Online Program Home
My Program

Abstract Details

Activity Number: 402 - HPSS Student Paper Competition Winners: Statistics Advancing Policy
Type: Topic Contributed
Date/Time: Tuesday, July 31, 2018 : 2:00 PM to 3:50 PM
Sponsor: Health Policy Statistics Section
Abstract #327243
Title: Bayesian Record Linkage Under Limited Linking Information
Author(s): Mingyang Shan* and Roee Gutman and Kali Thomas
Companies: Brown University and Brown University and Brown University
Keywords: Record Linkage; Bayesian Analysis; Missing Data; Multiple Imputation

Record linkage is a statistical technique that tries to identify individuals or entities that exist in two or more data sources. This is a challenging task when no unique identification variable is present. Bayesian record linking procedures have been developed that primarily focus on classifying links using the information that exist between both data files. These procedures struggle in performance when the amount of linking variables are few in number and when the variables are prone to error. We introduce an adaptation to the existing Bayesian record linkage methodology that incorporates associations between variables that exist in only one file, in addition to those shared between both files, to extract additional linking accuracy when identifying information is limited. We show various ways to incorporate such information into the existing framework and apply our method to link Meals on Wheels recipients to Medicare Enrollment records.

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

Back to the full JSM 2018 program