English  Français

Estimation of Infant and Child mortality across districts of India over the past 30 years: A cross-sectional birth history data application

Sourav Dey, International Institute for Population Sciences (IIPS)
Suryakant Yadav, International Institute for Population Sciences (IIPS)

Infant and Child Mortality is the most direct and vital indicator of health at the population level for the second most populous country. For preventing and controlling child deaths in India, we need to get the deeper into the throat (at district level) for the actual picture. Through Bayesian Discrete hazard model, all the infant and child mortality between 1985-2015/16 have been estimated from NFHS 2015-16 birth history information. Again, through Random walk models total variance of the observed data have also been decomposed and each of quantity will be the variance due to random effect. Indicators corresponding to survey designs are survey strata, cluster, household, survey weight, date of interview, date of child birth as well as death and age at death. Districts from Northern India have more cluster child death over the time-point than the southern states districts. Through mortality scenario got scattered from post-natal mortality to infant mortality. Positive spatial correlation has also been found between districts and interestingly it remains almost similar over the period. Northern and eastern state districts have least variance while southern state districts have highest variance in the child mortality estimate.

Keywords: Bayesian methods / estimation, Mathematical demography, Small area estimation

See extended abstract.

  Presented in Session P18.