Automatic Summarization of Doctor-Patient Encounter Dialogues Using Large Language Model through Prompt Tuning
arXiv:2403.13089v2 Announce Type: replace 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
The proliferation of advanced large language models (LLMs) and the increasing data burden on healthcare providers are creating a critical need for efficient summarization tools.
This development can significantly improve workflow efficiency for clinicians, reduce administrative overhead, and enable more focused patient care by distilling complex dialogues into actionable summaries.
The ability of LLMs to generate high-quality summaries of doctor-patient encounters through prompt tuning shifts how unstructured clinical data can be processed and utilized in healthcare.
- · Healthcare providers
- · Patients
- · AI/LLM developers
- · Clinical software integrators
- · Medical transcription services (traditional)
- · Healthcare administrative staff (repetitive tasks)
Clinicians gain immediate access to concise summaries of patient interactions, enhancing decision-making and handover processes.
Reduced administrative workload could free up healthcare professionals to spend more time on direct patient care, improving overall care quality and patient satisfaction.
Sophisticated summarization capabilities could lead to new forms of proactive and personalized medicine, as AI can quickly surface critical insights from vast amounts of clinical dialogue data.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.CL