Maternal Mortality Simulation and Cost-Effectiveness Evaluation of Two Interventions in Rural South Asia

Objective: Estimate and compare the cost-effectiveness of iron-supplementation for pregnant woman to a program to teach traditional birth attendants to administer misoprostol suppositories during the third stage of labor.

Methods: We designed an agent-based computer simulation running in Stata 10.0 to predict probability of maternal death for 1,000 hypothetical pregnant women in rural India. The distribution of antenatal hematocrit is drawn from the National Survey of Families and Households in India. The responsiveness of a population level hematocrit distribution to iron supplementation is drawn from data collected in India. The distribution of natural history and misoprostol-treated post-partum blood losses is drawn from the literature. One thousand iterations of the model were run to produce means and uncertainty intervals for the costs and number of lives saved for each intervention.

Results: Results of the model show that both interventions could be considered cost effective depending on the threshold set for the number of dollars per life saved. Sensitivity analyses are on-going to define the performance parameters for these interventions that require better data from field studies.

Discussion: Given the limited resources of Ministries of Health in developing nations where both the maternal mortality ratio and iron-deficiency anemia prevalence are high, the results of this simulation and the cost-effectiveness calculations will allow for governments to select interventions that will lead to the greatest reduction in maternal mortality and morbidity.