An empirical comparison of record linkage procedures

Abstract

We consider the problem of record linkage in the situation where we have only non-unique identifiers, like names, sex, race etc., as common identifiers in databases to be linked. For such situations much work on probabilistic methods of record linkage can be found in the statistical literature. However, although many groups undoubtedly still use deterministic procedures, not much literature is available on deterministic strategies. Furthermore, there appears to exist almost no documentation on the comparison of results for the two strategies. In this work we compare a stepwise deterministic linkage strategy with a probabilistic strategy, as implemented in AUTOMATCH, for a situation in which the truth is known. The comparison was carried out on a linkage between medical records from the Regional Perinatal Intensive Care Centers database and educational records from the Florida Department of Education. Social security numbers, available in both databases, were used to decide the true status of each record pair after matching. Match rates and error rates for the two strategies are compared and a discussion of their similarities and differences, strengths and weaknesses is presented.

Impact of record-linkage methodology on performance indicators and multivariate relationships

“…Previous research suggests that the positive predictive value of all three linkage algorithms used in this research would in the low to high 90s Gomatam & Carter, 1999;Gomatam et al, 2002;Grannis et al, 2002) and that sensitivity is likely to range from the mid-80s (match-merge) to the mid-90s (probabilistic).…”

“…The FOP has adopted a Common Patient Identifier (CPI) constructed from the following data elements: gender, birth date, birthplace, first 3 characters of first name, and first 3 characters of last name. More complex deterministic algorithms are discussed in the literature Gomatam et al, 2002;Grannis et al, 2002).…”

Quality and Complexity Measures for Data Linkage and Deduplication

“…In practise, therefore, deterministic rule based systems are limited to ad-hoc linkages of smaller data sets. In a recent study [20] an iterative deterministic linkage system has been compared with the commercial probabilistic system AutoMatch [27]. Empirical results showed that the probabilistic approach resulted in better linkage quality.…”

Using an epidemiological approach to examine outcomes affecting young children with Down syndrome and their families

“…Two general strategies are used for these designations: deterministic and probabilistic (Fellegi & Sunter, 1969;Gomatam, Carter, Ariet & Mitchell, 2002;Jaro, 1995). Deterministic linkage strategies examine the number of agreeing identifiers in a pair of records and designate the one with the most agreements (above a minimum number of agreements) as the matching pair; all other pairs are non-matches.…”