APH 2003, 61, 101-125:

Occupational stress and incidence of sick leave in three sectors of activity of the Belgian workforce: the belstress study.

M. Moreau, F. Valente, R. Mak, E. Pelfrene, P. De Smet, G. De Backer, M. Kornitzer

Keywords: Job characteristics, job stress, sick leave, stress models, demands-control-support model, sectors of activity

Background: The Karasek-Johnson "Demands-Control-Social support" (DCS) model is one of the most widely used job stress models. It has been largely used to predict a wide range of health outcomes and to a lesser extent absenteeism. Our aim is to test the predictive power of the model in relation with one year incidence of sick leave in three sectors of activity, taking into account a wide range of socio-demographic and behavioural variables.

Methods: The baseline survey of the BELSTRESS study was conducted in 25 companies of the secondary, tertiary and public sectors across Belgium between 1994 and 1998. A cohort of 15,557 males (9621 from the secondary, 2640 from the tertiary and 3296 from the public sector) and 4906 females (1076 from the secondary, 1802 from the tertiary and 2028 from the public sector) were followed for absenteeism during a year. The relation of psychological job demands, job control (decision latitude), social support at work as well as the job strain concept (combination of high job demands and low control) and the iso-strain concept (job strain combined with high or low social support) was tested in relation with high annual sick leave incidence (total number of sick leave days above the percentile 75 of the distribution of total number of sick leave days), short (< 7 days) and long spells (> 7days) of sick leave by univariate and multivariate logistic regression models.

Results: Independent from major baseline confounding variables we observed a significant association between job characteristics and sick leave: thus, low job control was associated with short and long spells of sickness absence in both genders with OR of 1.17 (Cl 95% 1.08-1.27) and 1.21 (Cl 95% 1,09-1.34) in men respectively and 1.16 (Cl 95% 1.00-1.33) and 1.27 (Cl 95% 1.06-1.51) in women respectively. We also found an association between high strained jobs with low social support and long spells of sickness absence in both genders with odds ratios of 1.22 (Cl 95% 1.05-1.41) in men and 1.35 (Cl 95% 1.05-1.74) in women. Independently from a wide range of potential confounders, the public sector is also associated with a significant 24% increase of sickness absence incidence in men. Finally, we estimated the number of sickness absence days in men avoided if control would be increased above the median: from 529 days per 1000 workers per year (upper white collars of the tertiary sector) to 3797 days per 1000 workers per year (lower white collars of the public sector).

Conclusions: We have shown evidence that job stress is an independent risk factor of sick leave whatever the gender, the occupational class and the sector of activity. Moreover, we were able to identify within the Karasek model those job characteristics that play a major role in the relation between job stress and sick leave: job control and social support at work. Sick leave being an indicator of morbidity and productivity, the results of this study have implications in the domain of both public health and economy and should help to "re-think" the work and its organisation in Belgium.