In-depth Mortality Patterns for Africa

Samuel Clark, University of Pennsylvania

A collection of mortality data from Africa compiled by the INDEPTH Network and including over 6.4 million person years of exposure is used to identify new mortality patterns. Seven distinct age-patterns of mortality emerge from the data, two of which may result from excess mortality due to HIV/AIDS. The emergent patterns are compared to the model mortality patterns produced by Coale and Demeny and the United Nations and are demonstrated to be substantially different. The principal components technique is used to calculate fifteen principal components that account for all of the variation in the data. The resulting component model is demonstrated through constructing a hypothetical set of life tables that combine the HIV/AIDS pattern of mortality with an underlying pattern of mortality that is not affected by HIV/AIDS. This yields mortality patterns that may prevail if the population described by the underlying mortality pattern were to be affected by HIV/AIDS.