Century Health and Eye Health America to develop real-world retinal disease dataset

A new collaboration between Century Health and Eye Health America will structure more than a decade of ophthalmology data spanning over 200,000 de-identified patient records for research into chronic retinal diseases.

 The Baseline

• Century Health and Eye Health America are cooperating to develop longitudinal datasets for research into chronic retinal diseases.
• The work will use AI to structure clinical information from more than 200,000 de-identified ophthalmology records spanning over a decade.
• The resulting datasets will support real-world evidence research into disease progression and treatment response.

Century Health and Eye Health America (EHA) have announced a new alliance to develop research datasets from longitudinal clinical data on chronic retinal conditions, including diabetic retinopathy and age-related macular degeneration (AMD).

The project will focus particularly on clinical information that is currently difficult to use for large-scale research as it is captured in unstructured sources such as clinical notes, PDF attachments and scanned reports rather than structured electronic health record (EHR) fields. This can include quantitative retinal measurements and imaging findings used to monitor disease progression and treatment response over time.

Century Health will apply its Century Health Abstraction and Retrieval Model (CHARM) to EHA records, including reports dating back more than 10 years. The AI model will extract key retinal measurements and disease progression metrics from unstructured clinical information and convert them into structured datasets for research.

Vish Srivastava, Co-Founder and CEO of Century Health, commented on the availability of retinal disease data for research:

“The clinical data that matters most in retinal disease, like progression measurements, imaging findings, treatment response over years, has remained locked inside ophthalmology practices.”

The collaboration reflects the importance of longitudinal data in chronic retinal diseases, where changes can occur gradually and patients may be monitored and treated over many years. An estimated 9.6 million people with diabetes in the US have diabetic retinopathy, including nearly 1.84 million with a vision-threatening form of the disease, while AMD affects an estimated 19.8 million people. Treatment response can vary between patients, and changes in retinal structure over time can provide information on disease progression and response to treatment.

Drawing on more than 200,000 de-identified records, the resulting dataset is intended to support real-world evidence (RWE) generation and life sciences research examining disease progression, treatment response, and other variables in routine clinical care.

Dr Cathleen McCabe, Chief Medical Officer at Eye Health America, a leading integrated eye care platform in the Southeast US, emphasized the role of longitudinal information in monitoring chronic eye conditions:

"Caring for patients with chronic eye conditions means following subtle changes over many years, yet much of that longitudinal history isn't available for large-scale research. By transforming more clinical history into usable data, we can better understand how these diseases progress, identify which treatments work best for different patients, and ultimately improve the care we deliver."