Programmatic Theme: Clinical Research Informatics

Abstract: Record linkage refers to a family of computational methods that determine if two or more electronic health records belong to the same individual using non-unique demographics data elements, called linkage variables. Because the quality and characteristics of linkage variables directly impact the performance of linkage methods, the objective of this work is to develop linkability measures that evaluate the potential capability of variables in datasets to be used for linking records.

Learning Objective: - Understand the concept of linkability in record linkage and the impact of the data quality on linkability
- Understand methods to perform synthetic linkage data corruption

Authors:

Toan Ong, University of Colorado Anschutz Medical Campus
Andrew Hill (Presenter)
University of Colorado Denver

Shaun Grannis, Indiana University
Chan Voong, University of Colorado Anschutz Medical Campus
Michael Kahn, University of Colorado Anschutz Medical Campus