Aim: Social equity in health is an important goal of public health policies in the Nordic countries. Infant mortality is often used as an indicator of the health of societies, and has decreased substantially in the Nordic welfare states over the past 20 years. To identify social patterns in infant mortality in this context the authors set out to review the existing epidemiological literature on associations between social indicators and infant mortality in Denmark, Finland, Norway, and Sweden during the period 1980-2000. Methods: Nordic epidemiological studies in the databases ISI Web of Science, PubMed, and OVID, published between 1980 and 2000 focusing on social indicators of infant, neonatal, and postneonatal mortality, were identified. The selected keywords on social indicators were: education, income, occupation, social factors, socioeconomic status, social position, and social class. Results: Social inequality in infant mortality was reported from Denmark, Finland, Norway, and Sweden, and it was found that these increased during the study period. Post-neonatal mortality showed a stronger association with social indicators than neonatal mortality. Some studies showed that neonatal mortality was associated with social indicators in a non-linear fashion, with high rates of mortality in both the lowest and highest social strata. The pattern differed, however, between countries with Finland and Sweden showing consistently less social inequalities than Denmark and Norway. While the increased inequality shown in most studies was an increase in relative risk, a single study from Denmark demonstrated an absolute increase in infant mortality among children born to less educated women. Conclusions: Social inequalities in infant mortality are observed in all four countries, irrespective of social indicators used in the studies. It is, however, difficult to draw inferences from the comparisons between countries, since different measures of social position and different inclusion criteria are used in the studies. Nordic collaborative analyses of social gradients in infant death are needed, taking advantage of the population-covering registers in longitudinal designs, to explore the mechanisms behind the social patterns in infant mortality.
