Social inequality in health--the role of the healthy worker effect.

Current occupational status is frequently used as an indicator of position in the socioeconomic structure in analyses of health inequalities. People outside the work-force are thereby ignored. One may hypothesize that due to a 'healthy worker effect' economically active people are a positively selected health group and that health-related exits from the labour marked are concentrated among low socioeconomic groups. One may therefore expect that this system of classification will show artificially small differences in health between socioeconomic groups. Analyses of material from a nation-wide representative survey and a complete set of mortality statistics confirm these expectations. When previously employed are included among currently employed, the differentials in health between the extreme groups in the occupational hierarchy widen. The system of classification has only a moderate impact on the main conclusion regarding men. For women, however, the conclusion depends more heavily on how they are classified. If one relies only on current occupation, small or inconsistent patterns emerge. Inclusion of previously employed has the effect of showing that low status women suffer from poorer health than high status women. This applies to several measures of health, i.e. morbidity, restricted activity and mortality.