Utility of Large Language Models to Produce a Patient-Friendly Summary From Oncology Consultations

Area of Science:

• Medical Informatics
• Natural Language Processing

Background:

• Oncologic clinical notes contain complex medical terminology.
• Effective communication of cancer information is crucial for patient understanding and treatment adherence.

Purpose of the Study:

• To evaluate the utility of large language models (LLMs) in generating plain-language summaries of oncologic notes.
• To assess the potential of LLMs to improve communication and comprehension in oncology.

Main Methods:

• Utilized advanced large language models (LLMs) to process and rewrite de-identified oncologic clinical notes.
• Compared LLM-generated summaries against original notes for accuracy and clarity.

Main Results:

• LLMs successfully transformed technical oncologic note content into accessible, plain-language summaries.
• The generated summaries demonstrated potential for enhanced patient and clinician understanding.

Conclusions:

• Large language models offer a promising tool for simplifying complex medical information in oncology.
• This approach may significantly improve health literacy and shared decision-making in cancer care.