The practice patterns in neonatology are often a reflection of training programs, and care may be regionalized. Different styles of neonatal care may result in different outcomes and length of stay (LOS). We hypothesized that regional differences in neonatal practice exist, and that these differences result in variations in LOS.

Methods: Neonatal patients admitted to an intensive care nursery or who remained in full-term nursery after their mother's discharge (ICN) in one of three major metropolitan areas (Chicago, Louisville, and Philadelphia), and managed by the same neonatal disease management system (DM) were included(n=2381). All infants included were discharged from an ICN between 7/1/96 and 11/30/97. Each patient was evaluated by an independent neonatal nurse case manager employed by the DM on a frequent basis, and data was validated by randomized chart review. Patients were divided into one of 30 groups based on birthweight and diagnostic category (Groupers). The LOS for patients in each group were case mix adjusted and compared with data obtained from the National Perinatal Information Center (NPIC) as a benchmark. We used ANOVA to test for differences in the values of the ratio between the case mix adjusted LOS and the case mix adjusted NPIC.

Results: The ratios for each city are as follows: CityA(110%), CityB(98%), and CityC(81%). Using multiple comparison tests (Tukey), CityA and CityC were shown to be significantly different (p<0.05).

Conclusions: These data demonstrate a marked difference in neonatal LOS between metropolitan areas, that was not found to be uniform across Groupers. This suggests that there is variation in the practice of neonatology among cities. There may also be, however, unmeasured parameters in neonates, other psychosocial variables, or some combination of these that may differ in the patient mix across regions. We speculate that there may be a need for group adjustment to include severity and other psychosocial variables. Despite that, variation in the clinical practice of neonatology may be present.