Led by Dr. Iris M. Kim from the Department of Ophthalmology, University of California, San Francisco, the prognostic study aimed to assess whether clinical data available through EHRs could be harnessed to anticipate the onset of PDR—a severe, vision-threatening complication of diabetic retinopathy. Researchers analyzed health records of 7,739 adults with type 2 diabetes who had been diagnosed with NPDR or diabetic macular edema (DME).
The study utilized de-identified data from the University of California Health Data Warehouse, which houses information on approximately 10 million patients. Participants were included if they were 18 years or older and had no previous diagnosis of PDR before the study index date. Patients were split into development and external test cohorts, with the development set further divided into training and internal test groups.
Three predictive models—Cox proportional hazards regression, Cox with least absolute shrinkage and selection operator (LASSO), and random survival forest (RSF)—were trained and tested on these datasets. The models evaluated a broad range of factors, including demographic data, comorbid conditions, laboratory values, medications, and ophthalmic diagnoses.
The findings of the study are summarized as follows:
- During a mean follow-up period of approximately two years, 9.3% of participants progressed to proliferative diabetic retinopathy (PDR).
- The average time to progression was 1.89 years.
- All three survival models demonstrated strong predictive performance, with concordance index (C-index) values between 0.73 and 0.75 in both internal and external test sets.
- The Cox proportional hazards regression and random survival forest (RSF) models showed good calibration for up to two years.
- Key predictors of progression to PDR identified by the models included older age, racial and ethnic background, presence of diabetic macular edema (DME), greater severity of nonproliferative diabetic retinopathy (NPDR) at baseline, higher mean hemoglobin A1c levels, insulin use, and diabetic nephropathy.
- The findings highlight the combined impact of systemic and eye-related factors in determining the risk of PDR progression.
According to the authors, these predictive tools offer an opportunity for earlier detection and personalized care planning. By anticipating which patients are at highest risk, clinicians can tailor monitoring intervals and allocate healthcare resources more efficiently. The study highlights the growing role of EHR-integrated machine learning in improving chronic disease management.
The authors concluded, "As EHR systems continue to evolve, the integration of predictive modeling into routine care could significantly enhance outcomes for patients with diabetes at risk of vision loss."
Reference:
Kim IM, Radgoudarzi N, Chen EM, et al. Time to Progression to Proliferative Diabetic Retinopathy in Patients With Type 2 Diabetes. JAMA Netw Open. 2025;8(7):e2521150. doi:10.1001/jamanetworkopen.2025.21150
Disclaimer: This website is primarily for healthcare professionals. The content here does not replace medical advice and should not be used as medical, diagnostic, endorsement, treatment, or prescription advice. Medical science evolves rapidly, and we strive to keep our information current. If you find any discrepancies, please contact us at corrections@medicaldialogues.in. Read our Correction Policy here. Nothing here should be used as a substitute for medical advice, diagnosis, or treatment. We do not endorse any healthcare advice that contradicts a physician's guidance. Use of this site is subject to our Terms of Use, Privacy Policy, and Advertisement Policy. For more details, read our Full Disclaimer here.
NOTE: Join us in combating medical misinformation. If you encounter a questionable health, medical, or medical education claim, email us at factcheck@medicaldialogues.in for evaluation.