Current Pediatric Research

Journal Banner

Child Survival Model in Reference to First Birth Order: Statistical Validation and Prediction

Child mortality is very high in India and it varies from state to state. Development of child survival models are lacking for these states. Hence, there is an immense importance and need to develop these models to design better and efficient health care systems and managements to prevent child mortality in India. The main aim of this study was to develop a child survival model to validate and predict survival probabilities for future perspectives. A sample of 627 children for the first birth order born during the four years preceding the survey was included in this study. The complete details of the children have been taken from the National Family Health Survey (NFHS-1992) for ever married women aged 13-49 years of age. Multivariate Cox PH analysis revealed that breastfeeding and immunization were the significant protective factors for child survival. Two hundred random samples with replacement from original sample were taken for the validation of the developed model. Shrinkage coefficient and Somer’s Dxy rank correlation were 78% and -0.89 having only 22% noise in the model. Validation Indices were found good enough for internal validity and the resulted model is found to be adequate to get accurate predictive survival probabilities. Therefore, this proposed model may be used by health policy planners for better child health care management in the state.

Author(s): Rajvir Singh, Abdulbari Bener, , Kalpana Singh, S.N. Dwivedi