Emerging challenges in the application of artificial intelligence for the eye disease screening in Chinese primary healthcare institutions

Abstract

Breakthroughs have been achieved recently in the application of artificial intelligence (AI) for the eye disease screening in Chinese primary healthcare institutions, but challenges have also emerged. First, AI software has continuously evolved, expanding the range of eye diseases that can be screened, enhancing diagnostic accuracy, and progressing towards predicting the course of eye diseases. However, inadequate infrastructure such as 5G and a shortage of specialized personnel have hindered the coverage of screenings. Second, while the cost-effectiveness of AI is well-established, new screening models have impacted the equity of screenings. It is essential to tailor AI application models to the specific context of China. Third, AI screening guidelines have been increasingly improved, providing direction for AI development and reference for the promotion and application of AI technologies. Nonetheless, high-quality empirical research is urgently needed to provide scientific evidence for policymaking related to AI in the eye disease screening. Therefore, it is suggested to develop multimodal AI models that integrate basic data such as symptoms and medical history with simple ophthalmic examinations, to accelerate the construction of infrastructure like 5G and focus on cultivating interdisciplinary talents, to explore suitable service systems and models for the large-scale eye disease screening tailored to local conditions, and to conduct long-term, multi-center, empirical studies.