Record linkage: making the most out of errors in linking variables

Area of Science:

• Health Informatics
• Biostatistics
• Data Linkage

Background:

• Accurate medical record linkage is crucial for epidemiological studies.
• Administrative data often contain errors affecting linkage accuracy.
• Probabilistic record linkage algorithms traditionally struggle with minor data discrepancies.

Purpose of the Study:

• To refine a probabilistic medical record linking algorithm.
• To introduce a "close agreement" parameter to handle common administrative data errors.
• To evaluate the impact of "close agreement" on linkage accuracy and reduce uncertain links.

Main Methods:

• Developed a probabilistic record linking algorithm incorporating a "close agreement" category.
• Applied the algorithm to link early pregnancy determinants with late child outcomes.
• Assessed the reduction in uncertain links and the discriminating power of the linking key.

Main Results:

• The addition of "close agreement" for postal code and date of birth significantly improved linkage.
• Achieved a 95% reduction in the number of pairs with an uncertain linking status.
• Demonstrated that a "close" agreement outcome substantially enhances probabilistic record linkage.

Conclusions:

• The "close agreement" extension is a major improvement for probabilistic record linkage studies.
• This refinement effectively handles typical errors in administrative variables.
• Expected to yield similar improvements in other studies using large databases with comparable error types.