India has the second-largest diabetic population on earth — over 100 million people living with diabetes, and millions more undiagnosed. Of these, a significant proportion will develop diabetic retinopathy (DR), the leading cause of preventable blindness in working-age adults. The tragedy is that most of these cases are entirely avoidable with timely screening. The barrier has never been treatment — it has been access. AI is now removing that barrier at scale. The Scale of the Problem in India India has roughly one ophthalmologist for every 100,000 people in rural areas. A fundus camera examination requires a trained specialist to interpret retinal images. In a country where the diabetic population is concentrated in tier-2 cities, towns, and rural districts — where specialist access is minimal — routine DR screening has been systematically out of reach. The result: patients present with advanced, sight-threatening DR that could have been caught and treated at an early, reversible stage. How AI DR Screening Works AI-based DR screening works by analysing retinal fundus photographs using deep learning models trained on large datasets of graded retinal images. The AI classifies the image into severity categories: no DR, mild, moderate, severe non-proliferative DR, and proliferative DR. High-sensitivity models flag referable cases for ophthalmologist review, while non-referable cases are cleared with a high negative predictive value. This means a trained technician with a fundus camera can run the screening; the AI provides the interpretation; only flagged cases require a specialist. MadhuNetrAI: India's Own Screening Tool MadhuNetrAI is an indigenously developed AI tool for diabetic retinopathy screening that has already screened over 7,100 patients across India. It was developed with Indian patient data, which matters significantly: retinal pathology can differ across ethnic populations, and models trained on Western datasets may underperform on Indian eyes. MadhuNetrAI is designed for deployment in primary health centres, community settings, and diabetes camps where specialist ophthalmologists are absent. Where AI DR Screening Is Being Deployed in India AI DR screening programmes are now active or piloted across multiple channels in India: (1) NCD clinics under Ayushman Bharat Health and Wellness Centres, (2) diabetes camps run by state health departments and NGOs, (3) corporate and urban diabetes management programmes, (4) integrated into private hospital diabetology departments as a standard annual review tool, and (5) via eSanjeevani-linked teleconsultation workflows where a district-level technician captures images and AI flags referrals to centrally based retina specialists. What Non-Ophthalmologist Doctors Need to Know General physicians, diabetologists, endocrinologists, and family medicine doctors are the frontline for diabetes management in India. The responsibility to recommend annual DR screening falls to them, not to ophthalmologists. With AI screening tools now accessible at community level, the clinical protocol has changed: every Type 2 diabetic patient at diagnosis should have a baseline retinal screen, then annually if no DR is detected. Doctors should proactively counsel patients on this, especially those with HbA1c above 8%, duration of diabetes above 5 years, or co-existing hypertension. Clinical Accuracy: What the Evidence Shows AI DR screening tools validated on Indian datasets have demonstrated sensitivity above 90% for detecting referable diabetic retinopathy. This compares favourably with human grader performance in community settings, where fatigue, lighting conditions, and image quality variation can lower diagnostic accuracy. The IndiaAI Mission has specifically funded the curation of Indian retinal imaging datasets to improve and validate these models further, with AIIMS Delhi serving as a Centre of Excellence for AI diagnostics evaluation. Integrating AI DR Screening Into Your Diabetic Clinic Workflow For diabetologists and general physicians running high-volume diabetes clinics, the practical pathway is: (1) At diagnosis or first annual review, flag the patient for retinal screening via a co-located or nearby fundus photography station. (2) The fundus image is captured by a trained technician or ophthalmic assistant. (3) The AI tool processes the image and produces a graded output. (4) Referable cases are sent to an ophthalmologist with the AI report attached. (5) Non-referable cases are documented in the patient's ABDM record with a 12-month recall. This workflow can be run without an on-site ophthalmologist. Q&A: What Indian Doctors Are Asking About AI DR Screening Q: Can AI replace a retina specialist for DR screening? A: For population-level screening to identify referable vs non-referable cases, yes — AI performs comparably to trained graders. It does not replace a retina specialist for diagnosis, staging confirmation, or treatment planning. The clinical value of AI is in dramatically expanding who gets screened, not in eliminating specialist review of complex cases. Q: What fundus camera do I need to run AI DR screening at my clinic? A: Most validated AI tools work with non-mydriatic fundus cameras. Entry-level devices from manufacturers like Topcon, Remidio (India), and Forus (India) are compatible with leading AI screening platforms. Remidio's smartphone-based fundus camera is specifically designed for Indian primary care settings and pairs with cloud-based AI grading. Q: Is AI DR screening covered under Ayushman Bharat PM-JAY? A: As of 2026, specific AI-based screening services are being piloted under the Ayushman Bharat Health and Wellness Centre framework. Formal PM-JAY package codes for AI-graded fundus screening are under development. Check with your state NHM office for current coverage status in your district. Q: How do I counsel a diabetic patient about DR screening? A: Use direct language: 'Diabetes can damage the blood vessels in your eyes silently. You can lose vision without any warning symptoms. A simple 5-minute eye photograph once a year can detect this early when it is still treatable. I am referring you for this test today.' Patients who understand the asymptomatic nature of early DR are more likely to comply. How Doctrust Connects Preventing blindness in diabetic patients starts with making sure every eligible patient gets referred for annual retinal screening — and that the referral, result, and recall are documented. Doctrust helps diabetologists and general physicians build this into their clinic workflow automatically: flag high-risk diabetic patients, track annual screening compliance, and maintain ABDM-linked longitudinal records that ensure no patient falls through the cracks. The best AI diagnostic tool in the world only works if the patient shows up. Doctrust makes sure they do. Book a free demo at www.doctrust.in.