Record Linkage of Health Care Insurance Claims
- Affiliations of the authors: Healthcare Informatics (TWV) and Health Economics and Outcomes Research (RMM), SmithKline Beecham, Collegeville, Pennsylvania
- Correspondence and reprints: Timothy W. Victor, Assistant Director, Research and Biostatistics, Healthcare Informatics, SmithKline Beecham, MS UP4305, 1250 South Collegeville Road, Collegeville PA 19426-2990; e-mail: 〈timothy.w.victor{at}sbphrd.com〉
- Received 19 September 2000
- Accepted 9 January 2001
Abstract
Objective This paper provides a detailed description of a method developed for purposes of linking records of individual patients, represented in diverse data sets, across time and geography.
Design The procedure for record linkage has three major components—data standardization, weight estimation, and matching. The proposed method was designed to incorporate a combination of exact and probabilistic matching techniques.
Measurements The procedure was validated using convergent, divergent, and criterion validity measures.
Results The output of the process achieved a sensitivity of 92 percent and a specificity that approached 100 percent.
Conclusions The procedure is a first step in addressing the current trend toward larger and more complex databases.
Footnotes
-
↵* An arbitrary number which is large enough such that the sample is representative of the dataset, but not so large as to become computationally prohibitive.
