Predicting mortality in a resource limited neonatal unit: University Teaching Hospital of Kigali (CHUK) experience.

Abstract

INTRODUCTION Disease severity risk evaluation is an important concept in neonatal practice. For a long time, birth weight and Apgar score were used to assess that risk. Later on, a new way of evaluation through the use of another score was developed. Our study aimed at evaluating the mortality risk of newborns in CHUK neonatal unit using this more modern and complex scoring system. METHODS From September 21st 2015 to February 12th 2016 we conducted a longitudinal prospective study in the CHUK neonatal unit. Our objectives were to assess the performance of a simplified SNAPPE II score and additional different risk factors in predicting neonatal mortality in this particular setting. RESULTS The overall mortality rate was 10.4%. The mortality rate in neonates who had hyperglycemia was 100%. A simplified SNAPPE II score of 30 was found to be the best cut off in mortality prediction. A univariable analysis of the simplified SNAPPE II score, birth weight and gestational age significantly predict mortality. By multivariable analysis and adjusting for confounders, the predictive performance was insignificant.