Automatic Summarization of Doctor-Patient Encounter Dialogues Using Large Language Model through Prompt Tuning

TL;DR

This paper demonstrates that prompt tuning of large clinical language models like GatorTronGPT can effectively summarize doctor-patient dialogues, offering a low-cost and efficient solution for clinical text summarization.

Contribution

It introduces prompt-tuning strategies for clinical LLMs and compares their performance to traditional fine-tuning methods, highlighting the effectiveness of prompt tuning in clinical summarization.

Findings

GatorTronGPT-20B achieved top performance across metrics.

Prompt tuning requires less computational resources than fine-tuning.

Generative clinical LLMs are effective for automatic clinical dialogue summarization.

Abstract

Automatic text summarization (ATS) is an emerging technology to assist clinicians in providing continuous and coordinated care. This study presents an approach to summarize doctor-patient dialogues using generative large language models (LLMs). We developed prompt-tuning algorithms to instruct generative LLMs to summarize clinical text. We examined the prompt-tuning strategies, the size of soft prompts, and the few-short learning ability of GatorTronGPT, a generative clinical LLM developed using 277 billion clinical and general English words with up to 20 billion parameters. We compared GatorTronGPT with a previous solution based on fine-tuning of a widely used T5 model, using a clinical benchmark dataset MTS-DIALOG. The experimental results show that the GatorTronGPT- 20B model achieved the best performance on all evaluation metrics. The proposed solution has a low computing cost as the LLM parameters are not updated during prompt-tuning. This study demonstrates the efficiency of generative clinical LLMs for clinical ATS through prompt tuning.