Clinical risk index for babies (CRIB) II score as a predictor of neonatal mortality among low birthweight babies at Kenyatta National Hospital

Abstract

Background: Neonatal deaths, especially among the LBW babies, are of major concern in the Newborn unit of Kenyatta National Hospital. Several instruments have been developed to predict initial mortality among the LBW babies, Among them is CRIB II score, It has been evaluated in several countries but not in Kenya, Objective: To evaluate CRIB II score as a predictor of neonatal mortality among the LBW babies at KNH, Methodology: A prospective cohort study was carried out in the NBU at KNH, CRIB II score was assigned to all LBW babies admitted to the newborn unit between 8am and 8pm during the study period who met the selection criteria until a sample size of one hundred and thirty five was achieved, Data on other neonatal mortality predictors such as APGAR score, maternal age and parity was collected and ana lysed, Results: One hundred and thirty five newborns were enrolled, Birthweight ranged from 600 - 2500g, with a median of 1600g, Total CRIB II score ranged from 1 - 15, with a median of 5,5, Gestational age ranged from 26 - 38 weeks, Total mortality was 45,9%, Birthweight <1500g, gestational age <30 weeks, base excess <-12mmol/l, temperature at admission >37,5 or <35 and total CRIB II score of> 4 were all found to be significantly associated with hospital neonatal mortality, Using a cut off point of 4, CRIB II score was found to have a sensitivity of 80,6%, specificity of 75.3%, and a predictive value of 77.7% compared to 72,5%, 71.2%, and 71,8% respectively for birthweight. Gestational age was found to have even lower figures; 56%, 75%, 66% for sensitivity, specificity and predictive values respectively, Areas under ROC curve were found to be 0,692,0,608, and 0,682 for CRIB II score, birthweight and gestational age respectively, Discussion: The sensitivity, specificity and predictive value of CRIB II score were found to be better than any of the traditional models independently, The area under the receiveroperating characteristic (ROC) curve for predicting death was greater for CRIB II score than for birthweight or gestational age alone, CRIB II cut off point of 4 was found to be optimal for predicting mortality, Conclusion: CRIB II score is a better predictor of hospital neonatal mortality among LBW babies at KNH than birthweight and gestational age independently, Based on these findings, we recommend that CRIB II score be included in the routine assessment of newborns admitted at the newborn unit of Kenyatta National Hospital.