Leveraging Language Models for Automated Patient Record Linkage

2025-04-21

AI Summary

Fragmented medical data severely impedes cross-institutional patient record linkage, necessitating efficient and robust automated entity resolution techniques. This study presents the first systematic evaluation of diverse language models—including RoBERTa, Mistral-7B/24B, and DeepSeek-R1—for both blocking and matching tasks in real-world medical record linkage, under supervised fine-tuning and zero-shot inference paradigms. We innovatively integrate sentence embeddings with large language model capabilities to replace conventional probabilistic or rule-based approaches. Results show that fine-tuned RoBERTa achieves near-perfect blocking recall (≈100%), reducing candidate pairs by 92%; fine-tuned Mistral-7B yields only six matching errors; and zero-shot Mistral-Small-24B commits just 55 errors—substantially outperforming generic baselines. The work characterizes fundamental trade-offs among accuracy, computational efficiency, and scalability in large-model-based entity linking, establishing a novel paradigm for healthcare entity resolution.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Data Mining & Knowledge Management: Linked Open Data, Knowledge Graphs & KB Completion

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Objective: Healthcare data fragmentation presents a major challenge for linking patient data, necessitating robust record linkage to integrate patient records from diverse sources. This study investigates the feasibility of leveraging language models for automated patient record linkage, focusing on two key tasks: blocking and matching. Materials and Methods: We utilized real-world healthcare data from the Missouri Cancer Registry and Research Center, linking patient records from two independent sources using probabilistic linkage as a baseline. A transformer-based model, RoBERTa, was fine-tuned for blocking using sentence embeddings. For matching, several language models were experimented under fine-tuned and zero-shot settings, assessing their performance against ground truth labels. Results: The fine-tuned blocking model achieved a 92% reduction in the number of candidate pairs while maintaining near-perfect recall. In the matching task, fine-tuned Mistral-7B achieved the best performance with only 6 incorrect predictions. Among zero-shot models, Mistral-Small-24B performed best, with a total of 55 incorrect predictions. Discussion: Fine-tuned language models achieved strong performance in patient record blocking and matching with minimal errors. However, they remain less accurate and efficient than a hybrid rule-based and probabilistic approach for blocking. Additionally, reasoning models like DeepSeek-R1 are impractical for large-scale record linkage due to high computational costs. Conclusion: This study highlights the potential of language models for automating patient record linkage, offering improved efficiency by eliminating the manual efforts required to perform patient record linkage. Overall, language models offer a scalable solution that can enhance data integration, reduce manual effort, and support disease surveillance and research.

Problem

• Healthcare data fragmentation challenges patient record linkage.
• Investigates language models for automated blocking and matching tasks.
• Compares model performance against traditional probabilistic linkage methods.

Innovation

• Fine-tuned RoBERTa for blocking via sentence embeddings
• Mistral-7B achieved best matching performance when fine-tuned
• Hybrid rule-based and probabilistic approach outperformed language models