Development and validation of a multivariate predictor of mortality in very low birth weight

Abstract

Accurate prognosis is critical to the design of all prospective research aimed at improving survival. Predictions based on birth weight, gestational age, or any other single variable, fail to take into account the potentially important contribution of other factors. In order to develop a practical and accurate multivariate model, we studied all singleton pregnancies resulting in viable liveborn infants who weighed ⩽ 1500 g at birth during 1984 and 1985 at the New York Hospital-Cornell Medical Center. When gestational age, birth weight, and/or crown-heel length were considered, no maternal characteristics were significant predictors of mortality. The model with the maximal predictive accuracy (84.5%) used birth weight and 5-minute Apgar score to calculate a probability of mortality. This prognostic model was then validated in a separate cohort of singletons born in 1986. We conclude that clinical trials should require stratification before randomization, using the calculated probability of mortality, rather than birth weight or gestational age alone. Given the ability of models, such as the one presented here, to generate reasonable estimates of mortality, this information might also be used in the clinical setting to assist parents and physicians in individualized decision-making processes for a given infant.