Evaluation of a measure of neonatal morbidity

The neonatal risk score of Hobel et al. was evaluated as a measure of neonatal morbidity in a sample of 1,600 singleton infants born alive at two institutions. The score was evaluated for internal consistency, the appropriateness of weight assigned to factors, criterion validity, and conformity to the negative binomial distribution. The results are encouraging for the use of the score as a measure of neonatal morbidity. It appears internally consistent. All risk factors occur with some frequency with at least one other and, in most instances, with several others. With one exception, all factors distinguish infants with high and low scores. The neonatal score, defined by a weighted score and by the number of factors, is highly correlated with length of infant stay in the nursery. Correlations are higher for scores of events measured after birth than for scores of events measured directly at birth. The authors recommend that the number of risk factors be used in preference to the weighted neonatal score since it conforms to the negative binomial distribution and is simpler to calculate. The potential uses of a quantitative, comprehensive measure of neonatal morbidity ar discussed.