role in the disparity in African American and white infant mortality rates (Allen et al., 2000) (Appendix B).
Access to neonatal intensive care was recognized as an important issue in the 1970s. Schlesinger (1973) was the first to report differences in the rates of neonatal survival among hospitals. The dearth of physicians and nurses skilled in the new techniques and support services needed to care for sick neonates led to the development of regional programs with NICUs with prescribed structures and functions, formalized arrangements for obstetric referrals, and transportation systems for sick neonates (Blackwell et al., 2005) (see Chapter 14). Regionalization initially involved transporting sick newborns from community hospitals to regional medical centers and outreach community education on the stabilization of acutely ill newborns.
The main arguments in favor of regionalized care rested largely on improved neonatal survival after its introduction into a geographically defined region (McCormick and Richardson, 1995). Low birth weight infants born in hospitals without a NICU had higher risk-adjusted mortality rates than those born in hospitals with an intermediate- or high-level NICU, and the mortality rate only marginally improved with subsequent transfer of the infant to an NICU (Cifuentes et al., 2002). The advantage of the earlier identification of high-risk pregnancies and referral to tertiary perinatal centers before delivery is supported by the more favorable outcomes for infants whose mothers were transported to perinatal centers before delivery compared with the outcomes for infants transported after birth (Doyle et al., 2004a; Kollée et al., 1988; Levy et al., 1981).
The concept of regionalized services has evolved to include the prenatal period and a fully integrated system of consultation, referral, and transport (McCormick et al., 1985).
Guidelines for designating levels of perinatal care (Level I, II, or III, depending on the resources available, the delivery volume, and geographic need) have been developed. As the proportion of infants with birth weights of less than 1,000 grams born at Level III perinatal centers has increased, their survival has improved, and the gap in survival between infants born in and out of such centers has increased (Saigal et al., 1989). In addition to gains in safety and expertise, the development of highly integrated vertical networks is inherently cost-effective because of the elimination of fragmented and redundant services.
This level of integration of regionalized perinatal services is difficult to achieve. In Georgia, the rate of delivery of infants with birth weights of less than 1,500 grams at recommended perinatal centers was better for urban mothers than for rural mothers who lived farther away from regional cen-