Comprehensive Summary
Tappeiner conducted a survey to assess the opinions of ophthalmologists and patients in Switzerland on the clinical use of AI in ophthalmology. An anonymous survey was conducted through the online platform findmind.ch and distributed either through WhatsApp or by directly inviting ophthalmologists to participate. The survey was built based on literature review and contained questions regarding understanding of AI, current concerns of clinical use of AI, and possible applications. Over the course of 3 weeks, 106 valid responses were compiled, consisting of 24.5% Switzerland residents and 75.5% ophthalmologists. Many participants indicated that they lacked sufficient understanding to fully commit to using AI in clinical practice and many also supported the integration of AI education. Participants also indicated the areas where they believed AI would be most useful, with image analysis and pre-screening being the highest, and intraoperative assistance not being selected at all. Finally, commonly noted concerns about clinical AI use were liability and data protection issues, reliability, and lack of knowledge or training, and many participants also preferred for AI to be used in a subordinate role in the clinic. These are the primary factors that suggest that further structured AI education is necessary to increase confidence in AI use in ophthalmology clinics.
Outcomes and Implications
This survey provides an accurate, current representation of perspectives on clinical AI use among both ophthalmologists and residents in Switzerland. Tapperiner proposed to integrate AI into clinics by gradually introducing AI to select clinics with a focus on image analysis, along with establishing institutional training and guidelines in order to make the process as transparent to ophthalmologists and patients as possible. Moving forward, AI in ophthalmology could provide benefits to doctors and patients alike, but will require careful systematic implementation to eliminate security and accuracy concerns.