Score for neonatal acute physiology: validation in three Kaiser Permanente neonatal intensive care units.

BACKGROUND

Measurement of the severity of illness is a research area of growing importance in neonatal intensive care. Most severity of illness scales have been developed in tertiary care settings. Their applicability in community neonatal intensive care units has not been tested.

OBJECTIVES

Our goal was to assess the operational characteristics of the score for neonatal acute physiology (SNAP): the relationship to birth weight, the length of total hospital stay, and in-hospital mortality.

METHODS

We assigned SNAP scores prospectively to all inborn admissions at three community neonatal intensive care units during an 11-month period. Data on other neonatal predictors (eg, birth weight and the presence of congenital heart disease) were also collected. We measured in-hospital mortality, the experience of interhospital transport to a higher level of care, and total hospital stay.

RESULTS

We found that the SNAP's relationship to birth weight was similar to previous reports. The SNAP's perinatal extension is a reliable predictor of newborn in-hospital mortality, with an area under the receiver operator characteristic curve of 0.95. The SNAP is also a good predictor of total hospital length of stay, whether by itself (by which it can explain 31% of the total stay) or in combination with other variables. Its predictive ability is better among infants of low birth weight (<2500 g) than among those of normal birth weight (> or = 2500 g). The SNAP's predictive power was most limited among infants admitted to rule out sepsis. The predictive ability of a model containing birth weight, the SNAP, and transport status was not improved by the inclusion of two major diagnostic categories, the presence of congenital heart disease or complex illness.

CONCLUSION

Although it has definite limitations among infants who weight 2500 g or more, the SNAP is a potent tool for outcomes research. Modification of some of its parameters could result in a multifunctional scale suitable for use with all birth weights.