The Kaplan Meier estimates of mortality and its predictors among newborns admitted with low birth weight at public hospitals in Ethiopia

Abstract

Background Globally, every year greater than twenty million newborns are born weighing less than 2500grams, of which over 96% of them were in developing countries. It contributes to sixty to eighty percent of all neonatal deaths, annually. These infants weighing less than 2500gram will have a poor immune system and are at risk of morbidity and mortality. This study was aimed to investigate the survival status and predictors of mortality among neonates admitted with low birth weight at public hospitals in Ethiopia. Method A prospective cohort study was conducted among selected 216 newborns admitted with low birth weight who were admitted in neonatal intensive care unit at Arba Minch General hospital, Sawla General Hospital and Chencha district hospital from 1st March 2018 to 28st February 2019. Data were entered into Epi data version 3.02 and exported to STATA V 14 for analysis. The Kaplan Meier survival curve together with log rank test was used to estimate the survival time of the newborns. Variables which had p-value <0.05 in multivariable analysis using the cox proportional hazard model were declared as statistically significant. Result The cumulative proportion of surviving at the end of third days was 99.01% (95%CI: 96.11, 99.75). Similarly it was 97.81% (95%CI: 94.25, 99.18), 96.27% (95%CI: 91.76, 98.33) and 94.33% (95%CI: 88.72, 97.19) at the end of fourth, fifth and sixth day respectively. The overall mean survival time was 17.13 days (95%CI: 12.76, 21.49). The incidence of mortality among neonates admitted with low birth weight was 83 per 1000 live births. Mothers with DM (AHR:4.79; 95%CI:1.15, 19.8), maternal HIV infection(AHR:6.47;95%CI:1.43,29.3), not keeping the newborn under KMC(AHR:13.0;95%CI:3.42, 49.5) and initiating exclusive breast feeding within one hour(AHR:0.19;95%CI:0.04, 0.95) were statistically significant in multivariable cox regression analysis. Conclusion The risk of mortality among neonates admitted with low birth weight was high at the early admission period and the incidence of mortality was high. Maternal history of diabetes mellitus, HIV/AIDS infection, keeping the newborn under kangaroo mother care and exclusive breast feeding initiation were statistically significant predictors of mortality. Special emphasis should be given for neonates with low birth weight at the early follow up periods and Kangaroo mother care practice should be a usual practice.