Predictive Power of Serial Neonatal Therapeutic Intervention Scoring System Scores for Short-term Mortality in Very-low-birth-weight Infants

Abstract

The Neonatal Therapeutic Intervention Scoring System (NTISS) is used to indicate disease severity for neonates who need intensive care. We examined the ability of serial NTISS scores to predict mortality in very-low-birth-weight preterm infants. We reviewed the medical records of all preterm infants that had a birth weight of <1500g and were admitted to a neonatal intensive care unit from 2007 to 2011. We calculated the NTISS scores at 24 hours, 48 hours, and 72 hours after admission, and assessed the predictive power for mortality using receiver-operating characteristic curve and area under the curve analysis. We also constructed a predictive model with gestational age, birth weight, and NTISS scores to predict mortality in these very-low-birth-weight infants. In total, 172 infants were enrolled into this study. Eighteen (10.5%) infants died in the first 7 days after birth. The area under the curve of the NTISS score was 0.913 at 24 hours, 0.955 at 48 hours, and 0.958 at 72 hours. However, there was no significant difference in the overall average NTISS scores between 48 hours and 72 hours. The NTISS score at 48 hours was a better predictor of mortality than that at 24 hours after admission. Combining gestational age, birth weight, and NTISS score at 48 hours, birth weight was found to contribute little to the predictive power of mortality. The model with gestational age and NTISS score at 48 hours had a better predictive power than the NTISS score alone (area under the curve=0.99). The NTISS score at 48 hours seemed to be effective to predict mortality in preterm infants whose birth weight was less than 1500g. In addition, gestational age played a more important role in predicting mortality than birth weight.