Improving Informed Consent Models for Endobronchial Ultrasound With Artificial Intelligence

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Patient Communication

Background:

• Informed consent (IC) is crucial for patient understanding of medical procedures, including endobronchial ultrasound (EBUS).
• Artificial intelligence (AI) offers potential solutions to enhance the traditional IC process.
• This study investigates AI's role in improving IC documents and evaluating AI-generated video consents.

Purpose of the Study:

• To assess the efficacy of AI-generated informed consent (IC) documents compared to traditional ones.
• To explore the feasibility and patient acceptance of AI-generated video consents (AIV-IC) as an alternative to verbal IC.
• To determine if AI can improve clarity, explanation of benefits, and discussion of complications in IC.

Main Methods:

• Development of an AI-generated IC (AI-IC) using a generative AI model.
• Phase I: Participants compared AI-IC and traditional IC (H-IC) texts via Likert scale questionnaires.
• Phase II: Patients evaluated AI-IC (text) and AIV-IC (video) formats through questionnaires.

Main Results:

• AI-IC texts scored higher in language clarity, benefits explanation, and addressing complications than H-IC.
• A significant majority preferred AI-IC for its mention of alternative procedures.
• Patients reported high satisfaction with both AI-IC and AIV-IC, with AIV-IC being accepted as a replacement for verbal IC.

Conclusions:

• AI-generated materials enhance the accessibility of IC for EBUS procedures.
• AI-assisted and video-based consent tools show potential for clinical integration.
• Human supervision remains essential for AI-driven consent processes.