In a country where a single doctor may stand between millions of patients and the care they need, India is preparing to place a quiet new companion beside the clinician: an artificial intelligence that offers a short list of likely diagnoses. Built under the national digital health mission with AIIMS and the Centre for Chronic Disease Control, and woven into the eSanjeevani telemedicine network, the Clinical Decision Support System is meant to steady judgment rather than supplant it. Its promise rests on a familiar human hope, that fewer illnesses slip past unnoticed and that care arrives soon
AI Decision-Support Tool for Doctors Set for Rollout Across Indian Hospitals
Doctors can accept or reject AI suggestions, and their feedback improves the system
Mark
What is this tool, in plain terms?
Mimi
It's an AI assistant for doctors. A health worker at a village-level clinic records symptoms, and a doctor at a bigger hospital gets three suggested diagnoses to consider.
Mark
Why split it into two parts?
Mimi
It mirrors how care is organised. The spoke is the Ayushman Arogya Mandir, where mid-level providers use the patient assistance form. The hub is the primary, secondary or tertiary hospital where the doctor reviews the case.
Luke
Before we get comfortable, the official says it has high accuracy in its top three. Is there a number?
Mimi
No. The report gives no figure and no independent study.
Luke
So that's the ministry's characterisation, not a verified result.
Mimi
Right. What is concrete is the scale: 282 million eSanjeevani consultations from April 2023 to November 2025, per the official.
Mark
Does that mean 282 million diagnoses were made by AI?
Luke
Not as stated. The official says those consultations have benefited, and the report doesn't explain how that was measured.
Mimi
The safeguard they emphasise is that doctors can accept or reject suggestions, and their feedback improves the system.
Mark
What happens next?
Mimi
States and Union Territories have been told to update their software and get it approved under ABDM. How fast they do that will shape the rollout.
O Pulso
- A nationwide rollout of an AI diagnostic tool is moving toward public and private hospitals alike, raising the stakes for every clinic that adopts it.
- Frontline health workers at spoke clinics now enter symptoms through a guided multilingual form, a shift that reshapes how patients are first recorded and referred.
- At hub hospitals, doctors receive three probable diagnoses drawn from structured data, testing whether a machine's shortlist can sharpen decisions under heavy workloads.
- States and Union Territories are being asked to update their software and secure ABDM approval, and the speed of that work will determine how quickly the system reaches patients.
- Officials insist doctors retain final control, yet the accuracy claims still await independent validation, leaving the outcome unsettled.
In a country where a single doctor may stand between millions of patients and the care they need, India is preparing to place a quiet new companion beside the clinician: an artificial intelligence that offers a short list of likely diagnoses. Built under the national digital health mission with AIIMS and the Centre for Chronic Disease Control, and woven into the eSanjeevani telemedicine network, the Clinical Decision Support System is meant to steady judgment rather than supplant it. Its promise rests on a familiar human hope, that fewer illnesses slip past unnoticed and that care arrives sooner, while the question of how far such trust should travel remains open.
India's Health Ministry is preparing to introduce an artificial intelligence tool that suggests possible diagnoses to doctors, with deployment planned across both public and private hospitals. The tool, the Clinical Decision Support System, is meant to help doctors reach faster, evidence-based decisions, and it is offered free of charge under the Ayushman Bharat Digital Mission. It was developed with AIIMS and the Centre for Chronic Disease Control, and it is already linked to eSanjeevani, the national telemedicine service.
The system was trained on conditions commonly seen in teleconsultations, such as respiratory infections, gastritis, fever and diabetes. It works in two stages that mirror how care is organised. In the Ayushman Arogya Mandirs, mid-level providers record a patient's complaints through a form available in 13 languages, which guides them through simple multiple-choice questions and suggests the right specialty for referral. At hub hospitals, doctors then receive three probable diagnoses based on age, gender, symptoms and their details.
Officials stress that clinicians remain in charge. They can accept or reject the suggestions, and their feedback helps refine the tool. Yet the accuracy claims come from the ministry itself, and independent validation has not yet been published. As states update their software and seek approval, the coming months will show whether the promise of fewer missed diagnoses holds in practice.
India's Health Ministry is preparing to roll out an artificial intelligence tool that suggests possible diagnoses to doctors, with deployment planned across both public and private hospitals. The tool is called the Clinical Decision Support System, or CDSS, and it is meant to help doctors reach faster, evidence-based decisions.
CDSS is an initiative under the Ayushman Bharat Digital Mission (ABDM), developed with the All-India Institute of Medical Sciences (AIIMS) and the Centre for Chronic Disease Control (CCDC). It is offered to doctors as free digital assistance. A Health Ministry official, who described the system, said it is integrated with eSanjeevani, the national telemedicine service.
The diagnostic model was trained on the symptoms and diseases most often seen in eSanjeevani teleconsultations. The official named respiratory tract infections, gastritis, fever and diabetes among them. CDSS was integrated into eSanjeevani in April 2023, and the official said that from then until November 2025, 282 million eSanjeevani consultations have benefited.
The system works in two parts, matching the way care is organised. At the spoke level, in the Ayushman Arogya Mandirs, mid-level healthcare providers use a patient assistance form (PAF) to record a patient's complaints. At the hub level, in primary, secondary and tertiary care hospitals, doctors get the differential diagnosis feature.
The PAF has been validated by AIIMS Rishikesh. It is available in 13 languages and follows the international Systematized Nomenclature of Medicine - Clinical Terms standards. Providers enter only minimal input: a symptom and its attributes. After a symptom is selected, the form leads the provider through a sequence of multiple-choice questions on details such as duration, severity and location, which the official said reduces typing errors. Based on the data, the form also recommends which specialty department the patient should be referred to for teleconsultation, aiming for more appropriate referrals from the spoke.
At the hub, the differential diagnosis tool takes what was captured in the PAF and offers the doctor three possible diagnoses. The official said the model analyses structured data, including age, gender, symptoms and their attributes. According to him, this lowers the chance of a missed or delayed diagnosis and helps hub doctors decide on treatment and advice. He said the model has high accuracy in its top three recommendations and acts as a smart filter that reduces doctors' workload.
The official stressed that doctors stay in control. They can accept or reject the AI's suggestions, and their feedback continuously improves the system. He described the design as a Human-in-the-Loop approach and said AI does not replace clinical judgment.
As part of the rollout, all States and Union Territories have been advised to update their software and get it approved under ABDM so the integration can work. The accuracy claims come from the ministry official; the report does not cite independent validation or give figures, so those are worth watching as deployment proceeds, along with how quickly states complete the software updates.
Citações Notáveis
AI doesn't replace clinical judgment; doctors can accept or reject the AI's suggestions, and their feedback continuously improves the system.— Health Ministry official
Quer a matéria completa?
Leia o original em The Hindu ↗
Health & Medicine