Monitoring Child Mortality Change through Household Surveys

Ken Hill, Johns Hopkins University
Agbessi Amouzou, United Nations Children's Fund (UNICEF)
Eoghan Brady, Johns Hopkins University
Linnea Zimmerman, Johns Hopkins University
Livia Montana, Harvard University
Romesh Silva, Johns Hopkins University

Under-5 mortality is high on the global development agenda; there is increasing need for rapid mortality monitoring to evaluate program impact. Most low- and middle-income countries do not have fully functional civil registration systems, so monitoring has largely been through periodic household surveys using full birth histories. However, such surveys are impractical for rapid monitoring because of cost considerations. Surveys using summary birth histories (SBHs) are much cheaper, and can therefore in principle be carried out more frequently for larger samples, but existing analysis methods do not provide estimates for short time periods. We have developed and tested in several developing country settings two new methods of analysis that can provide estimates for single year periods, one using imputed birth histories and the other using cohort changes between surveys. Estimates from both methods are critically dependent on the quality of the underlying SBHs, giving good estimates from high quality surveys.