High risk obstetrics is rapidly evolving into a high cost, technology intensive field, especially when coupled with neonatal intensive care. Although high risk cases are a small percentage of the total obstetric population, they account for the majority of maternal, fetal and neonatal morbidity and mortality. Furthermore, most high risk obstetric cases are concentrated in a limited number of tertiary centers. There is enormous variation in outcomes among these tertiary centers as evidenced by marked differences in newborn morbidity and mortality. This variation may reflect differences in initial obstetric risk, in the effectiveness of high risk obstetric treatments or in the success of neonatal care. Despite this, few studies in this field have focused on maternal risk status, outcomes, effectiveness or resource use. Evaluation of effective treatments and assessment of perinatal services requires characterizing the risk of the mother when she initially presents to a perinatal center and measuring neonatal status at the time of birth. Traditional outcome measures for obstetrics such as birth weight, gestational age, Apgars and neonatal mortality may be inadequate for assessing the success of high risk obstetric care because they reflect events occurring before or subsequent to perinatal interventions. The proposed project will develop tools to quantitate initial obstetric risk upon entry to a perinatal center. In addition, it will examine the utility of an established newborn severity of illness score, the Score for Neonatal Acute Physiology (SNAP), as an obstetric outcome measure. The study sample will consist of approximately 1200 women delivering premature infants for whom extensive data are available from a related Agency for Health Care Policy Research study on newborn illness severity and outcome variation in neonatal intensive care. Logistic models of initial obstetric risk will be developed using specific obstetric events, maternal conditions and demographic factors as predictors of neonatal outcomes. Parallel analyses will use linear models to establish which obstetric events/treatments are associated with higher or lower SNAP scores and to quantitate their contribution to the overall variation of the SNAP score. Finally, the associations between obstetric risk, neonatal outcome and hospital resource use will be explored by analyzing length of stay and hospital costs for the mother, the infant(s) and the mother-baby pair(s). By characterizing initial risk and validating an outcome measure that reflects the immediate condition of the newborn, this project will provide tools for comparison of high risk obstetric services, identification of effective practice styles and evaluation of resource use.
