Is MMR Really Declining -- Challenges in Monitoring Maternal Mortality

Tracking of maternal deaths is crucial to the planning and implementation of maternal health programs. In India, one of the countries with a relatively high maternal mortality ratio (MMR), a series of national programs have been launched since the early 90s to reduce maternal deaths. However, there is little reliable data documenting the time trend of maternal mortality at the district-level. The lack of district-level data poses considerable challenge to the evaluation of intervention programs. The goal of this study was to explore the use of a spatial-temporal model to synthesize information from various sources in order to derive reliable MMR estimates. We focused on Madhya Pradesh, one of the states in India with high MMR. Data on maternal deaths were gathered from multiple sources including the National Family Health Survey, District-Level Household Survey, the Annual Health Survey, Sample Registration System and the annual reports by the National Rural Health Commission. Descriptive analysis was performed to examine the differences in the MMR estimates from these sources. We then explored the use of a spatial-temporal model to integrate information from the different sources and to estimate complete MMR time-series for each district. Descriptive analysis revealed substantial differences between the MMR reported by the state health department and other surveys such as the Annual Health Survey. In one instance, the MMR reported by the state health department was nearly 7 times lower than that reported by AHS. Using a spatial-temporal model, we obtained MMR time series for each districts in Madhya Pradesh. The changes in MMR across districts ranged from 20% to 50%. This study demonstrates wide differences in MMR estimates from different surveys in the state. Although MMR continued to decline in Madhya Pradesh, the progress varies across districts.