| The first 300 words of the full text of this article appear below. |
To the Editor.
I read with interest the article you have published on the outcome of neonates cared for in different size units.1 As a provider of care who has worked at both community-based and regional medical centers, I have several questions and points of discussion.
The highest mortality rate is found in neonates weighing <2 kg delivered at hospitals without neonatal intensive care unit (NICU) care (159 sites). I think few would argue that no infant should be delivered at a hospital without a neonatal resuscitation team. How many of these "no NICU units" have regular neonatal resuscitation programs (NRPs)? The median number of neonates with a birth weight <2 kg delivered at the "no NICU sites" was 4.5 per year. However, it is also clear that at least 1 unit had 31 such deliveries per year. What are the maternal demographics of mothers delivering at these sites? Do they have access to (do they receive) prenatal care? Have the authors visited this unit to determine if the problem identified in 1992 and 1993 still exists in 2002?
Clinical recommendations have been made based on statistically insignificant results. It is one thing to report these results; it is another to make policy decisions that are not supported by the evidence presented. If 1.29 (95% confidence interval [CI]: 0.742.25) presented in Table 3 for neonates (birth weight [BW]: 12501999 g) cared for in intermediate NICUs is considered clinically important, then the finding that "major teaching hospitals" did worse must also be considered clinically important (odds ratio [OR]: 1.39; 95% CI: 1.041.86). If we are to use nonstatistically important differences, I could use the data in Table 2 to suggest that neonates should be born in for-profit hospitals (0.82