OncoChat: LLM-driven Patient Support and Healthcare Communication

Abstract

The deployment of large language models (LLMs) is significantly transforming healthcare communication, particularly in oncology use cases where precision and timeliness are critical. In this study, we introduce OncoChat, which refers to the finely-tuned LLM developed on the foundational large language model Meta AI (LLaMA) for oncology-specific dialogues. We integrate a distinctive feature of the sentence similarity model, which filters out irrelevant inquiries, ensuring that only oncology-related questions are processed. This functionality not only enhances the accuracy but also the reliability of the provided information. The effectiveness of the proposed framework is presented through comparative performance metrics with the benchmark models. The results show that the OncoChat achieves a precision of 82%, recall of 86% and an F1-score of 84%, outperforming OncoGPT and ChatDoctor. These results hihglight the OncoChat’s potential to assist healthcare professionals by providing accurate information, supporting patient understanding, and timely communication with the healthcare providers.

Ähnliche Arbeiten