REDUCING PERINATAL-MORTALITY IN DEVELOPING-COUNTRIES - HIGH-RISK OR IMPROVED LABOR-MANAGEMENT

International organizations have advocated screening and classification of pregnancies based on maternal characteristics (the 'risk approach') as a strategy for reducing maternal and perinatal death rates in developing countries. Such a strategy assumes that the predictive values of such maternal characteristics are sufficiently high to be useful in allocating the scarce obstetric resources that typically exist in such countries. We used the data from an intensive one-year longitudinal study of births in a traditional indigenous rural community of Guatemala to examine this assumption. The most common adverse outcome of pregnancy, death during labour and delivery or on the first day of life of an infant apparently alive at the onset of labour, was the variable used to evaluate predictive values. We collected complete information on the mother's obstetric history and sociodemographic characteristics, pregnancy, labour and delivery and the newborn for the 23 births resulting in potentially preventable intrapartum/day one death and for 329 (86%) of all other births during the study period. For mothers of infants experiencing intrapartum/day one death, 30% were primigravidas, 30% had greater-than-or-equal-to 3 previous pregnancies, 43% had experienced a bad outcome of a previous pregnancy, and 70% were illiterate. However, the positive predictive values for intrapartum/day one death of these and other maternal characteristics were low (from 0 to 0.13). By combining first pregnancy with the three characteristics of multigravidas having highest positive predictive values and greatest prevalence among mothers of intrapartum/day one deaths (prior obstetric problems, short birth interval, less than 50% survival of liveborn children) 100% of intrapartum/day one deaths were identified. Also identified as 'high risk' were 70% of all other women giving birth. The occurrence of a labour and delivery complication was a much stronger predictor of intrapartum/day one infant death. Obstetric complications subject to improved detection and management accounted for 73% of the 26 intrapartum/day one infant deaths delivered in the community, but occurred in only 18% of study births. Of these complications, fetal malpresentations had the highest case-fatality rate and accounted for over half of the potentially preventable intrapartum/day one infant deaths. These results suggest that in developing communities such as the one studied, the maternal risk screening and classification approach does not have adequate predictive value to assign obstetric resources. Interventions based on improved detection, management, and referral of complicated labours and deliveries might be effective interim strategies for obstetric risk management in such settings.