Model-based estimates of adult mortality in sub-Saharan Africa: comparisons with data on the survival of parents and siblings

Bruno Masquelier, Institut National d'Études Démographiques (INED)

Data on adult survival remain scant and defective in Sub-Saharan Africa. Adult mortality is still mostly estimated by combining summary measures of child survival with model mortality schedules. The death rates are in turn used to derive orphanhood prevalences and maternal mortality levels. Such model-based estimates remain overly sensitive to the choice of the age pattern to infer the background mortality. Their accuracy depends also on the validity of assumptions made to estimate the number of aids-related deaths. In this paper, UN estimates are compared with sibling survival estimates from 62 DHS. UNAIDS orphan prevalences are also compared with proportions of orphaned children observed in censuses and sample surveys. The background mortality appears to be underestimated by model outputs in a few countries because of the use of the South and West model life tables. In countries with severe HIV/AIDS epidemic, the aids-related mortality also seems underestimated, especially for males.