Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Lightweight Clinical Decision Support System using QLoRA-Fine-Tuned LLMs and Retrieval-Augmented Generation

Mohammad Shoaib Ansari, Mohd Sohail Ali Khan, Shubham Revankar, Aditya Varma, Anil S. Mokhade

TL;DR

The paper addresses the need for accurate, context-aware clinical decision support using resource-efficient AI. It presents a two-component framework that fuses Retrieval-Augmented Generation with hospital-specific data and Quantized Low-Rank Adaptation to fine-tune a compact Llama model. The authors demonstrate benefits in disease prediction, treatment suggestion, and medical report summarization, supported by evaluations on medical QA benchmarks and practical workflows. They also discuss data privacy, clinical liability, and the challenges of real-world deployment, highlighting pathways for scalable adoption in low-resource healthcare settings. Overall, the work shows that lightweight, ground-grounded LLMs can deliver practical, governance-conscious AI assistance in healthcare with meaningful impact on clinician efficiency and patient care.

Abstract

This research paper investigates the application of Large Language Models (LLMs) in healthcare, specifically focusing on enhancing medical decision support through Retrieval-Augmented Generation (RAG) integrated with hospital-specific data and fine-tuning using Quantized Low-Rank Adaptation (QLoRA). The system utilizes Llama 3.2-3B-Instruct as its foundation model. By embedding and retrieving context-relevant healthcare information, the system significantly improves response accuracy. QLoRA facilitates notable parameter efficiency and memory optimization, preserving the integrity of medical information through specialized quantization techniques. Our research also shows that our model performs relatively well on various medical benchmarks, indicating that it can be used to make basic medical suggestions. This paper details the system's technical components, including its architecture, quantization methods, and key healthcare applications such as enhanced disease prediction from patient symptoms and medical history, treatment suggestions, and efficient summarization of complex medical reports. We touch on the ethical considerations-patient privacy, data security, and the need for rigorous clinical validation-as well as the practical challenges of integrating such systems into real-world healthcare workflows. Furthermore, the lightweight quantized weights ensure scalability and ease of deployment even in low-resource hospital environments. Finally, the paper concludes with an analysis of the broader impact of LLMs on healthcare and outlines future directions for LLMs in medical settings.

Lightweight Clinical Decision Support System using QLoRA-Fine-Tuned LLMs and Retrieval-Augmented Generation

TL;DR

The paper addresses the need for accurate, context-aware clinical decision support using resource-efficient AI. It presents a two-component framework that fuses Retrieval-Augmented Generation with hospital-specific data and Quantized Low-Rank Adaptation to fine-tune a compact Llama model. The authors demonstrate benefits in disease prediction, treatment suggestion, and medical report summarization, supported by evaluations on medical QA benchmarks and practical workflows. They also discuss data privacy, clinical liability, and the challenges of real-world deployment, highlighting pathways for scalable adoption in low-resource healthcare settings. Overall, the work shows that lightweight, ground-grounded LLMs can deliver practical, governance-conscious AI assistance in healthcare with meaningful impact on clinician efficiency and patient care.

Abstract

This research paper investigates the application of Large Language Models (LLMs) in healthcare, specifically focusing on enhancing medical decision support through Retrieval-Augmented Generation (RAG) integrated with hospital-specific data and fine-tuning using Quantized Low-Rank Adaptation (QLoRA). The system utilizes Llama 3.2-3B-Instruct as its foundation model. By embedding and retrieving context-relevant healthcare information, the system significantly improves response accuracy. QLoRA facilitates notable parameter efficiency and memory optimization, preserving the integrity of medical information through specialized quantization techniques. Our research also shows that our model performs relatively well on various medical benchmarks, indicating that it can be used to make basic medical suggestions. This paper details the system's technical components, including its architecture, quantization methods, and key healthcare applications such as enhanced disease prediction from patient symptoms and medical history, treatment suggestions, and efficient summarization of complex medical reports. We touch on the ethical considerations-patient privacy, data security, and the need for rigorous clinical validation-as well as the practical challenges of integrating such systems into real-world healthcare workflows. Furthermore, the lightweight quantized weights ensure scalability and ease of deployment even in low-resource hospital environments. Finally, the paper concludes with an analysis of the broader impact of LLMs on healthcare and outlines future directions for LLMs in medical settings.
Paper Structure (33 sections, 2 equations, 3 figures, 3 tables)

This paper contains 33 sections, 2 equations, 3 figures, 3 tables.

Table of Contents

  1. Introduction
  2. System Architecture
  3. Hospital Data Preprocessing and Embedding
  4. Runtime Query Processing
  5. Fine-Tuning with QLoRA
  6. Quantized Low-Rank Adaptation: Principles and Implementation
  7. Implementation with Llama 3.2-3B-Instruct
  8. Dataset Selection
  9. LoRA Adapter Configuration
  10. Training Details
  11. Model Evaluation
  12. Benefits of QLoRA Fine-Tuning over Traditional Approaches
  13. RAG on Hospital Data
  14. Medical Data Preprocessing and Embedding
  15. Hybrid Retrieval System
  16. ...and 18 more sections

Figures (3)

  • Figure 1: System Architecture Diagram
  • Figure 2: Loss Value and Learning Rate during Finetuning
  • Figure 3: Data Preprocessing Pipeline for RAG