Transformation of Home Physical Therapy Guidance with Generative AI.

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare

Background:

• Patient guidance software requires trustworthy medical knowledge, user-friendliness, and adaptability.
• Traditional decision support tools for home physical therapy have limitations in enhancing patient autonomy.

Purpose of the Study:

• To explore the integration of generative artificial intelligence (AI) into patient guidance software for home-based physical therapy.
• To evaluate the impact of generative AI on the quality and readability of rehabilitation guidance.
• To assess the potential of AI to increase patient autonomy in rehabilitation.

Main Methods:

• Integrated a GPT-4.1-mini model API into an existing software application.
• Conducted a comparative evaluation of guidance provided with and without generative AI input.
• Utilized the Flesch-Kincaid index to assess guideline complexity and readability.

Main Results:

• Generative AI significantly improved the readability and reduced the complexity of patient guidance.
• AI-powered prompts enhanced the clarity and understandability of physical therapy instructions.
• The Flesch-Kincaid index indicated a more accessible user experience with AI integration.

Conclusions:

• Generative AI frameworks represent a paradigm shift in patient guidance for physical therapy.
• AI integration can effectively increase patient autonomy and improve the quality of home-based rehabilitation.
• Optimized prompting strategies are key to maximizing the benefits of AI in medical guidance.