An assessment of DHS maternal mortality indicators.

This study presents an assessment of the quality of data relating to maternal mortality collected in 14 Demographic and Health Surveys (DHS) for 13 countries that included a complete sibling history. Four aspects of data quality are considered: completeness of the data for reported events, evidence of omission in the reporting of events, plausibility of the pattern of sibling deaths, and sampling errors of the maternal mortality estimates. Although the data relating to reported events are complete for most variables, comparisons of sibling-history-based estimates of adult mortality for both males and females with other independent estimates suggest that sibling estimates are more likely to be underestimates than overestimates. The downward bias is probably greater for female mortality than for male mortality. The sampling errors associated with maternal mortality ratios are substantially larger than those associated with other frequently used DHS indicators. This lack of precision precludes the use of these data for trend analysis and has led to the recommendation that this DHS module not be used more than once every ten years in the same country.