Advancing Sexual and Reproductive Health and Rights
International Family Planning Perspectives
Volume 31, Number 3, September 2005

Safe-Delivery Intervention In Rural Pakistan Reduces The Risk of Perinatal Death

In rural Pakistan, an intervention in which traditional birth attendants were trained in basic delivery care and were issued safe-delivery kits was associated with a 30% reduction in the odds of perinatal death when compared with usual pregnancy and delivery care.1 The likelihood of maternal death was not significantly altered, but women cared for under the intervention had 30-80% reductions in the odds of spontaneous abortion, stillbirth, hemorrhage and childbirth-related infection relative to their counterparts who received usual care.

In the six-month, cluster-randomized controlled trial, three subdistricts of Larkana, a predominantly rural district, were assigned to an intervention and four were assigned to usual care, as a control group. All pregnant women in the subdistricts were eligible for the study. In the intervention subdistricts, traditional birth attendants were trained by medical staff, issued disposable safe-delivery kits and asked to visit women during pregnancy. They were instructed to refer women with complications to emergency care, and to inform Lady Health Workers about the women they were caring for. The Lady Health Workers acted as liaisons between the trained traditional birth attendants and the health services. In addition, obstetric teams held outreach clinics for antenatal care about once a month. In all of the districts, Lady Health Workers collected data from birth attendants, women and the women's families during pregnancy and for six weeks after delivery.

Analyses were based on 10,114 women in the intervention group and 9,443 women in the control group. Across subdistricts and across study groups, maternal characteristics were generally similar. On average, women were about 27 years old, had had 3-4 previous live births, lived 2-4 km from the nearest primary health care facility and were recruited to the trial in the fifth month of their pregnancy. The mean number of years of education was slightly higher in the control group (1.4) than in the intervention group (1.1).

In the subdistricts conducting the intervention, 91% of women received antenatal care from a trained traditional birth attendant, and 16% of women visited an outreach clinic for antenatal care at least once. In addition, the trained attendants used 8,172 safe-delivery kits.

Most of the women in both the intervention group and the control group gave birth at home (81-83%) and had a normal vaginal delivery (91% in each group). However, significantly smaller proportions of women in the intervention group than of those in the control group were cared for by a traditional birth attendant who was not trained according to the intervention protocol (6% vs. 76%) or whose training status was unknown (2% vs. 3%). Spontaneous abortion (fetal loss before six months of gestation) was significantly less common in the intervention group (2.5% vs. 3.3%).

In an analysis restricted to singleton pregnancies, the unadjusted perinatal death rate (number of deaths per number of stillbirths and live births) was about 85 per 1,000 in the intervention group and 120 per 1,000 in the control group. After adjustment for characteristics of the subdistricts, the difference corresponded to a significant reduction in the odds of perinatal death in the intervention group relative to the control group (odds ratio, 0.7). The intervention was associated with equally reduced odds of stillbirth (0.7) and neonatal death (0.7).

The odds of maternal death were reduced in the intervention group, but not significantly so. However, relative to their counterparts in the control group, women in the intervention group had reduced odds of hemorrhage after 28 weeks of gestation (odds ratio, 0.6) and markedly reduced odds of experiencing infection associated with childbirth (0.2). The odds of pregnancy- or childbirth-related convulsions and of spontaneous abortion with complications were not significantly affected, but few women experienced these events. Women in the intervention group had higher odds than their control counterparts of experiencing labor lasting longer than 18 hours (1.3), and of being referred for emergency obstetrical care (1.5)—not surprising findings, according to the researchers, because the intervention promoted recognition of and referral for complications.

Training traditional birth attendants and integrating them into health services is feasible and improves perinatal outcomes, the investigators contend. They note that the study was too small to detect a meaningful reduction in maternal death, but the trend for this outcome seemed to parallel the reduction in perinatal death. The investigators point out that the benefits of the intervention were achieved by using the existing infrastructure; factors that may have contributed to the intervention's success included distributing the kits through primary care centers, which increased contact between the traditional birth attendants and health services, and some effect of the kits in elevating the attendants'standing in the community. They conclude, "This model could result in large improvements in perinatal and maternal health in developing countries."—S. London


1. Jokhio AH et al., An intervention involving traditional birth attendants and perinatal and maternal mortality in Pakistan, New England Journal of Medicine, 2005, 352(20):2091-2099.