To assess patient comfort and acceptance of AI in making retreatment decisions for macular disease based on retinal imaging.
Approach:
Key Findings:
Error rate (34.4%) and presence of a second reader/checker (33.6%) were the most important factors for patient confidence in AI.
Participants showed no significant preference for human versus AI as the first reader.
Many participants suggested that a human making final decisions with AI support could be ideal.
Interpretation:
Patients prioritize reliability and verification in AI-led treatment decisions over the identity of the decision-maker, indicating a potential pathway for successful AI integration in clinical practice.
Limitations:
The study had a limited sample size with only 181 complete responses.
Responses may not fully represent the broader patient population due to recruitment method.
Conclusion:
To facilitate AI adoption in macular clinics, focus should be on ensuring high performance, accuracy, and transparency rather than on promoting AI technology itself.
