Online Program Home
My Program

Abstract Details

Activity Number: 122 - Clinical Trial Design and Missing Data
Type: Contributed
Date/Time: Monday, July 30, 2018 : 8:30 AM to 10:20 AM
Sponsor: Biopharmaceutical Section
Abstract #330411 Presentation
Title: Linking Medicare Current Beneficiary Survey (MCBS) to Augment Post-Market Real World Data from Medicare Claims: a Multiple Imputation Approach
Author(s): Yun Lu* and Xiyuan Wu and Yoganand Chillarige and Michael Wernecke and Hector Izurieta and Jeffrey Kelman and Richard Forshee
Companies: FDA and Acumen LLC and Acumen LLC and Acumen LLC and FDA and CMS and FDA
Keywords: Multiple Imputation; Medicare Current Beneficiary Survey (MCBS); Medicare Data; Post-Market; Real World Evidence (RWE))
Abstract:

With the increasing importance placed by the FDA on real world evidence of regulated products, it is essential to address bias due to unmeasured confounding. Medicare claims data, which are frequently used by the FDA/CBER to investigate post-market vaccine safety and effectiveness, do not contain comprehensive information on behavioral and frailty measures. This study piloted an approach to augment Medicare claims with the Medicare Current Beneficiary Survey (MCBS), a continuous survey of a nationally representative sample of the Medicare population. We linked a Medicare Herpes Zoster vaccine effectiveness study population to MCBS to obtain additional information such as health seeking behavior. Since MCBS comprises < 1% of the Medicare population, to fully utilize MCBS information, we used multiple imputation by chained equations to impute potential confounders. We observed imbalances for some MCBS variables between the matched cohorts from the original study, and compared analysis results with and without imputation. This multiple imputation approach using linked Medicare and MCBS data provides a potential solution to address bias issues related to unmeasured confounding.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2018 program