Two-Step Imputation of Linked National Health Interview Survey and Medicare Data Files

• Activity Number: 23
• Type: Topic Contributed
• Date/Time: Sunday, August 4, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Survey Research Methods Section
• Author(s): Guangyu Zhang*+ and Jennifer D. Parker and Nathaniel Schenker
• Companies: National Center for Health Statistics and National Center for Health Statistics and National Center for Health Statistics
• Keywords: record linkage ; multiple imputation ; NHIS ; Medicare

Abstract

Record linkage is a valuable tool for combining information from different data sources. The National Center for Health Statistics has developed a record linkage program to link the center's population-based surveys with administrative data, including Medicare. However, not all survey participants provide key information for record linkage. In addition, for Medicare linkages, data are available for the Fee-for-Service program, but less consistently available for the managed care programs, such as Medicare Advantage. In this talk we discuss multiple imputation of missing data in linked National Health Interview Survey (NHIS)-Medicare files. We study mammography status based on Medicare claims for women 65 years and older. In our study, mammography and Medicare Advantage status are missing for NHIS respondents not linked to Medicare; and mammography status is missing for some linked respondents who have Medicare Advantage coverage. To address both issues, we first impute mammography status for unlinked respondents; then we impute Medicare Advantage status based on the imputed mammography status. We conduct simulations and apply our method to the linked NHIS-Medicare files.