An Intervention Involving Traditional Birth Attendants and Perinatal and Maternal Mortality in Pakistan

Much remains to be learned about how to encourage the development of effective maternity services so as to lower current very high rates of neonatal and maternal deaths in developing countries. In this randomized, controlled trial, women in 3 of 7 subdistricts in Pakistan were randomly assigned to an intervention in which female health workers linked traditional birth attendants, equipped with disposable delivery kits, with established services. Processes and outcomes were documented. Obstetric teams provided antenatal care at outreach clinics. Women in the other 4 subdistricts were cared for conventionally in the setting of monthly home visits to women and children. The intervention was designed to further antepartum care using available infrastructure and to be a low-cost and sustainable process. Obstetricians and female paramedics trained traditional birth attendants over a 3-day period using picture cards conveying proper antepartum, intrapartum, and postpartum care. Trainees were instructed in how to perform clean deliveries and were told when to refer women for emergency care. The attendants were asked to visit each woman at least 3 times during pregnancy, at 3, 6, and 9 months. The 10,114 women in the intervention group and 9443 control women represented 84% and 79%, respectively, of those eligible. Nearly all participants were followed up to 42 days postpartum. Rates of spontaneous abortion were 2.5% in the intervention group and 3.3% in the control group. Baseline maternal features were comparable in the 2 groups. Crude perinatal death rates were 85 per 1000 and 120 per 1000, respectively, in the intervention and control groups. Analysis by random-effects multilevel modeling yielded an odds ratio for perinatal death in the intervention group, compared with control women, of 0.70 (95% confidence interval CI, 0.59–0.82). Similar results were obtained for stillbirths and neonatal deaths. Maternal mortality rates were 268 and 360 per 100,000 pregnancies, respectively, in the intervention and control groups. The odds ratio for maternal death in the intervention group was 0.74 (95% CI, 0.45–1.23). Puerperal sepsis and hemorrhage were less frequent in the intervention group, but obstructed labors were significantly more frequent in this group. There were no significant differences in rates of eclampsia. Women in the intervention group were likelier than control women to be referred to emergency obstetric care. Sustainable and substantial improvement in pregnancy outcomes was achieved in this large-scale trial by training traditional birth attendants and making them part of an improved healthcare system. This approach holds promise for major improvements in perinatal and maternal health in developing countries.