Related Experiment Videos
Practical introduction to record linkage for injury research
1Center for Outcomes Research and Evaluation, Maine Medical Center and the Harvard Injury Control Research Center, Harvard School of Public Health, USA. clarkd@mmc.org
Summary
Linking multiple health databases enhances injury research and trauma system evaluation. This review covers record linkage methods, mathematical foundations, practicalities, and ethical considerations for robust population-based studies.
Area of Science:
- Public Health
- Biostatistics
- Health Informatics
Background:
- Single databases are insufficient for population-based injury research due to early fatalities and transient emergency care.
- Linking multiple data sources is crucial for comprehensive injury surveillance and trauma system evaluation.
Purpose of the Study:
- To review the historical development of record linkage techniques.
- To provide a mathematical foundation for record linkage methods.
- To discuss practical issues and ethical concerns associated with data source integration.
Main Methods:
- Overview of deterministic record linkage (clerical and computer-assisted).
- Explanation of probabilistic record linkage, simulating human reasoning for record comparison.
- Discussion on combining deterministic, probabilistic, and clerical methods based on study objectives.
Main Results:
- Probabilistic methods are particularly effective for large-scale studies.
- The choice of method (deterministic vs. probabilistic) depends on whether the goal is statistical estimation or individual inference.
- Careful clerical review enhances the positive predictive value for individual-level data linkage.
Conclusions:
- Effective population-based injury research necessitates linking multiple data sources.
- Understanding the mathematical and practical aspects of record linkage is essential for accurate surveillance and evaluation.
- Integrating data sources introduces significant ethical responsibilities for researchers.