Childhood Mortality Levels and Trends in Kenya: New Estimates Based on the Own Children Method

Collins Opiyo, University of Pennsylvania

The Brass technique is widely used to estimate childhood mortality among Sub-Saharan African countries with limited and defective data. However, two fundamental assumptions – of constant fertility and mother-child mortality independence – are violable by the ongoing fertility transitions and the high prevalence of HIV/AIDS observed in many countries. This study provides new estimates of childhood mortality using the own children method based on Kenyan Census data. The technique uses the age distribution of surviving children and does not require any assumptions on the recent fertility patterns. The new estimates are robust because they are purged of biases associated with the contemporary application of the Brass method and give plausible trends across censuses. They suggest that the greatest increase in childhood mortality risk attributable to HIV/AIDS – of about 10 % – occurs at age five and that HIV/AIDS does not have a discernible impact on the age pattern of childhood mortality.