A0760
Title: Comparison of statistical methods for analyzing missing survival times in studies of census and death registry data
Authors: Wei-Ting Hwang - University of Pennsylvania (United States) [presenting]
Arielle Marks-Anglin - University of Pennsylvania (United States)
Frances Barg - University of Pennsylvania (United States)
Abstract: Observational studies that evaluate the effect of risk factors on mortality are sometimes limited to population registries for outcome information. Such studies could have missing death dates, particularly when the population under study is identified by census, where the distributions of age and other risk factors that impact mortality vary significantly. To date, there are no well-developed statistical methods to address this situation. Motivated by this gap in methodology, we conduct simulation studies to compare the performance of four analytical approaches (complete case analysis, censoring at the age of the census, inverse probability weighting, and multiple imputation of the residual lifetime), in estimating the exposure effects and median survival times in the presence of missing survival times. We also explore the effects of different missing data mechanisms (e.g., MCAR, MAR, MNAR), exposure-survival associations, censoring and missing percentages on their performance. The methods are applied to a cohort of the residents in Ambler, PA established using the 1930 US census, from which only 2,440 out of the total 4,514 individuals had death records retrievable from publicly available data sources and death certificate. Using this cohort, we examine the effects of occupational and paraoccupational asbestos exposure on survival and disparities in mortality by race and gender.
