Poverty and health in different contexts : Social inequalities in child mortality in Mozambique and 19th century Stockholm

Infant and child mortality has declined in many low-income countries. However, in sub Saharan Africa, childhood mortality is still one of the major public health problems, which is worsening as some countries experience new increases in mortality due to HIV/AIDS. On the other hand, historically, in many developed countries infant and child mortality declined as social and economic changes (modernisation) took place. However, the mechanisms that brought about the decline are still not well understood. The levels of mortality and many major causes of death among poor children in the Europe of the past were similar to those presently prevailing in lowincome countries, and were socially patterned. Hence, comparisons between social differentials in child mortality in the two settings might contribute to knowledge of how to reduce mortality among poor children in low-income countries. The overall aim of the thesis was to increase understanding of the mechanisms underlying high levels of mortality and inequalities in child mortality by investigating social differentials in child mortality in Mozambique 1973-1997 and Stockholm 1878-1925, and their determinants. The thesis is based on five papers. Two data sources were employed. Papers I-IV used data from the 1997 Mozambique Demographic and Health Survey. Study V used data from the Roteman archives for Stockholm 1878-1925. The first paper used demographic methods to describe urban/rural differentials in infant and child mortality in Mozambique. In papers II-V calculation of mortality rates and Cox regression analysis were used to investigate the relationship between selected variables and mortality. There were social differentials in child mortality in both settings. Nevertheless, it is important to contextualise the study of indicators of social position/disadvantage, since the meaning of a certain social position may vary according to context, creating different health divides. In Mozambique, children born to mothers who lived in rural areas (Paper I), the children of illiterate fathers (Paper II) and children whose mothers had an Emakua or Cisena ethnic affiliation (Paper III) experienced the worst survival chances. The ethnic differentials were partly explained by other factors. On the other hand, in Stockholm 1878-1925 it was the children of unmarried mothers who had the highest mortality risks (Paper V). Parental occupation seems to have influenced child mortality in a similar way. In both contexts children who lived in households in which the mother, the father or the head of the household (in Stockholm 1878-1925) was in a lower socioeconomic group showed higher mortality. The Mozambican studies also showed that external influences, such as war or structural adjustment programmes and their consequences, may be detrimental to child survival. The study of Stockholm shows that it is possible to reduce high levels of and social differentials in child mortality through both general and targeted interventions, and that a reduction in mortality did not occur only via economic improvement per se, but through implementation of specific policies and public health interventions. Household environment factors, such as access to clean water and adequate sanitation, played an important role in both settings (papers IV and V). Greater availability of such services seems important in order to achieve lower child mortality in Mozambique. The cause-of-death pattern in the two settings was partly similar, suggesting that further study of the specific aspects of the decline in child mortality in Stockholm that are common to the two settings may contribute knowledge of use in a Mozambican setting. However, the age structure of under-five mortality differed between the two settings