MedPlan: A Two-Stage RAG-Based System for Personalized Medical Plan Generation
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908836917411840 |
|---|---|
| author | Hsu, Hsin-Ling Dao, Cong-Tinh Wang, Luning Shuai, Zitao Phan, Thao Nguyen Minh Ding, Jun-En Liao, Chun-Chieh Hu, Pengfei Han, Xiaoxue Hsu, Chih-Ho Luo, Dongsheng Peng, Wen-Chih Liu, Feng Hung, Fang-Ming Wu, Chenwei |
| author_facet | Hsu, Hsin-Ling Dao, Cong-Tinh Wang, Luning Shuai, Zitao Phan, Thao Nguyen Minh Ding, Jun-En Liao, Chun-Chieh Hu, Pengfei Han, Xiaoxue Hsu, Chih-Ho Luo, Dongsheng Peng, Wen-Chih Liu, Feng Hung, Fang-Ming Wu, Chenwei |
| contents | Despite recent success in applying large language models (LLMs) to electronic health records (EHR), most systems focus primarily on assessment rather than treatment planning. We identify three critical limitations in current approaches: they generate treatment plans in a single pass rather than following the sequential reasoning process used by clinicians; they rarely incorporate patient-specific historical context; and they fail to effectively distinguish between subjective and objective clinical information. Motivated by the SOAP methodology (Subjective, Objective, Assessment, Plan), we introduce \ours{}, a novel framework that structures LLM reasoning to align with real-life clinician workflows. Our approach employs a two-stage architecture that first generates a clinical assessment based on patient symptoms and objective data, then formulates a structured treatment plan informed by this assessment and enriched with patient-specific information through retrieval-augmented generation. Comprehensive evaluation demonstrates that our method significantly outperforms baseline approaches in both assessment accuracy and treatment plan quality. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_17900 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | MedPlan: A Two-Stage RAG-Based System for Personalized Medical Plan Generation Hsu, Hsin-Ling Dao, Cong-Tinh Wang, Luning Shuai, Zitao Phan, Thao Nguyen Minh Ding, Jun-En Liao, Chun-Chieh Hu, Pengfei Han, Xiaoxue Hsu, Chih-Ho Luo, Dongsheng Peng, Wen-Chih Liu, Feng Hung, Fang-Ming Wu, Chenwei Computation and Language Despite recent success in applying large language models (LLMs) to electronic health records (EHR), most systems focus primarily on assessment rather than treatment planning. We identify three critical limitations in current approaches: they generate treatment plans in a single pass rather than following the sequential reasoning process used by clinicians; they rarely incorporate patient-specific historical context; and they fail to effectively distinguish between subjective and objective clinical information. Motivated by the SOAP methodology (Subjective, Objective, Assessment, Plan), we introduce \ours{}, a novel framework that structures LLM reasoning to align with real-life clinician workflows. Our approach employs a two-stage architecture that first generates a clinical assessment based on patient symptoms and objective data, then formulates a structured treatment plan informed by this assessment and enriched with patient-specific information through retrieval-augmented generation. Comprehensive evaluation demonstrates that our method significantly outperforms baseline approaches in both assessment accuracy and treatment plan quality. |
| title | MedPlan: A Two-Stage RAG-Based System for Personalized Medical Plan Generation |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2503.17900 |