IPF-RMF: Intelligent Patient Follow-up Supported by RAG and Multi-Model Fusion

EasyChair Preprint 15757

Abstract

In modern healthcare services, intelligent patient follow-up is a critical approach to improving the quality of medical services and the efficiency of patient health management. Our study proposes an intelligent patient follow-up method based on Retrieval Augmented Generation (RAG) and multi-model fusion.  Patient data and case information are collected using custom forms, forming an RAG-supported retrieval database to store follow-up records and relevant details. A multi-model framework was designed, using machine learning algorithms to predict key information such as follow-up schedules, and leveraging multiple large language models to generate initial follow-up recommendations. A decision-making large language model was utilized to integrate the initial follow-up recommendations from various language models, optimizing and developing the final personalized follow-up plan. Manual assessments were conducted to comprehensively analyze the quality of the final follow-up plan in terms of readability, professionalism, and other dimensions to evaluate the proposed method. Experimental results demonstrate that the proposed method significantly enhances the scientific validity and personalization of the follow-up plan, providing a reliable technical foundation for intelligent health management.

Keyphrases: Intelligent Patient Follow-Up, LLM, Retrieval Augmented Generation, intelligent patient follow, multi-model fusion, rag and multi model fusion, rag technology

@booklet{EasyChair:15757,
  author    = {Junyu Jia and Yue Qian and Zhe Shen and Zhi Wang and Dacheng Sang and Lanxin Yang and Lin Xu and Fulong Chen},
  title     = {IPF-RMF: Intelligent Patient Follow-up Supported by RAG and Multi-Model Fusion},
  howpublished = {EasyChair Preprint 15757},
  year      = {EasyChair, 2025}}