MMOOZZGG.com

Fields

AI diabetic retinopathy screening where there is no eye doctor

2 min read3 sources

AI diabetic retinopathy screening works with a photo of the retina. From it, researchers find signs of eye diseases and learn to predict heart failure and stroke. AI screening is launched in countries short of doctors. Below are three programs and what follows from them for Russia.

On this page · 3 sections

Three programs

According to the developer, a model that detects diabetic retinopathy has supported more than 600,000 screenings in clinics around the world. For India and Thailand, the plan is 6 million screenings over 10 years, free for patients, and partners will obtain local regulatory approvals (2024). Singapore: a national pilot processed 38,215 cases in 6 clinics, and a preliminary analysis found accuracy comparable to the standard of care; a second stage checks whether technicians with AI become more accurate (2021). London: a foundation model trained on 1.6 million images can help detect retinopathy and glaucoma and predict heart failure, stroke and Parkinson’s disease, and it was released openly (Nature, 2023).

What they share

Screening is placed where there is no specialist, not instead of one, and the result goes to a doctor. Singapore ran its pilot in stages: 6 clinics in the first and 6 more in the second. Open models speed up the field: one model gives a signal for several diseases.

For a clinic in Russia

AI analysis of a retina image is a medical device that needs registration. Without it, a clinic can automate booking for screening, reminders about scheduled exams by the list and dates the doctor set, and documents. Follow-up reminders by the doctor’s dates take 14 days to roll out.

Questions and answers

Does AI replace an eye doctor’s exam?

No. It selects the patients who need a doctor, and it does this where no doctor is nearby.

Sources and fact-checking

Checked by the editors on 29 September 2026.

  1. Google, The Keyword, 17 October 2024: Google Health, India and Thailand (blog.google)
  2. Ministry of Health Singapore, 5 July 2021: Singapore Ministry of Health (moh.gov.sg)
  3. UCL News, 13 September 2023: University College London and Moorfields Eye Hospital (ucl.ac.uk)

Found a mistake? Tell us

Publications

Cases on this topic: 3

  • Market case · World

    Google Health, India and Thailand more than 600,000 diabetic retinopathy screenings worldwide, partners aim for 6 million in 10 years

    According to the company, its diabetic retinopathy AI screening model, which reads a photo of the retina, has supported more than 600,000 screenings in clinics in different countries. The studies began with an eye clinic in India and a hospital in Thailand. The company licenses the model to partners in these two countries, and their goal is 6 million screenings, free for patients, in 10 years.

    Ophthalmology

    17 October 2024 · source: blog.google

    Read the case
  • Market case · World

    Singapore Ministry of Health 38,215 cases of AI screening for diabetic eye complications at 6 polyclinics in the first phase

    The national pilot of AI retinal screening for diabetic eye complications processed 38,215 cases at 6 polyclinics in the first phase. The second phase, from December 2020, added 6 more polyclinics. By the ministry’s preliminary analysis, accuracy in finding sight-threatening conditions is comparable to the current standard of care.

    Ophthalmology

    5 July 2021 · source: moh.gov.sg

    Read the case
  • Market case · World

    University College London and Moorfields Eye Hospital a foundation model trained on 1.6 million retinal images, open to every clinic

    The first retinal foundation model for eye medicine learned from 1.6 million retinal images of a London eye hospital and was released openly. According to the authors, it can help diagnose diabetic retinopathy and glaucoma and predict diseases of the whole body: heart failure, stroke and Parkinson’s disease.

    Ophthalmology

    13 September 2023 · source: ucl.ac.uk

    Read the case
All cases: 62
Contact us