Integrating artificial intelligence in ophthalmology: a pilot study of clinical understanding and adoption

Abstract

Purpose: The purpose of this pilot study was to explore the current understanding and application of artificial intelligence (AI) within clinical ophthalmology. Design: This study used a qualitative research approach. One-on-one interviews were conducted with ophthalmologists (including residents/fellows/students) and medical professionals involved in ophthalmology. Methods: Participants were recruited via professional networks, and an interview guide informed by prior research and expertise of the interdisciplinary research team led the question-asking process. Transcribed interviews were analyzed using qualitative thematic analysis methods with Nvivo12 software. Results: Participants (N = 18) included attending clinicians (44%, n = 8), residents (44%, n = 8), a fellow (6%, n = 1), and a medical student (6%, n = 1). In-depth analysis of the interviews yielded 3 overarching themes: 1) AI has high utility in ophthalmology; 2) AI is a tool, but a balance between AI and the clinician is important; and 3) several challenges to integrating and accessing AI need to be addressed. Overall, participants believed an AI informed clinical practice is important and participants described ways AI could be incorporated into their own patient management. However, the majority of participants do not presently use AI in patient care, noting concerns about the current state of AI in research and clinical practice. Participants also described balance between AI and the provider as essential, suggesting AI applications are not currently able to replace the human element of clinical practice. AI applications in ophthalmic clinical practice are viewed positively across all participants, with noted caution towards the current ability to use AI as an automated tool and challenges for its integration into clinical management. Conclusions and future perspectives: Although findings yielded generally favorable views, suggesting high potential for benefit with integration of AI systems, several barriers to adoption were noted by participants. While participants believe AI is the future of ophthalmology, a balance between the clinician and the computer is vital and concerns related to trustworthiness of the data were a consistent finding. This research lays important groundwork for developing future research that can bridge the gap between the development of AI systems and its translation to more effective clinical practice.

Ähnliche Arbeiten