Study shows non-random association between missing electronic health record data regarding the severity of diabetic retinopathy

The study demonstrates how de-identified real-world data (RWD) from electronic health records (EHRs) in medical registries can allow for a deeper understanding of data trends, patterns, or gaps. (Image Credit: Adobe Stock/Andrey Popov)

Verana Health® shared that its recent ophthalmic research has been published in the peer-reviewed Journal of the American Medical Informatics Association (JAMIA). The study demonstrates how de-identified real-world data (RWD) from electronic health records (EHRs) in medical registries can allow for a deeper understanding of data trends, patterns, or gaps by helping to fill in the missing pieces and providing deeper context.

Conducted by a team of Verana Health researchers, the study shows a non-random association between missing EHR data regarding the severity of diabetic retinopathy (DR) in de-identified patients and specific demographic, clinical, and practice-level characteristics. The study, titled: "Identifying the Mechanism of Missingness for Unspecified Diabetic Retinopathy Disease Severity in the Electronic Health Record: An IRISⓇ Registry Analysis," leveraged de-identified, curated RWD from the American Academy of Ophthalmology (Academy) IRIS® Registry (Intelligent Research in Sight).

The goal of this study was to understand the instances in which disease severity is not specified using diagnosis codes alone and whether patterns existed for the missing data. Investigating how codes capture disease severity provides further insight into creating cohorts for future studies.

The research team identified more than 10 million clinical encounters with ICD-9 codes for DR and 24 million clinical encounters with ICD-10 codes for DR in the IRIS Registry occurring from 2014 to 2021 resulting in 3.6 million unique de-identified patients.

Among the major findings of the study:

• Although disease severity for DR is routinely captured in EHR notes, nearly 20% of ICD-9 cases had unspecified DR severity. This significantly improved with the transition to ICD-10 codes with only 5% of clinical encounters with unspecified DR severity.
• Encounters with unspecified disease severity were associated with significantly older patient age, better visual acuity (VA), and lower utilization of ophthalmic procedures.
• Higher-volume practices and retina specialist practices had lower proportions of clinical encounters with unspecified disease severity.
• A higher proportion of clinical encounters with specified disease severity is associated with anti-VEGF injections and panretinal photocoagulation compared to encounters with unspecified disease severity.
• Visual acuity (VA) for encounters with specified disease severity was significantly lower than the mean VA for encounters with unspecified disease severity. This finding supports the assessment that clinical encounters with unspecified disease severity are associated with less severe DR severity.

Toward earlier intervention

The volume and quality of RWD from the IRIS Registry allowed the Verana Health team to make meaningful comparisons of unspecified DR severity across multiple structured data elements including demographic, procedure, ocular comorbidity data, and practice-level characteristics. These study results are generalizable to studies using various EHR registries in ophthalmology because more than 70% of practicing ophthalmologists in the U.S. contribute data to the IRIS Registry.

"Diabetic retinopathy treatment is headed toward earlier intervention, so quality real-world data is invaluable," said Durga S. Borkar, M.D., MMCi, a Verana Health medical advisor, a vitreoretinal surgeon at Duke University Eye Center, and senior author of the study. "Recognizing the associations between DR severity and other clinical and demographic characteristics through RWD will help enable investigators to build more accurate predictive models."

Established in 2014, the IRIS Registry is the first comprehensive eye disease clinical registry in the U.S. It includes more than nine years of longitudinal patient RWD amassed from more than 490 million billable patient encounters, 79 million unique patients, and more than 15,000 ophthalmologists and other eye care providers across the country. Verana Health is the Academy's exclusive end-to-end data curation and analytics partner for the IRIS Registry.

"This study demonstrates the deep value of robust and representative medical registries such as the IRIS Registry," said Verana Health CEO Sujay Jadhav. "The curated real-world data from clinicians provides a longitudinal and detailed view of the patient journey that can't be found elsewhere, and which can unlock new insights for research and development of therapies to help improve patient outcomes and quality of life."