Open Access Open Badges Research article

The demand control model and circadian saliva cortisol variations in a Swedish population based sample (The PART study)

Magnus Alderling1*, Töres Theorell2, Bartolomé de la Torre2 and Ingvar Lundberg13

Author Affiliations

1 Division of Occupational Medicine, Department of Public Health Sciences, Karolinska Institutet, Stockholm, Sweden

2 National Institute for Psychosocial Factors and Health, Karolinska Institutet, Stockholm, Sweden

3 National Institute for Working Life, Stockholm, Sweden

For all author emails, please log on.

BMC Public Health 2006, 6:288  doi:10.1186/1471-2458-6-288

Published: 27 November 2006



Previous studies of the relationship between job strain and blood or saliva cortisol levels have been small and based on selected occupational groups. Our aim was to examine the association between job strain and saliva cortisol levels in a population-based study in which a number of potential confounders could be adjusted for.


The material derives from a population-based study in Stockholm on mental health and its potential determinants. Two data collections were performed three years apart with more than 8500 subjects responding to a questionnaire in both waves. In this paper our analyses are based on 529 individuals who held a job, participated in both waves as well as in an interview linked to the second wave. They gave saliva samples at awakening, half an hour later, at lunchtime and before going to bed on a weekday in close connection with the interview. Job control and job demands were assessed from the questionnaire in the second wave. Mixed models were used to analyse the association between the demand control model and saliva cortisol.


Women in low strain jobs (high control and low demands) had significantly lower cortisol levels half an hour after awakening than women in high strain (low control and high demands), active (high control and high demands) or passive jobs (low control and low demands). There were no significant differences between the groups during other parts of the day and furthermore there was no difference between the job strain, active and passive groups. For men, no differences were found between demand control groups.


This population-based study, on a relatively large sample, weakly support the hypothesis that the demand control model is associated with saliva cortisol concentrations.