What question did this study set out to answer?

This research aims to create an effective medical symptom triage system for Uzbek speakers using LLM technology.

June 23, 20260 citationsOpen Access

LLM-Enhanced Medical Symptom Triage System for Low-Resource Languages: A Case Study in Uzbek

BRBahrom Reyimberganov BTBurxon Turayev SNSirojiddin Nuraddinov

Key Points

This research aims to create an effective medical symptom triage system for Uzbek speakers using LLM technology.
Developed CareRoute AI with a hypothesis-driven adaptive questioning engine.
Utilized DeepSeek LLM for context-aware follow-up and urgency classification.
Evaluated on 150 synthetic clinical scenarios.
Achieved 89.3% agreement with physician triage decisions.
Demonstrated 100% sensitivity for emergency-level cases.

Abstract

This paper presents CareRoute AI, an LLM-enhanced medical symptom triage system designed for the Uzbek language, addressing the critical gap in digital health tools for low-resource language speakers. The system implements a hypothesis-driven adaptive questioning engine combined with DeepSeek LLM integration for context-aware follow-up generation and urgency classification. The architecture follows a safety-first design pattern with red-flag symptom detection triggering immediate emergency escalation. Evaluation on 150 synthetic clinical scenarios demonstrated 89.3% agreement with physician triage decisions, with 100% sensitivity for emergency-level cases.

问 AI

Bookmark

View Full Paper