Evaluating Privacy-Preserving Record Linkage in Merging Cancer Registry Data and Electronic Health Records

Privacy-preserving record linkage (PPRL) can improve the utility of clinical data for research and discovery by merging disparate data sources to provide a more complete record of a patient. We assess the performance of record linkage across two data sets consisting of cancer registry data and electronic health records (EHRs). Using two data sources-EHR extracts and records from the Missouri Cancer Registry (MCR)-we performed various validation exercises, including an audit by the MCR and a comparative evaluation using multiple matching algorithms. Results indicated substantial matching performance, with high recall and precision rates. We conclude that the refined data linkage process significantly supports enhanced research capabilities without compromising patient privacy, thereby enabling reliable data integration for clinical research in oncology.