Error Adjustments for File-Linking Methods Using Encrypted Unique Client Identifier (eUCI) with Application to Recently Released HIV+ Prisoners
*Roee Gutman, Brown University
Keywords: File Linking, Bayesian, Multiple Imputation, eUCI
Incarceration provides an opportunity to test for HIV, provide treatment, and link infected persons to comprehensive HIV care upon their release. Linkage time is a key factor in assessing the success of a program that links released individuals to care. To estimate this time, records from correction systems are linked to Ryan White Clinic data using eUCI. Generally, eUCI matching identifies most of the true records in the two data sources; however, it also links records incorrectly, or misses records that should have been linked. We propose a Bayesian procedure that relies on the relationships between variables that exist in either of the data sources and variables that exist in both to identify correctly linked records among all linked records. The procedure generates K data sets in which each pair of linked records is identified as a true or false link. The K data sets are analyzed independently and the results are combined using Rubin’s multiple imputation rules. A validation study is used to identify an appropriate model and to inform the prior distributions of the parameters. The proposed method is flexible, efficient, and can be applied to other file-linking applications.