Modelling weight of a Newborn based on baby's characteristics for low birth weight Babies

Abstract

Neonatal mortality rate (NMR) is an increasingly important public health issues in many developing countries. Neonatal death now accounts for about two-thirds of the eight million infant deaths that occur globally each year. It is well-documented that low birth weight (LBW) is the most significant factor influencing NMR. This paper deploys multi-regression models to identify the significant factors for forecasting the weight of the LBW babies. The model explores the relationship between weights, baby's characteristics, gestation age and mother pre-pregnancy BMI. The results indicate that 65.9% of the variations in the weight of the LBW babies can be explained by baby's characteristics such as the height, head and chest circumferences, the gestation age and mother's BMI. The proposed model was then used to estimate the recorded weights together with their corresponding 95% confidence and predication interval. Analysis of the prediction errors shows that the mean prediction error for the recorded data is one gram. The research is based on a case study in Indonesia intended to improve the mortality rate.