Which health services reduce maternal mortality? Evidence from ratings of maternal health services

We report cross-national regressions for maternal mortality in 49 developing countries, using indices of the adequacy of maternal health services derived from ratings by at least 10 experts per country. As in previous such regressions, a socioeconomic factor - in this case per capita income - has a significant effect, but having a trained attendant at delivery does not. Instead, the ratings index for access to services has a consistent, significant effect regardless of which estimates of maternal mortality ratios are predicted. Further analysis suggests that access to treatment for pregnancy complications and to services that help avoid pregnancy and birth are most closely related to lower mortality. Service ratings are interdependent, however, so that focusing only on individual services may not be productive.