Securing Safety and Quality in AI-generated Patient Education: A Nurse-led Methodological Framework Integrating...

Area of Science:

• Digital Health
• Artificial Intelligence in Healthcare
• Nursing Informatics

Background:

• Generative AI in healthcare offers personalized patient education.
• Clinical inaccuracies and lack of theoretical grounding in AI-generated content pose safety risks.
• Existing AI tools lack a robust framework for ensuring clinical safety and theoretical alignment.

Purpose of the Study:

• To validate a nurse-led AI protocol for generating safe, theory-guided digital patient education materials.
• To ensure AI-generated content is grounded in Kolcaba's Comfort Theory.
• To establish a quality-control framework for AI in healthcare education.

Main Methods:

• A three-stage iterative prompt engineering process: initial, clinical refinement, and theoretical alignment.
• Utilized stroke, chronic kidney disease, and COPD as case models for developing educational materials.
• Assessed quality using expert panel reviews (n=3) with Content Validity Index (CVI) and Ateşman's Readability Index.

Main Results:

• The nurse-led AI protocol effectively mitigated AI hallucinations and ensured theoretical integration across comfort domains.
• Readability scores significantly improved from 48.5 to 66.8.
• High expert consensus (CVI: 0.93-0.95) demonstrated clinical safety and effectiveness.

Conclusions:

• Nursing expertise is a mandatory safety layer in AI-generated healthcare education.
• The validated protocol provides a replicable quality-control mechanism for nurses.
• This framework ensures AI-generated digital materials are clinically safe and theoretically sound.