Background: Outcomes and resource uses differ among NICUs. While this depends on case mix, institutional practice styles may also differ. We characterized practice style, adjusting for case mix, in a cohort of premature infants. Outcomes are reported elsewhere.

Methods: We studied 1420 first admissions <31 weeks gestational age (GA) in 7 NICUs in MA and RI (10/94-6/96). Illness severity was measured in the first 12 hours of admission using the Score for Neonatal Acute Physiology (SNAP). We recorded a checklist of treatments in the first 24 hrs (Neonatal Therapeutic Intervention Scoring System-NTISS) which included subscores for respiratory intervention, cardiovascular, procedures & access, transfusions, drugs & nutrition, and monitoring. For both scores, higher is more severe. We used analysis of covariance to determine mean NTISS and subscores for each site, adjusted for GA, birth weight (BW), and SNAP.

Results: GAs, BWs, and SNAPs differed among sites. After risk adjustment, mean therapeutic response differed significantly by site (Table, ranked most aggressive to least). Sites differed in rankings in subscores but site T was distinctly more aggressive and site Z distinctly less aggressive.

Table 1 No caption available.
Full size table

Conclusion: Practice styles are complex, with significantly different mixes of treatments among NICUs for similar babies. Site T generally utilizes more resources. If this is not matched by better outcomes, it would suggest inefficient resource use.