Long working hours and subsequent use of psychotropic medicine: a study protocol

Mental ill health is the most frequent cause of long-term sickness absence and disability retirement in Denmark. Some instances of mental ill health might be due to long working hours. A recent large cross-sectional study of a general working population in Norway found that not only "very much overtime", but also "moderate overtime" (41-48 work hours/week) was significantly associated with increased levels of both anxiety and depression. These findings have not been sufficiently confirmed in longitudinal studies.

The objective of the study is to give a detailed plan for a research project aimed at investigating the possibility of a prospective association between weekly working hours and use of psychotropic medicine in the general working population of Denmark.

People from the general working population of Denmark have been surveyed, at various occasions in the time period 1995-2010, and interviewed about their work environment. The present study will link interview data from these surveys to national registers covering all inhabitants of Denmark. The participants will be followed for the first occurrence of redeemed prescriptions for psychotropic medicine. Poisson regression will be used to analyze incidence rates as a function of weekly working hours (32-40; 41-48; > 48 hours/week). The analyses will be controlled for gender, age, sample, shift work, and socioeconomic status. According to our feasibility studies, the statistical power is sufficient and the exposure is stable enough to make the study worth the while.

The publication of the present study protocol ends the design phase of the project. In the next phase, the questionnaire data will be forwarded to Statistics Denmark where they will be linked to data on deaths, migrations, socioeconomic status, and redeemed prescriptions for psychotropic medication. We expect the analysis to be completed by the end of 2014 and the results to be published mid 2015.

The proposed project will be free from hindsight bias, since all hypotheses and statistical models are completely defined, peer-reviewed, and published before we link the exposure data to the outcome data. The results of the project will indicate to what extent and in what direction the national burden of mental ill health in Denmark has been influenced by long working hours.