Unemployment and Early Cause-Specific Mortality: A Study Based on the Swedish Twin Registry

Margaretha Voss, PhD, MPH; Lotta Nylén, MPH; Birgitta Floderus, PhD; Finn Diderichsen, MD, PhD; Paul D. Terry, PhD


Am J Public Health. 2004;94(12) 

In This Article


Among women subjects, elementary school, smoking, use of alcohol, use of tranquilizers, shift work, personality factors, illness, and low socioeconomic status showed at least a 10% higher prevalence among those who had experienced short-term unemployment or long-term unemployment compared with those who had no experience of unemployment ( Table 1 ). Being divorced and use of sleeping pills was also more prevalent among those experiencing unemployment.

Unemployed men more often were unmarried, were childless, smoked, used tranquilizers, had introvert personalities, had long-lasting illnesses, and had low socioeconomic status compared with those without experience of unemployment. Furthermore, being divorced, using sleeping pills, and doing shift work were more prevalent among the unemployed. The latter results were statistically significant, but the difference in prevalence was less than 10%.

Among women, external causes of death -- suicide in particular -- showed a strong association with unemployment, with an almost threefold relative risk for the 24-year follow-up period and a sixfold increase for the first 10 years of follow-up ( Table 2 ). The increased rates among the unemployed remained after adjustment for possible confounding factors, although at a somewhat lower level. Disregarding specific causes, the ever unemployed showed an increased mortality rate based on the 24-year follow-up, with identical but less precise estimates for the first 10 years.

For men, a strong association between unemployment and death from external undetermined cause was found, even after adjustment for potential confounding factors ( Table 2 ). An association on a lower level was also found for all external causes. There tended to be an association between malignant neoplasms and unemployment, which was weakened and still imprecise when potential confounders were controlled for. Total mortality over the 2 follow-up periods was higher among the ever unemployed; adjusting for potential confounding factors somewhat lowered the estimates.

In analyses in which the confidence intervals were adjusted for correlated outcomes, the confidence intervals of the mortality rate ratios were in general identical to those of the unadjusted analyses, showing little influence from correlation within twin pairs (data not shown).

For women, a synergy index of 7.0 was found for the joint presence of unemployment and use of sleeping pills ( Table 3 ), indicating a mortality effect 7 times higher than expected from additivity. For women in the exposed group, use of tranquilizers, extravert personality, and unstable personality combined with unemployment showed a synergistic relation to mortality, although the statistical precision was weak owing to small numbers. For men, elementary school education, use of sleeping pills, and long-lasting or serious illness combined with unemployment were associated with a mortality rate ratio that exceeded expectancy based on additivity.

In the mortality analysis within twin pairs, where 1 twin had experienced unemployment while the other had not, the estimated relative risk of death for unemployed was 1.5 (95% confidence interval [CI]=0.7, 3.1) among women. A corresponding estimate of 1.4 (95% CI=1.0, 2.0) was found for men. Controlling for social, behavioral, health, and personality factors from the 1973 questionnaire lowered the relative risks marginally, to 1.4 (CI=0.6, 3.4) and 1.3 (CI=0.9, 1.9) for women and men, respectively.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.