Assessing Patient Education Guide Generated by ChatGPT vs Google Gemini on Common Hepatology Conditions: A...

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Hepatology Patient Education

Background:

• Artificial intelligence (AI) is increasingly used in healthcare, necessitating quality assessment of AI-generated health information.
• Patient education guides are crucial for managing liver conditions like hepatitis, cirrhosis, and non-alcoholic fatty liver disease.

Purpose of the Study:

• To compare patient education guides on common hepatology conditions generated by ChatGPT and Google Gemini.
• To evaluate the accessibility, clarity, and comprehensiveness of AI-generated health content.

Main Methods:

• Comparative analysis of AI-generated patient education guides.
• Utilized Flesch-Kincaid readability tests, Quillbot for similarity scores, and DISCERN for reliability.
• Quantitative analysis of word count, sentence count, and readability metrics.

Main Results:

• ChatGPT produced more extensive content (words and sentences) than Google Gemini.
• No significant differences were found in readability, complexity (words per sentence, syllable count, grade level, ease score), or reliability between the two AI models.
• Similarity scores were also comparable, indicating consistent output.

Conclusions:

• AI-generated health information requires refinement to meet diverse patient needs.
• While AI shows potential for patient education, continuous evaluation is vital for clarity and balance.
• Recommendations focus on enhancing content accessibility and reliability for better patient engagement.