Objectives: To compare associations of health-related behaviours with self-certified and medically confirmed sickness absence, and to examine whether these associations can be explained by psychosocial and physical working conditions and occupational social class.
Methods: The study included 5470 female and 1464 male employees of the City of Helsinki surveyed in 2000-2002. These data were linked to sickness absence records until the end of 2005, providing a mean follow-up time of 3.9 years. Poisson regression analysis was used to examine associations of smoking, alcohol use, physical activity, dietary habits and relative weight (body mass index) with self-certified (1-3 days) and medically confirmed (> or =4 days) absence spells. Population attributable fractions (PAFs) were calculated to quantify the sickness absence burden related to the behaviours.
Results: Smoking and high relative weight were most strongly associated with sickness absence, while the associations of other studied health-related behaviours were weaker. The associations were stronger for medically confirmed sickness absence spells for which heavy smoking and obesity more than doubled the risk of sickness absence in men and nearly doubled it in women. Adjusting for psychosocial working conditions had little or no effect on the associations. Physical working conditions and social class somewhat attenuated the associations, especially for smoking and relative weight. In self-certified sickness absence the PAF for smoking (16.4 in men, 10.3 in women) was largest, while in medically confirmed absence relative weight had the largest PAF (23.5 in men, 15.0 in women).
Conclusions: Health-related behaviours, smoking and high relative weight in particular, were associated with subsequent sickness absence independently of psychosocial and physical working conditions and social class. Decreasing smoking and relative weight is likely to provide important gains in work ability and reduce sickness absence.