Demographic, Programmatic and Socioeconomic Correlates of Maternal Mortality in Matlab, Bangladesh

Mizanur Rahman, Pathfinder International
Julie DaVanzo, RAND Corporation
Abdur Razzaque, ICDDR,B: Centre for Health and Population Research
Kapil Ahmed, Pathfinder International
Lauren Hale, Stony Brook University, State University of New York (SUNY)

We investigate the association between maternal mortality and demographic, programmatic and socioeconomic factors using longitudinal data on nearly 143,000 pregnancy outcomes during 1982-2002 from the Matlab Demographic Surveillance System. Among the four “too’s” that are commonly believed to increase maternal mortality risk – too young, too short an interval, too many children and too old -- only too old is found to be a risk factor. Controlling for gravidity, women under age 25 have the lowest risk of maternal mortality, and then it increases sharply with age. We find evidence of one additional “too” -- too long; interpregnancy intervals of five years or longer have nearly double the maternal mortality risk of intervals of two-five years. First pregnancies and non-live-births pregnancies also have high risks as do women with a history of child deaths and pregnancy losses, low levels of education, low socioeconomic status, and living in the Matlab Comparison Area.