Department of Agricultural, Food, and Resource Economics Food Security Group

IDWP 99. Antony Chapoto, T.S Jayne, L. Kirimi, and S. Kadiyala. Michigan State University. 2009. Characteristics Associated with Prime-Age Mortality in Eastern and Southern Africa: Evidence from Zambia and Kenya

EXECUTIVE SUMMARY:
Campaigns to prevent the spread of HIV require accurate knowledge of the characteristics of
those most likely to contract the disease. Studies conducted in Sub-Saharan Africa during the
1980s generally found a positive correlation between socioeconomic characteristics such as
education, income, and wealth and subsequent contraction of HIV. As the disease has
progressed, the relationship between socioeconomic status and HIV contraction may have
changed, although there is little evidence to support this. An emerging strand of the literature
on the AIDS epidemic in Africa posits that poverty is increasingly associated with the spread
of the disease. However, this conclusion is somewhat contentious, as other recent studies
find mixed evidence of a poverty-AIDS connection.

This study attempts to shed light on these issues by reporting findings from two linked
studies on the socioeconomic characteristics of prime-age individuals (defined as ages 15 to
59) dying of disease-related causes in Zambia and Kenya. Longitudinal data from nationwide
rural household surveys are analyzed using probit models to determine how the probability of
individual mortality is affected by initial socioeconomic characteristics of individuals and
their households. The survey periods covered in the two countries examined in this study
(Zambia, 2000 to 2004 and Kenya, 1997 to 2004) makes the findings more recent than the
studies in the 1980s which led to the generally accepted findings that AIDS-related mortality
disproportionately afflicted individuals with relatively high incomes and educational
attainment. Moreover, the panel nature of the data in this study provides the means to
examine individual and household characteristics prior to mortality, which provides a more
accurate picture than cross-sectional studies which measure socioeconomic attributes only
after mortality has occurred.

Findings in these two countries point to a more complex burden of mortality than generally
understood, with significant variations by gender, age and economic conditions of
individuals. In Zambia, there appears to be little correlation between indicators of individual
and household wealth and the probability of disease-related mortality. Nor is there a
significant relationship between individual educational attainment and mortality, for either
poor or non-poor men or women. Being away from home was associated with increased
probability of mortality. By contrast, results in Kenya where HIV prevalence is about half
that of Zambia are somewhat similar to those found in Zambia. Years of education reduces
the probability of mortality for men, especially non-poor men, but not for women. Diseaserelated
mortality rates for poor men and women (5.46 and 5.07 persons per 1000 person
years) are higher than for non-poor men and women (3.51 and 3.44 persons per 1000 person
years). However, we find that women who had salaried jobs or business activities in the
initial survey period were no more or less likely to die than women without such jobs.