Skip to content
DraconiticResearch
Search Sources API
197,761 studies31 sourcesUpdated 55 min ago
Peer-reviewedCC BYnpj Health Systems

MAP: evaluation and multi-agent enhancement of large language models for inpatient pathways

  • Zhen Chen1
  • Zhihao Peng2
  • Xusheng Liang2
  • Cheng Wang2
  • Peigan Liang3
  • Linsheng Zeng4
  • Minjie Ju5
  • Yixuan Yuan6
6 affiliations
  1. 1Department of Data Science and Artificial Intelligence, The Hong Kong Polytechnic University, Kowloon, Hong Kong SAR
  2. 2Department of Electronic Engineering, The Chinese University of Hong Kong, New Territories, Hong Kong SAR
  3. 3Guangzhou Hospital of Integrated Traditional Chinese and Western Medicine, Guangzhou, China
  4. 4Shenzhen Traditional Chinese Medicine Hospital, Shenzhen, China
  5. 5Department of Critical Care Medicine, Zhongshan Hospital Fudan University, Shanghai, China
  6. 6Department of Electronic Engineering, The Chinese University of Hong Kong, New Territories, Hong Kong SAR. [email protected]
Open at sourcePDFDOICite

Abstract

Inpatient pathways require complex clinical decision-making based on comprehensive patient information, yet research on medical LLMs is limited in this area due to the lack of large-scale datasets. Existing medical benchmarks primarily focused on question-answering and examinations, overlooking the multifaceted nature of inpatient decision-making. To address this gap, we developed the IPDS benchmark, comprising 51,274 cases across 9 triage departments, 17 major disease categories, and 16 treatment options. We further proposed the Multi-Agent Inpatient Pathways (MAP) framework, containing three specialized clinical agents: a triage agent for patient admission, a diagnosis agent for diagnostic decision-making, and a treatment agent for care planning. A chief agent guides and promotes these agents to ensure coordination. Experiments demonstrated that MAP achieved superior alignment with operational protocols compared to state-of-the-art LLMs. The MAP sets a foundation for advancing inpatient support systems, offering significant potential for enhancing operational efficiency and resource planning in healthcare facilities.

Related

  1. Self-harm incidents and restrictive practice use in an acute female inpatient unit: a service evaluation of a continuous multi-system improvement approach
    S. Johnson, M. Kelbrick, C. Griffiths +4
    medRxivPreprintCC BY-NC-NDFull textPsychiatry & mental healthmedRxiv
  2. Multi-site observational evaluation of Breast AI™-supported ultrasound risk stratification in breast cancer assessment pathways
    Kathryn Malherbe, Carol A. Benn, Gerhard Botha
    PLOS OnePeer-reviewedCC BY 4.0Full textMedical imaging & radiologyOncologyPLOS
  3. PharmaAgent: a real-world evaluation of high- and low-autonomy agentic systems for inpatient vancomycin dosing
    Yixing Jiang, Kameron C. Black, Timothy Keyes +11
    medRxivPreprintAll rights reservedFull textEpidemiology & public healthmedRxivPMC
  4. Evaluation of Resident Physicians Inpatient Clinical Performance Using a DRG-Based Weighted TOPSIS-RSR Model: A Single-Centre Pilot Study
    Duan Xiaoli, Wen Guohong
    Advances in Medical Education and PracticePeer-reviewedCC BY-NC 4.0Full textEpidemiology & public healthPMCEurope PMC
  5. Factors associated with the quality of urban geriatric healthcare services empowered by large language models and improvement pathways: a theoretical framework and cross-sectional study
    Xiaoyan Qi, Xuejiao Song, Jingkai Xu +3
    BMC Public HealthPeer-reviewedCC BY-NC-ND 4.0Epidemiology & public healthCrossref

Cite

BibTeXRIS
BibTeX
@article{chen2026evaluation,
  title = {{MAP: evaluation and multi-agent enhancement of large language models for inpatient pathways}},
  author = {Zhen Chen and Zhihao Peng and Xusheng Liang and Cheng Wang and Peigan Liang and Linsheng Zeng and Minjie Ju and Yixuan Yuan},
  journal = {npj Health Systems},
  year = {2026},
  month = jun,
  doi = {10.1038/s44401-026-00085-0},
  url = {https://doi.org/10.1038/s44401-026-00085-0},
  note = {CC BY},
}
RIS
TY  - JOUR
TI  - MAP: evaluation and multi-agent enhancement of large language models for inpatient pathways
AU  - Zhen Chen
AU  - Zhihao Peng
AU  - Xusheng Liang
AU  - Cheng Wang
AU  - Peigan Liang
AU  - Linsheng Zeng
AU  - Minjie Ju
AU  - Yixuan Yuan
PY  - 2026
DA  - 2026/06/01
JO  - npj Health Systems
DO  - 10.1038/s44401-026-00085-0
UR  - https://doi.org/10.1038/s44401-026-00085-0
AB  - Inpatient pathways require complex clinical decision-making based on comprehensive patient information, yet research on medical LLMs is limited in this area due to the lack of large-scale datasets. Existing medical benchmarks primarily focused on question-answering and examinations, overlooking the multifaceted nature of inpatient decision-making. To address this gap, we developed the IPDS benchmark, comprising 51,274 cases across 9 triage departments, 17 major disease categories, and 16 treatment options. We further proposed the Multi-Agent Inpatient Pathways (MAP) framework, containing three specialized clinical agents: a triage agent for patient admission, a diagnosis agent for diagnostic decision-making, and a treatment agent for care planning. A chief agent guides and promotes these agents to ensure coordination. Experiments demonstrated that MAP achieved superior alignment with operational protocols compared to state-of-the-art LLMs. The MAP sets a foundation for advancing inpatient support systems, offering significant potential for enhancing operational efficiency and resource planning in healthcare facilities.
SN  - 3005-1959
ER  -