Perinatal mortality in home births in rural Tanzania.

OBJECTIVE To compare perinatal mortality, stratified for risk level, in home births attended by a relative or traditional birth attendant without formal training with births attended by trained personnel in a dispensary or hospital. STUDY DESIGN A prospective community based study in five villages in Northwestern Tanzania during 1990, involving 222 women delivering at home and 199 in a dispensary or hospital. RESULTS Twenty-two of the 29 (76%) perinatal deaths occurred in home births. Perinatal mortality, stratified for risk level, was three times higher in home births than it was in births in dispensaries or in hospital (Mantel-Haenszel odds ratio, 3.29; 95% CI, 1.28-9.22). CONCLUSIONS This study re-emphasises that all births should be attended by adequately trained personnel. More effective strategies are needed to convince women with high risk pregnancies to deliver in hospital. Emergency referral services are required to deal with unexpected complications arising in low risk births at home.