Revolutionizing Diabetic Retinopathy Screening: Integrating AI-Based Retinal Imaging in Primary Care

Area of Science:

• Ophthalmology
• Public Health
• Medical Informatics

Background:

• Diabetic retinopathy (DR) is a significant public health concern in the US and Europe.
• Suboptimal screening rates (50-70%) persist despite professional recommendations, hindering early detection.
• Barriers include lack of awareness, socioeconomic factors, fragmented healthcare, and workforce shortages.

Purpose of the Study:

• To evaluate the implementation of an AI-based retinal screening program integrated with quality improvement and continuing medical education.
• To assess the effectiveness of AI tools in improving DR detection within primary care settings.

Main Methods:

• Deployment of 198 AI-equipped cameras across 5 health systems starting in 2020.
• Screening of approximately 151,000 patients with diabetes, completing over 20,000 screenings.
• Integration with a comprehensive education and process improvement initiative.

Main Results:

• Over 3,450 individuals detected with more than mild DR, leading to specialist referrals.
• AI screening facilitated timely follow-up care for a significant number of patients.
• Demonstrated feasibility and utility of AI in primary care for DR detection.

Conclusions:

• AI-based DR screening offers a promising solution to improve detection rates and patient care in primary care.
• Integration of AI with educational initiatives can overcome screening barriers and optimize resource allocation.
• This approach warrants consideration for enhancing patient outcomes in diabetes management.