Related Experiment Videos
Avoiding doubles in distributed nominative medical databases: optimization of the needleman and wunsch algorithm.
Loïc Le Mignot1, Claude Mugnier, Mohamed Ben Saïd
1Université Paris-Descartes, Faculté de Médecine; Assistance Publique-Hôpitaux de Paris; Hôpital Necker, EA222, Service de Biostatistique et d'Informatique Médicale, Hôpital Necker, 149 rue de Sèvres, Paris, France.
Studies in Health Technology and Informatics
|September 15, 2005
Summary
Patient identification errors in health systems create duplicate records. The Needleman-Wunsch algorithm efficiently detects these duplicates in distributed databases, improving data accuracy.
Area of Science:
- Health Informatics
- Bioinformatics
- Database Management
Background:
- Patient identification errors are a significant challenge in public health information systems.
- Duplicate patient records hinder accurate trajectory reconstitution and data analysis.
- Distributed web databases exacerbate the issue of managing unique patient identities.
Purpose of the Study:
- To evaluate the efficacy of the Needleman-Wunsch (N&W) algorithm for detecting duplicate patient records in distributed databases.
- To develop and implement optimized variants of the N&W algorithm for a multi-source information system.
- To assess the performance of N&W algorithm variants in identifying patient record duplications within a simulated dataset.
Main Methods:
- Explored the Needleman-Wunsch (N&W) algorithm for string matching optimization.
- Developed five variants of the N&W algorithm.
- Implemented algorithms within a web-based Multi-Source Information System for End-Stage Renal Disease patient tracking.
- Utilized a simulated database of 73,210 records with simulated character insertions/suppressions.
Main Results:
- A 2% rate of double entries was observed with an acceptable distance set to 5 modifications.
- The N&W algorithm demonstrated sensitive and specific detection with acceptable processing times.
- The algorithm detected up to 10% of modifications, exceeding the estimated error rate.
- A 'cut-off heuristic' variant proved particularly efficient for identifying duplicate entries.
Conclusions:
- The Needleman-Wunsch algorithm, particularly the 'cut-off heuristic' variant, is effective for detecting duplicate patient records in distributed databases.
- Optimized N&W algorithm variants enhance the accuracy and reliability of patient tracking in health information systems.
- This approach improves data integrity for managing patient trajectories, especially for chronic diseases like End-Stage Renal Disease.