Background Adverse psychosocial working conditions have been linked with a range of indicators of poor health, including hospitalisation for both somatic and mental disorders. However, it is unknown whether these associations may be due to individuals with poor health being selected into specific job-groups, or if the associations reflect a causal relationship. This study aims to investigate whether an adverse psychosocial work environment characterised by high levels of violence, threats, emotional demands and low levels of influence is associated with increased risk of being treated for mental and somatic disorders, when accounting for mental and physical health factors prior to labour market entry.

Methods The sample consists of all Danish residents, aged 15–30 years who entered the workforce in Denmark during the years 1995–2014. By using data from the Danish Work Environment Cohort Study and applying job exposure matrix (JEM) methodology, we will generate dynamic measures of psychosocial work environment characteristics aggregated at the job group level during 1995–2010. These matrices will, in turn, be linked to registry data to indicate a typical exposure for the different job groups throughout the time-period of the study.

Results This study is in its early stages but data on the performance of the job exposure matrices will be presented at the time of the conference.

Discussion By using register data and applying JEM we address several sources of bias that are otherwise common in the psychosocial work environment literature. Reporting bias is largely mitigated as we aggregated assessments of the psychosocial work environment. Selection bias into the different professions will be assessed by including health information from all individuals before, during and after the exposures. Finally, work environment exposures prior to baseline assessment will not bias the results, as all participants are followed-up from the first day they entered the workforce.