Wanted: studies on mortality estimation methods for humanitarian emergencies, suggestions for future research

Abstract

Measuring rates and circumstances of population mortality (in particular crude and under-5 year mortality rates) is essential to evidence-based humanitarian relief interventions. Because prospective vital event registration is absent or deteriorates in nearly all crisis-affected populations, retrospective household surveys are often used to estimate and describe patterns of mortality. Originally designed for measuring vaccination coverage, the two-stage cluster survey methodology is frequently employed to measure mortality retrospectively due to limited time and resources during humanitarian emergencies. The method tends to be followed without considering alternatives, and there is a need for expert advice to guide health workers measuring mortality in the field. In a workshop in France in June 2006, we deliberated the problems inherent in this method when applied to measure outcomes other than vaccine coverage and acute malnutrition (specifically, mortality), and considered recommendations for improvement. Here we describe these recommendations and outline outstanding issues in three main problem areas in emergency mortality assessment discussed during the workshop: sampling, household data collection issues, and cause of death ascertainment. We urge greater research on these issues. As humanitarian emergencies become ever more complex, all agencies should benefit from the most recently tried and tested survey tools.

Figure 1

The effect of the number of clusters and number of individuals per cluster on the 95% CI for a mortality rate of 1/10 000/day (N = 3600 individuals). Note: The ICC can be used to calculate Deff for different numbers of individuals per cluster (m) using the following formula: ICC = (Deff -1)/(m - 1).