Back to Journals » Clinical Ophthalmology » Volume 19
Artificial Intelligence Improves Patient Follow-Up in a Diabetic Retinopathy Screening Program [Letter]
Authors Bhavsar AP 
, Minerd CI, Piri N
Received 19 July 2025
Accepted for publication 5 August 2025
Published 9 August 2025 Volume 2025:19 Pages 2659—2660
DOI https://doi.org/10.2147/OPTH.S537960
Checked for plagiarism Yes
Editor who approved publication: Dr Scott Fraser
Avi P Bhavsar,1 Charles I Minerd,1 Niloofar Piri2
1Saint Louis University School of Medicine, Saint Louis, MO, USA; 2SSM Health Medical Group, Department of Ophthalmology, Saint Louis University School of Medicine, Saint Louis, MO, USA
Correspondence: Niloofar Piri, MD, SSM Health Medical Group, Department of Ophthalmology, Saint Louis University School of Medicine, Saint Louis, MO, USA, Email [email protected]
View the original paper by Dr Dow and colleagues
Dear editor
We commend Dow et al for demonstrating that an artificial-intelligence (AI) driven diabetic retinopathy (DR) screening workflow led to a threefold increase in follow-up rates compared to a human workflow.1 Utilizing IDx-DR, the first FDA-approved autonomous DR screening system, marks a promising advancement in access to DR screening and care.2,3 After all, DR is a leading cause of vision loss in the United States,4 and its prevalence remains an epidemiological concern.5,6 Despite the positive results illustrated by Doe et al, we wish to contribute some additional factors that merit consideration prior to implementation.
First, study participants were seen in both primary care and endocrinology settings. Patients screened at endocrinology clinics represent a population with poorly controlled or advanced diabetes,7 along with other systemic conditions. These patients may be referred to ophthalmology regardless of screening results, and we must consider whether AI screening affects referral behavior to begin with. After all, AI screening cannot detect several relevant ocular conditions. Namely, patients with diabetes are at moderate risk of developing open angle and neovascular glaucoma and retinal vascular occlusions.8,9 Accordingly, there is a benefit of over-referral. While the study did also screen primary care patients, DR screening of high-risk populations may cause other missed diagnoses.
Second, although this study was performed in an urban, ethnically diverse setting, rural populations may benefit significantly from point-of-care AI screening as they are less likely to meet the standard of diabetes care.10 Lacking proximity to specialty care, rural patients face barriers to attending follow-up appointments, receiving procedures, and seeking urgent care.11 While these nuances are not captured in this study, they should inform equitable and cost-effective deployment of screening initiatives.
Finally, the authors surmise that the positive findings may be due to more rapid delivery of results in the AI workflow.1 However, patients also cited knowledge gaps as to why they did not follow up.1 Therefore, real-time delivery of results could be bolstered with patient education. Future studies should consider a stepwise protocol for DR screening in rural primary care offices that reduces patient burden by combining novel screening tools. This may involve high sensitivity screening by IDx-DR2,3 prior to smartphone OCT (SCANLY),12 reviewed remotely by a retina specialist who determines the need for referral.
Dow et al’s work encourages the implementation of AI DR screening in existing workflows to improve patient follow-up rates. Therefore, the high prevalence of DR and the screening benefits to rural populations warrant ponderance on selection bias, diagnostic blind spots, and thoughtful implementation as we await integrated screening tools.
Throughout the field of ophthalmology, there is an advent of AI screening tools, including for NAION,13 glaucoma,14 and other pathologies.15 However, isolated screening comes with the risk of neglecting comorbid pathologies. Although with its own challenges, future multimodal AI screening16,17 may more effectively address this concern.
Disclosure
The authors report no conflicts of interest in this communication.
References
1. Dow ER, Chen KM, Zhao CS, et al. Artificial intelligence improves patient follow-up in a diabetic retinopathy screening program. Clin Ophthalmol. 2023;17:3323–3330. doi:10.2147/opth.S422513
2. Abràmoff MD, Lavin PT, Birch M, Shah N, Folk JC. Pivotal trial of an autonomous AI-based diagnostic system for detection of diabetic retinopathy in primary care offices. NPJ Digit Med. 2018;1:39. doi:10.1038/s41746-018-0040-6
3. Khan Z, Gaidhane AM, Singh M, et al. Diagnostic accuracy of IDX-DR for detecting diabetic retinopathy: a systematic review and meta-analysis. Am J Ophthalmol. 2025;273:192–204. doi:10.1016/j.ajo.2025.02.022
4. About common eye disorders and diseases. Available from: https://www.cdc.gov/vision-health/about-eye-disorders/index.html.
5. Varma R, Bressler NM, Doan QV, et al. Prevalence of and risk factors for diabetic macular edema in the United States. JAMA Ophthalmol. 2014;132(11):1334–1340. doi:10.1001/jamaophthalmol.2014.2854
6. Lundeen EA, Burke-Conte Z, Rein DB, et al. Prevalence of diabetic retinopathy in the US in 2021. JAMA Ophthalmol. 2023;141(8):747–754. doi:10.1001/jamaophthalmol.2023.2289
7. Beaser RS, Okeke E, Neighbours J, Brown J, Ronk K, Wolyniec WW. Coordinated primary and specialty care for type 2 diabetes mellitus, guidelines, and systems: an educational needs assessment. Endocr Pract. 2011;17(6):880–890. doi:10.4158/EP10398.OR
8. Feldman-Billard S, Dupas B. Eye disorders other than diabetic retinopathy in patients with diabetes. Diab Metabo. 2021;47(6):101279. doi:10.1016/j.diabet.2021.101279
9. Zhao D, Cho J, Kim MH, Friedman DS, E G. Diabetes, fasting glucose, and the risk of glaucoma: a meta-analysis. Ophthalmology. 2015;122(1):72–78. doi:10.1016/j.ophtha.2014.07.051
10. Varghese JS, Ravi Kumar V, Bartelt J, Hendrick AM, Pasquel FJ. The role of urban residence, race and ethnicity, and glycemic control in receiving standards of care and progression to vision-threatening diabetic retinopathy. Diab Care. 2024;48(1):29–37. doi:10.2337/dci24-0024
11. Rao P, Ramchandran R, Baldonado K, et al. Patient perspectives on accessing eye-related healthcare from rural communities. Eye. 2024;38(17):3389–3391. doi:10.1038/s41433-024-03266-z
12. Bordbar DD, Bhatnagar A, Weng CY. Use of home optical coherence tomography for retinal diseases. Int Ophthalmol Clin. 2025;65(1):41–46. doi:10.1097/IIO.0000000000000546
13. Szanto D, Erekat A, Woods B, et al. Deep learning approach readily differentiates papilledema, non-arteritic anterior ischemic optic neuropathy, and healthy eyes. Am J Ophthalmol. 2025;276:99–108. doi:10.1016/j.ajo.2025.04.006
14. Al-Aswad LA, Ramachandran R, Schuman JS, Medeiros F, Eydelman MB. Artificial Intelligence for Glaucoma: creating and Implementing Artificial Intelligence for Disease Detection and Progression. Ophthalmol Glaucoma. 2022;5(5):e16–e25. doi:10.1016/j.ogla.2022.02.010
15. Hashemian H, Peto T, Ambrósio Jr R Jr, et al. Application of artificial intelligence in ophthalmology: an updated comprehensive review. J Ophthalmic Vis Res. 2024;19(3):354–367. doi:10.18502/jovr.v19i3.15893
16. Wang S, He X, Jian Z, et al. Advances and prospects of multi-modal ophthalmic artificial intelligence based on deep learning: a review. Eye Vis. 2024;11(1):38. doi:10.1186/s40662-024-00405-1
17. Shi D, Zhang W, Yang J, et al. A multimodal visual–language foundation model for computational ophthalmology. Npj Digital Med. 2025;8(1):381. doi:10.1038/s41746-025-01772-2
© 2025 The Author(s). This work is published and licensed by Dove Medical Press Limited. The
full terms of this license are available at https://www.dovepress.com/terms
and incorporate the Creative Commons Attribution
- Non Commercial (unported, 4.0) License.
By accessing the work you hereby accept the Terms. Non-commercial uses of the work are permitted
without any further permission from Dove Medical Press Limited, provided the work is properly
attributed. For permission for commercial use of this work, please see paragraphs 4.2 and 5 of our Terms.