IndicStack Reports · First published 8 October 2026
Silent until it isn’t
What AI can and can’t do for diabetic retinopathy screening in India.
For teams building healthcare AI, the screening journey extends beyond the model. This article examines the workflow around it, the evidence behind it, and the questions that remain.
Dr Prasanth Ghanta, PhD
Founding Scientist, Life Sciences & Healthcare
Inside the report
The screening workflow
Image quality, the AI’s first read, human rechecks and the path to the eye clinic.
The evidence and its limits
Studies from India and Rwanda, and what they tell us about screening and follow-up.
Designing for real settings
Practical considerations for camps, primary health centres and people’s homes.
The photograph as health data
Consent, records and the questions to ask whoever hosts your AI.
Take a look inside
A preview of the article’s opening page. The complete report includes the figures, references and author details.
Written for healthcare AI builders, life-sciences teams and researchers.
Email me the report