Article Title

Record Linkage With Washington State Cancer Registry

Publication Date



linkage, cancer


Background/Aims: At Group Health, the virtual data warehouse’s (VDW) tumor registry table consists of tumor cases ascertained by Cancer Surveillance System (CSS) at the Fred Hutchinson Cancer Research Center (FHCRC). This data is limited to residents of 13 Puget Sound counties in the SEER region only. On the other hand, Washington State Cancer Registry (WSCR) consists of tumor cases for the entire Washington State residents. The poster describes the data linkage protocol and the results of linkage of Group Health members with WSCR database in the absence of unique identifier.

Methods: Group Health data consists of 1,420,334 members residing in Washington State. We linked Group Health data with WSCR database consisting of 613,631 cancer cases diagnosed during January 1, 1993–December 31, 2011. We used a probabilistic linkage software application such as CDC’s Registry Plus Link Plus® to link patient information such social security number, date of birth, gender, and last and first names with corresponding data elements in WSCR. We assessed the quality of the linkage results by comparing the linkage outcome with VDW’s tumor registry. We calculated the sensitivity and positive predictive value by comparing the WSCR linkage results with VDW’s tumor registry. We used Venn diagrams for visualization of results of comparison.

Results: Sensitivity and positive predictive value are 90% and 96%, respectively. We found 35,166 (59%) additional tumor cases as a result of WSCR linkage in comparison to the VDW tumor registry. Of the above, 21,600 (61%) tumor cases were reported within SEER region and 13,566 (39%) cases diagnosed in members residing outside of Puget Sound region. Of 21,600 SEER cases, 17,248 (88%) tumor cases were diagnosed after disenrollment from Group Health, and 2,090 SEER cases were found missing from VDW’s tumor registry. Due to lack of interpretable information, we were unsuccessful in linking 5,752 records in the VDW tumor registry with WSCR data.

Discussion: In the absence of unique identifier, the project demonstrates a successful linkage using Link Plus. We now have cancer incidence and survival data for all Group Health members, not limited to Puget Sound region. This will augment cancer studies involving members from both SEER region and outside of Puget Sound counties.




April 2nd, 2015


April 28th, 2015