Comparison Between The Application Of Linkage Probabilistic Methods: Openreclink And R

LATORRE M. ^1,2 , TANAKA L. ^1,2 , TAHARA E. ¹

Purpose: The database relationship, also known as linkage, is frequently used in the studies of database management and its usage (duplicate identification cases, the completeness of the database, among others) has shown improvement in quality. However, there are several relationship techniques (probabilistic and deterministic), without a consensus, on what would be the best choice. The aim is to compare two linkage probabilistic methods, analyzing the time of each procedure, the number of true pairs found in each process and the ease of handling between the two procedures to be applied in a workplace, especially in cancer registries.
Methods: The information contained in the PROAIM database (Improvement Program for Mortality Information in São Paulo) and the RCBP-SP database (Population-Based Cancer Registry of São Paulo) related to ovarian cancer were used. The Openreclink and R computer programs were used for the linkage process.
Results: 2,944 true pairs in Openreclink were found in a timespan of 04h58min31s. On R, 1,719 true pairs were found in a timespan of 3h18min00s. The Openreclink program has a routine and a simpler understanding, which enhances and speeds its application in the work environment, such as the Population-Based Cancer Registry of São Paulo. However the R program, despite being more complex to understand (longer demand to understand the logic of the program and define the commands to be used), its steps are executed more quickly.
Conclusion: Although the R program is less time-consuming, the Openreclink program is more effective.
Funding Source:  Support Foundation of São Paulo Research