Privacy-Preserving Clinical Decision Support for Emergency Triage Using LLMs: System Architecture and Real-World Evaluation

This study presents a next-generation clinical decision-support architecture for Clinical Decision Support Systems (CDSS) focused on emergency triage. By integrating Large Language Models (LLMs), Federated Learning (FL), and low-latency streaming analytics within a modular, privacy-preserving framew...

Full description

Saved in:
Bibliographic Details
Main Authors: Alper Karamanlıoğlu, Berkan Demirel, Onur Tural, Osman Tufan Doğan, Ferda Nur Alpaslan
Format: Article
Language:English
Published: MDPI AG 2025-07-01
Series:Applied Sciences
Subjects:
Online Access:https://www.mdpi.com/2076-3417/15/15/8412
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849407531181408256
author Alper Karamanlıoğlu
Berkan Demirel
Onur Tural
Osman Tufan Doğan
Ferda Nur Alpaslan
author_facet Alper Karamanlıoğlu
Berkan Demirel
Onur Tural
Osman Tufan Doğan
Ferda Nur Alpaslan
author_sort Alper Karamanlıoğlu
collection DOAJ
description This study presents a next-generation clinical decision-support architecture for Clinical Decision Support Systems (CDSS) focused on emergency triage. By integrating Large Language Models (LLMs), Federated Learning (FL), and low-latency streaming analytics within a modular, privacy-preserving framework, the system addresses key deployment challenges in high-stakes clinical settings. Unlike traditional models, the architecture processes both structured (vitals, labs) and unstructured (clinical notes) data to enable context-aware reasoning with clinically acceptable latency at the point of care. It leverages big data infrastructure for large-scale EHR management and incorporates digital twin concepts for live patient monitoring. Federated training allows institutions to collaboratively improve models without sharing raw data, ensuring compliance with GDPR/HIPAA, and FAIR principles. Privacy is further protected through differential privacy, secure aggregation, and inference isolation. We evaluate the system through two studies: (1) a benchmark of 750+ USMLE-style questions validating the medical reasoning of fine-tuned LLMs; and (2) a real-world case study (<i>n</i> = 132, 75.8% first-pass agreement) using de-identified MIMIC-III data to assess triage accuracy and responsiveness. The system demonstrated clinically acceptable latency and promising alignment with expert judgment on reviewed cases. The infectious disease triage case demonstrates low-latency recognition of sepsis-like presentations in the ED. This work offers a scalable, audit-compliant, and clinician-validated blueprint for CDSS, enabling low-latency triage and extensibility across specialties.
format Article
id doaj-art-b47cfd2cf12f427788c173e8208eacfc
institution Kabale University
issn 2076-3417
language English
publishDate 2025-07-01
publisher MDPI AG
record_format Article
series Applied Sciences
spelling doaj-art-b47cfd2cf12f427788c173e8208eacfc2025-08-20T03:36:02ZengMDPI AGApplied Sciences2076-34172025-07-011515841210.3390/app15158412Privacy-Preserving Clinical Decision Support for Emergency Triage Using LLMs: System Architecture and Real-World EvaluationAlper Karamanlıoğlu0Berkan Demirel1Onur Tural2Osman Tufan Doğan3Ferda Nur Alpaslan4Department of Computer Engineering, Middle East Technical University, Ankara 06800, TürkiyeProduct and R&D, TURKSAT, Ankara 06839, TürkiyeEmergency Medicine, Zonguldak Bülent Ecevit University, Zonguldak 67100, TürkiyeR&D and Innovation, Innova IT Solutions, Ankara 06800, TürkiyeDepartment of Computer Engineering, Middle East Technical University, Ankara 06800, TürkiyeThis study presents a next-generation clinical decision-support architecture for Clinical Decision Support Systems (CDSS) focused on emergency triage. By integrating Large Language Models (LLMs), Federated Learning (FL), and low-latency streaming analytics within a modular, privacy-preserving framework, the system addresses key deployment challenges in high-stakes clinical settings. Unlike traditional models, the architecture processes both structured (vitals, labs) and unstructured (clinical notes) data to enable context-aware reasoning with clinically acceptable latency at the point of care. It leverages big data infrastructure for large-scale EHR management and incorporates digital twin concepts for live patient monitoring. Federated training allows institutions to collaboratively improve models without sharing raw data, ensuring compliance with GDPR/HIPAA, and FAIR principles. Privacy is further protected through differential privacy, secure aggregation, and inference isolation. We evaluate the system through two studies: (1) a benchmark of 750+ USMLE-style questions validating the medical reasoning of fine-tuned LLMs; and (2) a real-world case study (<i>n</i> = 132, 75.8% first-pass agreement) using de-identified MIMIC-III data to assess triage accuracy and responsiveness. The system demonstrated clinically acceptable latency and promising alignment with expert judgment on reviewed cases. The infectious disease triage case demonstrates low-latency recognition of sepsis-like presentations in the ED. This work offers a scalable, audit-compliant, and clinician-validated blueprint for CDSS, enabling low-latency triage and extensibility across specialties.https://www.mdpi.com/2076-3417/15/15/8412clinical decision supporttriagefair data principlesgenerative AIlarge language modelsfederated learning
spellingShingle Alper Karamanlıoğlu
Berkan Demirel
Onur Tural
Osman Tufan Doğan
Ferda Nur Alpaslan
Privacy-Preserving Clinical Decision Support for Emergency Triage Using LLMs: System Architecture and Real-World Evaluation
Applied Sciences
clinical decision support
triage
fair data principles
generative AI
large language models
federated learning
title Privacy-Preserving Clinical Decision Support for Emergency Triage Using LLMs: System Architecture and Real-World Evaluation
title_full Privacy-Preserving Clinical Decision Support for Emergency Triage Using LLMs: System Architecture and Real-World Evaluation
title_fullStr Privacy-Preserving Clinical Decision Support for Emergency Triage Using LLMs: System Architecture and Real-World Evaluation
title_full_unstemmed Privacy-Preserving Clinical Decision Support for Emergency Triage Using LLMs: System Architecture and Real-World Evaluation
title_short Privacy-Preserving Clinical Decision Support for Emergency Triage Using LLMs: System Architecture and Real-World Evaluation
title_sort privacy preserving clinical decision support for emergency triage using llms system architecture and real world evaluation
topic clinical decision support
triage
fair data principles
generative AI
large language models
federated learning
url https://www.mdpi.com/2076-3417/15/15/8412
work_keys_str_mv AT alperkaramanlıoglu privacypreservingclinicaldecisionsupportforemergencytriageusingllmssystemarchitectureandrealworldevaluation
AT berkandemirel privacypreservingclinicaldecisionsupportforemergencytriageusingllmssystemarchitectureandrealworldevaluation
AT onurtural privacypreservingclinicaldecisionsupportforemergencytriageusingllmssystemarchitectureandrealworldevaluation
AT osmantufandogan privacypreservingclinicaldecisionsupportforemergencytriageusingllmssystemarchitectureandrealworldevaluation
AT ferdanuralpaslan privacypreservingclinicaldecisionsupportforemergencytriageusingllmssystemarchitectureandrealworldevaluation