Bootstrap to Study Completeness of Adult Mortality Surveillance

Bhamidipati Narasimhamurthy, National Institute of Epidemiology
M. D. Gupte, National Institute of Epidemiology
A. K. Mathai, National Institute of Epidemiology
S. Jabbar, National Institute of Epidemiology
S. Venkata Subramanian, National Institute of Epidemiology

Surveillance of mortality in India is being carried out through vital registration system. Though there exists an act of law, underreporting of deaths is common because of lapse in reporting, and dysfunction of the system. Incomplete and inadequate registration of deaths can often affect estimates of disease burden. Further, differential under-registration of deaths by age and sex will bias both mortality and synthetic analyzes of disease burden.The present paper attempts to (i) estimate the extent of under registration of deaths by age (5 years and above) and sex and (ii) construct life tables after correcting for under registration of deaths using nonparametric bootstrap. The results suggest that the extent of under registration varies from 20% to 30% for deaths in middle ages. Further, the expectation of life at age 5 is biased upward by two years for females and by five years for males due to age misreporting.