Trends in social inequality in self-reported health in The Netherlands; does infant mortality in year of birth as a cohort indicator matter?

In this article, we study trends in self-reported health (general health and chronic conditions) and health inequality in the Netherlands between 1974 and 1998 using an age-period-cohort framework. We answer two questions: (1) to what extent can trends in self-reported health be explained by the current macro-context (period effect) and by infant mortality in year of birth (cohort effect)? And (2) do the effects of period and cohort differ for educational groups? Health indicators are self-reported poor health and chronic conditions. The use of 26 Dutch cross-sectional surveys makes it possible to estimate largely unbiased effects of period and cohort simultaneously (controlled for age effects) and thus to adequately describe trends in social inequality in health. Our results give rise to four conclusions. First, for men poor health has been more or less stable, for women there has been an increase. The prevalence of chronic conditions has increased for both sexes. Second, adding cohort specific experiences to a model including age and period effects is only relevant for women's poor health. Decreasing infant mortality in year of birth leads to better health and consequently the period effect initially found for women appears to be slightly underestimated. Third, we found no trends in social inequalities in self-reported health due to period effects. Fourth, our analyses do show socially unequal trends in health as a result of cohort specific experiences. Contrary to our hypothesis, we found that decreased infant mortality in year of birth makes for a stronger impact of educational differences on self-reported poor health. Concerning chronic conditions no trends for educational groups were found.