companies located in France.

Methods Three working populations were analysed: a surveillance network based on a longitudinal design (Cosali) and two samples from companies based on a cross-sectional design. All workers completed a self-administered questionnaire about musculoskeletal symptoms, individual factors and exposure to work constraints. A conceptual model was defined drawing from the literature and the expertise of the authors. Structural equation modelling was used to test the direct and indirect relationships among the variables.

Result Results obtained in the two companies were in majority consistent with those observed in Cosali. In the three samples, industrial constraints influenced physical factors and psychosocial factors whereas market constraints influenced only psychosocial factors. Psychosocial factors and physical factors were correlated and physical factors increased shoulder pain. Shoulder pain was influenced directly by perceived stress only in one sample.

Discussion The results provide a better comprehension of the complexity of the distal and proximal determinants of shoulder pain and highlight that workplace interventions should act on multiple dimensions (i.e. organisational, psychosocial and physical factors) to be more effective.