Application of large language models in clinical record correction : a comprehensive study on various retraining methods

Objectives: We evaluate the effectiveness of large language models (LLMs), specifically GPT-based (GPT-3.5 and GPT-4) and Llama-2 models (13B and 7B architectures), in autonomously assessing clinical records (CRs) to enhance medical education and diagnostic skills. Materials and Methods: Various tec...

Descripción completa

Detalles Bibliográficos
Autores: Maitin, Ana M., Maitín, Ana María, Nogales, Alberto, Fernández-Rincón, Sergio, Nogales Moyano, Alberto, Aranguren, Enrique, Cervera-Barba, Emilio, Denizon-Arranz, Sophia, Mateos-Rodríguez, Alonso, Cervera Barba, Emilio Juan, García-Tejedor, Álvaro J., Denizon Arranz, Sophia, Mateos-Rodríguez, Alonso A., García Tejedor, Álvaro José
Tipo de recurso: artículo
Fecha de publicación:2025
País:España
Institución:Universidad Francisco de Vitoria
Repositorio:DDFV. Repositorio Institucional de la Universidad Francisco de Vitoria
Idioma:inglés
OAI Identifier:oai:ddfv.ufv.es:10641/7515
Acceso en línea:https://hdl.handle.net/10641/7515
Access Level:acceso abierto
Palabra clave:LLMs
artificial intelligence
clinical records
retraining
Health Informatics
Journal Article
Research Support, Non-U.S. Gov't
Yes
yes
Descripción
Sumario:Objectives: We evaluate the effectiveness of large language models (LLMs), specifically GPT-based (GPT-3.5 and GPT-4) and Llama-2 models (13B and 7B architectures), in autonomously assessing clinical records (CRs) to enhance medical education and diagnostic skills. Materials and Methods: Various techniques, including prompt engineering, fine-tuning (FT), and low-rank adaptation (LoRA), were implemented and compared on Llama-2 7B. These methods were assessed using prompts in both English and Spanish to determine their adaptability to different languages. Performance was benchmarked against GPT-3.5, GPT-4, and Llama-2 13B. Results: GPT-based models, particularly GPT-4, demonstrated promising performance closely aligned with specialist evaluations. Application of FT on Llama-2 7B improved text comprehension in Spanish, equating its performance to that of Llama-2 13B with English prompts. Low-rank adaptation significantly enhanced performance, surpassing GPT-3.5 results when combined with FT. This indicates LoRA’s effectiveness in adapting open-source models for specific tasks. Discussion. While GPT-4 showed superior performance, FT and LoRA on Llama-2 7B proved crucial in improving language comprehension and task-specific accuracy. Identified limitations highlight the need for further research. Conclusion: This study underscores the potential of LLMs in medical education, providing an innovative, effective approach to CR correction. Low-rank adaptation emerged as the most effective technique, enabling open-source models to perform on par with proprietary models. Future research should focus on overcoming current limitations to further improve model performance.