Academic dissertation for the Degree of Doctor of Philosophy in Public Health Sciences
at Stockholm University to be publicly defended on Friday 14 June 2019, at 13.00 in Room 207, Stress Research Institute, Frescati Hagväg 16 A, Stockholm

Download the thesis from DiVA (Academic Archive On-line)


A work environment characterized by poor psychosocial working conditions may lead to stress and mental health problems such as depression, a common and burdensome public health problem with significant consequences for individuals and for society at large. A number of psychosocial working characteristics have been found to be associated with increased depressive symptoms or clinical depression. This thesis aims to further examine how certain psychosocial working conditions predict depressive symptoms over time, in working life and across retirement. This was done by using several repeated measures from the Swedish Longitudinal Occupational Survey of Health (SLOSH) in 2006–2018.

In study I, we investigated how long-term patterns of exposure to job demands and job control were associated with major depressive symptoms. Those with high strain (high demands, low control) and active (high demands, high control) jobs were more likely to have subsequent major depressive symptoms compared to those with low strain jobs (low demands, high control). However, after adjusting for baseline depressive symptoms and various demographic factors, the associations did not remain statistically significant.

In study II, we assessed how job demands, job control and workplace social support were related to long-term development of depressive symptoms. A perception of high job demands and low social support predicted higher or increasing depressive symptom trajectories. In addition, negative changes in job demands, job control and social support were associated with increased symptoms, indicating that the onset of poor working conditions could negatively impact depressive symptoms.

In study III, we investigated simultaneous and lagged bidirectional associations between job demands, job control, balance between demands and control, social support, procedural justice, effort, reward, balance between efforts and rewards, and depressive symptoms, while controlling for individual time-stable characteristics. There were associations between all work stressors and depressive symptoms when measured simultaneously, except for job control. However, only efforts, were prospectively associated with depressive symptoms measured later.

In study IV, we examined how the same psychosocial working characteristics as in study III were associated with the development of depressive symptoms across retirement. Generally, depressive symptoms appeared to decrease across retirement. Job demands, job strain, social support, rewards, effort-reward imbalance and procedural justice, but to a lesser extent job control and efforts, were associated with a more negative and positive course of depressive symptoms across retirement. Especially, depressive symptoms decreased in relation to retirement for a small group with previously high exposure to work stress.

In conclusion, this thesis indicates that particularly perceptions of high job demands, low workplace social support and high work effort predict subsequent higher levels of depressive symptoms, and/or influence the course of symptoms both in working life and past retirement. In addition, changes in these types of conditions seemed to influence the course of depressive symptoms. Especially, the relief from previous exposure to work stress at retirement seemed to have a clear positive impact on depressive symptoms. These results contribute to strengthen the evidence of causality between these types of work stressors and depressive symptoms.


Professor Tarani Chandola School of Social Sciences, University of Manchester, UK


Dr Linda Magnusson Hanson, Stress Research Institute, Stockholm University
Professor Hugo Westerlund, Stress Research Institute, Stockholm University