Prediction by maternal risk factors of neonatal intensive care admissions: evaluation of >59,000 women in national managed care programs.

OBJECTIVE: Managed care plans have adopted risk assessment tools as part of pregnancy disease state management strategies to assist in reducing poor pregnancy outcomes and related costs. We evaluated the relationships of maternal risk factors to determine which pregnancy risk factors were associated with neonatal intensive care unit (levels II and III) admission.

STUDY DESIGN: Risk assessments were performed through perinatal telephone interviews of nurses with 59, 861 pregnant women during 1996 and 1997 calendar years as part of managed care maternity risk screening and education programs. A series of 3 interviews was conducted, at 17 weeks and 28 weeks average gestational age and at 2 weeks post partum. Univariate chi(2) analysis was performed on >50 historical and pregnancy risk factors to determine the associations with neonatal intensive care unit admission. Significant factors were included in a stepwise logistic regression model. Receiver operating curves were generated for the use of significant factors in a risk scoring system in the prediction of neonatal intensive care unit admission, and the percentages of neonatal intensive care unit days attributable to significant risk factors were calculated.

RESULTS: Among the participants most women (90%) had their prenatal visit during the first trimester. The mean maternal age was 30.2 +/- 5.2 years, with 74% of women reportedly of white ethnicity, 86% married, and 44.3% primigravid. The mean gestational age at birth decreased with increasing number of fetuses from singletons to quadruplets. The chi(2) analysis identified 26 significant risk factors associated with neonatal intensive care unit admission. Of these, 14 remained significant by logistic regression. Multiple gestation, preterm premature rupture of membranes, diabetes, abruptio placentae, pregnancy-induced hypertension, and preterm labor were independently associated with at least a 3-fold risk of neonatal intensive care unit admission. A modeled risk scoring system that used these and other significant factors was poorly predictive of neonatal intensive care unit admission. However, an analysis of neonatal intensive care unit length of stay attributable to significant risk factors concluded that 19% of all neonatal intensive care unit days in this population were associated with multiple gestations. Furthermore, 85% of neonatal intensive care unit days were the result of infant lengths of stay >/=1 week.

CONCLUSION: This analysis of a managed care population showed similar risk factors to those traditionally associated with neonatal intensive care unit admission. Although many of these risk factors are not preventable, identification of neonatal intensive care unit admission risks with a screening program may be of use for focusing interventions, and earlier identification of these factors may allow maximum impact of interventions. Importantly, a reduction in the incidence of higher-order multiple gestations might help to reduce neonatal intensive care unit admissions and costs.