UPFRONT: Combating Retinal Disease With Data


“It is a capital mistake to theorize before one has data. Insensibly, one begins to twist the facts to suit theories, instead of theories to suit facts.” – Sherlock Holmes

In 2011, the American Academy of Ophthalmology started a new registry called the Intelligent Research in Sight Registry, or IRIS Registry, which went live on March 25, 2014. Within a year, 1 in 4 AAO members used the website and it has quickly become the largest single-specialty clinical database in the world. At first, it required manual entry of data through a web portal. It now interfaces with many common electronic medical record (EMR) systems, including Compulink, EyeMD EMR, NextGen, Modernizing Medicine, and Nextech, to automatically extract electronic data to the registry without any physician or staff involvement.

One key reason to enter your patient data into the registry is to avoid penalties from federal quality reporting programs, as it is a federally designated Qualified Clinical Data Registry. It is estimated that the IRIS Registry has saved ophthalmologists an estimated $90 million in 2014 and 2015, or roughly $19,000 per year per physician; not bad compared to the estimated $15 million that the AAO spent developing the system.

While some may enter data into the IRIS Registry to earn bonus points in the Merit-Based Incentive Payment System (MIPS) without having to pay your EMR company or for maintenance of certification requirements, there is a far greater power to the registry. With more than 200 million patient visits, one can learn a tremendous amount by mining the data for patterns. For example, we can investigate whether anti-VEGF injections lead to a higher incidence of glaucoma, whether cataract surgery lowers IOP, or what endophthalmitis rates are after injections or surgery.

With the advent of artificial intelligence and machine learning algorithms that will better be able to analyze the data, this treasure trove of data will help all ophthalmologists to answer questions only big data can. In 2017, the AAO announced a partnership with Verana Health (formerly DigiSight) to use machine learning to explore the registry. For readers who have a specific project that they would like to use the registry to explore, the AAO has set up a system to vet project ideas. Researchers can also benchmark their own data against their peers’ to evaluate their own outcomes. This may lead to improved practice patterns and outcomes in the future.

In this issue, we explore the IRIS Registry and some of the retina findings that have been gleaned from the data to date. We also tackle the recent outbreak of sterile inflammation seen after aflibercept injections in a peer-reviewed article. It goes without saying that these rare events could quickly be discovered using real-time data mining by artificial intelligence of the IRIS Registry. I look forward to deep dives into this data to improve our practice in the future. RP

Listen to episodes of Straight From the Cutter’s Mouth with discussion of Retinal Physician articles.