Résumé
<p><b>OBJECTIVE</b>To investigate the accuracy and clinical utility of neonatal critical illness score (NCIS) and score for neonatal acute physiology, perinatal extension, version II (SNAPPE-II) in predicting the "dead and abandoned" risk in critically ill neonates.</p><p><b>METHODS</b>A total of 269 critically ill neonates were divided into two groups according to their prognosis: dead/abandoned and improved/cured. The accuracy of these two scoring systems, NCIS and SNAPPE-II, in predicting the "dead and abandoned" risk was compared.</p><p><b>RESULTS</b>The dead/abandoned group had a significantly higher SNAPPE-II score than the improved/cured group (P<0.001), while there was no significant difference in the NCIS score between the two groups (P=0.091). The children who were in line with the individual indicator in the NCIS results had a significantly higher "dead and abandoned" risk than those who were not (P=0.005).</p><p><b>CONCLUSIONS</b>SNAPPE-II is more accurate in early prediction of the "dead and abandoned" risk in critically ill neonates compared with NCIS. NCIS has the ability to predict the "dead and abandoned" risk in children in line with the individual indicator.</p>