Abstract:
|
The National Center for Health Statistics recently updated the National Health Interview Survey (NHIS)-Linked Mortality Files. Eligible respondents in the 1986-2009 NHIS were linked to the National Death Index through probabilistic record linkage to obtain mortality follow-up through Dec 31, 2011. To assess how well mortality was ascertained, cumulative survival probabilities overall and by interview age, race, Hispanic origin, and sex were derived for each annual NHIS cohort using the Kaplan-Meier product limit method and compared with corresponding probabilities computed from annual U.S. life tables. Additionally, cause-of-death distributions were compared to those derived from U.S. vital statistics. Our analyses also address the impacts of length of follow-up, reduced NHIS response rates, and use of 4-digit instead of 9-digit social security number for linkage. Assessment of the overall and relative accuracy of the mortality ascertainment achieved in the NHIS-Linked Mortality Files is important because the files are frequently used by researchers to measure mortality associations for health outcomes and population subgroups.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.