Title presentation
Closing the Diabetic Retinal Exam Care Gap: Integration of Autonomous AI Technology Increases Access
5 key short sentences highlighting the essence of my presentation
- Compliance for regular diabetic retinopathy screening is low in the U.S. and around the world.
- Deep learning based image classification will help alleviate the bottle necks of current models of diabetic retinopathy screening.
- Deployment of AI-assisted diabetic retinopathy screening within an integrated healthcare system requires close collaboration between primary care and ophthalmology.
- Deployment of AI-assisted diabetic retinopathy screening within an integrated healthcare system improves diabetic retinopathy screening compliance rate.
- Efficient and successful implementation of such a program will be beneficial to both primary care clinics and specialized eye hospitals.
Summary of my presentation
We studied whether the implementation of autonomous AI for the diabetic retinopathy (DR) exam in a
community setting improved year-over-year DR screening compliance rate. The data was captured in
August 2020 (prior to implementation) and in August 2021 (1 year after implementation). Community
based clinics within the Baltimore County and Baltimore City, managed by Johns Hopkins Medicine,
were compared. Two clinics with autonomous AI DR screening (“AI clinics”) were compared to four
clinics without autonomous AI screening (“non-AI clinics”).
The DR exam adherence rate in the AI clinics improved from 35.4% (1,127/3,182) to 58.8% (1,995/3,395),
representing a 23.4% increase. Of the 1,995 exams, 300 (15.0%) were performed autonomously. The compliance rate in the non-AI clinics improved from 30.0% (751/2,504) to 41.7% (1,080/2,593), representing an 11.7% increase. The increase of 23.4% in AI clinics was statistically significant as compared to the increase of 11.7% increase in the non-AI clinics (p<0.0002).`
We conclude that the implementation of autonomous AI results in a statistically significant improvement in HEDIS compliance, made even more difficult during the pandemic. In addition, we performed in-depth analyses of the patients who were referred to the Wilmer Eye Institute, as a result of the autonomous screening program.