Measuring under-five mortality: validation of new low-cost methods

Julie Knoll Rajaratnam, University of Washington
Linda N Tran, Institute for Health Metrics and Evaluation
Alan D. Lopez, University of Queensland
Christopher J.L. Murray, University of Washington

In countries without complete vital registration systems, there is a demand for improvement of the methods employing summary birth history data to produce more accurate as well as sub-national estimates of child mortality. We use data from 169 DHS to develop new empirically-based methods of estimating under-five mortality using children ever born and children dead data. We focus on improvement of estimates in the most recent time periods prior to a survey where the traditional indirect methods fail, and incorporate uncertainty into these methods. Validated against under-five estimates generated from complete birth histories, our methods outperform the original indirect methods by an average of 45%, and in the five years prior to the survey, by 62%. To illustrate the value of this method for local-area estimation, we apply our new methods to the analysis of summary birth histories in the 2000 Brazilian Census, generating estimates of under-five mortality by municipality.