To determine the extent to which sickness absence is predictive of mortality. Prospective cohort study. Data on medically certified long term absences (>3 days), self certified short term absences (1-3 days), and sick days were derived from employers' records and data on mortality from the national mortality register. 10 towns in Finland. 12821 male and 28915 female Finnish municipal employees with a job contract of five consecutive years. The mean follow up was 4.5 years. After adjustment for age, occupational status, and type of employment contract, the overall mortality rate was 4.3 (95% confidence intervals 2.6 to 7.0) and 3.3 (2.1 to 5.3) times greater in men and women with more than one long term absences per year than in those with no absence. The corresponding hazard ratios for more than 15 annual sick days were 4.7 (2.3 to 9.6) and 3.7 (1.5 to 9.1). Both these measures of sickness absence were also predictive of deaths from cardiovascular disease, cancer, alcohol related causes, and suicide. Associations between short term sickness absences and mortality were weaker and changed to non-significant after adjustment for long term sickness absence. These findings suggest that measures of sickness absence, such as long term absence spells and sick days, are strong predictors of all cause mortality and mortality due to cardiovascular disease, cancer, alcohol related causes, and suicide.
