Related Experiment Videos
Determinants of homonym and synonym rates of record linkage in disease registration
1Department of Epidemiology, University of Ulm, Germany.
Insights
Accurate record linkage is crucial for disease registries. This study algebraically assesses homonym and synonym rates, identifying factors influencing linkage errors for better quality assessment.
Area of Science:
- Health Informatics
- Biostatistics
- Public Health
Background:
- Reliable record linkage is fundamental for high-quality population-based disease registration.
- Assessing disease registry quality necessitates quantitative methods to evaluate record-linkage errors.
- Homonym and synonym rates are proposed metrics for quantifying linkage errors.
Purpose of the Study:
- To provide an algebraic assessment of the determinants of homonym and synonym rates in disease registries.
- To elucidate the factors influencing the accuracy of record linkage in population-based health data.
- To inform the interpretation and comparison of record-linkage study results across different settings.
Main Methods:
- Algebraic analysis of homonym and synonym rate determinants.
- Evaluation of the impact of personal identifier discriminating power and linkage procedures.
- Consideration of disease registration specific circumstances like registry size and notification frequency.
Main Results:
- Homonym and synonym rates are determined by the discriminating power of personal identifiers and the linkage procedure.
- Factors such as the number of patients and notifications per case significantly influence these error rates.
- Understanding these determinants is key to accurate interpretation of linkage study outcomes.
Conclusions:
- The study provides a framework for understanding and quantifying record-linkage errors in disease registries.
- Accurate reporting and interpretation of linkage study results require consideration of both technical and contextual factors.
- This algebraic assessment aids in comparing record-linkage procedure performance across diverse environments.
Abstract:
Reliable record linkage is an essential component of the quality of population-based disease registration. Quality assessment of disease registries should, therefore, include quantitative approaches to describe the extent of record-linkage errors. The homonym and synonym rates have been proposed for this purpose. The homonym rate quantifies the proportion of distinct patients excluded from registration due to erroneous linkage with other patients. The synonym rate quantifies the proportion of unrecognized duplicate notifications on patients already registered in the registry. This paper provides an algebraic assessment of the determinants of both rates. It is shown how the homonym and the synonym rate are determined by the discriminating power provided by the personal identifiers and the record-linkage procedure on the one hand and the specific circumstances of disease registration on the other hand, such as the number of patients in the registry or the number of notifications per case. All these factors should be taken into account when reporting and interpreting results of record-linkage studies, particularly if comparisons are made between the performance of record-linkage procedures in different environments.