Page 74 - AIH-1-2

Artificial Intelligence in Health Perspective on AI in eye care practices

Figure 1. PRISMA diagram illustrating the process of article selection. Adapted from Page et al 67
Abbreviation: PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

prevalent eye conditions, such as diabetic retinopathy, be the most significant area for AI application (78.2%),
glaucoma, age-related macular degeneration, and cataracts. followed by the diagnoses of glaucoma (70.7%), age-related
The global survey involved 1176 ophthalmologists from macular degeneration (66.8%), and cataracts (51.4%).
70 countries, and the response rates were 78.8–85.8% 40
per question. According to the survey findings, 88.1% Ho et al. assessed the perspectives of optometrists on
of ophthalmologists expressed readiness to use AI the use of AI in the diagnosis of retinal disorders. A paper-
technology, particularly as clinical assisting tools. However, based survey was conducted among 133 optometrists
the preference for the use of AI as a tool for diagnosis and to determine the factors and obstacles affecting AI
assisting clinical decisions declined at a response rate of implementation in optometry, as well as their general
64.5% and 78.8%, respectively. Most of the respondents opinion toward AI technology. The primary results of the
expressed confidence that AI would not take their jobs survey revealed that the surveyed optometrists generally
(68.2%). Approximately 72.5% of respondents identified had an optimistic view toward using AI as a support tool
notable challenges in AI implementation, including to diagnose retinal disorders. The optometrists’ perception
concerns regarding medical liability resulting from errors. of AI-assisted diagnosis was positive, with a mean score of
The diagnosis of diabetic retinopathy was identified to 4.0 out of 5 (standard deviation [SD]: 0.8). Furthermore,

Volume 1 Issue 2 (2024) 68 doi: 10.36922/aih.2809