CMC Vellore, IIIT-H Build Multilingual AI Companion to Support Cancer Patients Between Hospital Visits
Advertisement
BandhuCare currently supports eight Indian languages through both text and voice interactions, including an AI-powered consent agent that explains medical consent forms in simple language.
Christian Medical College (CMC) Vellore and the International Institute of Information Technology Hyderabad (IIIT-H) have developed BandhuCare, a multilingual AI companion designed to improve communication between cancer patients and clinicians.
The platform has reached Technology Readiness Level (TRL) 5 and will soon begin clinical pilot studies among head and neck cancer patients.
The multilingual AI companion enables patients to ask routine post-treatment questions in their preferred language, report symptoms through natural conversations, and help clinicians monitor their condition between follow-up visits.
BandhuCare has been designed to rely only on clinician-approved medical information rather than internet sources.
The platform uses Retrieval-Augmented Generation (RAG) combined with an additional AI safety layer that validates every response against an approved medical knowledge base before it is shared with patients, reducing the possibility of AI hallucinations.
Unlike traditional symptom questionnaires, the platform allows patients to communicate through text or voice in a conversational format.
Patients can maintain daily health journals, describe symptoms as they occur, and record concerns between appointments.
The AI then converts these interactions into structured clinical summaries and standard Patient Reported Outcome Measures (PROMs), enabling doctors to review a patient's progress before consultations.
Dr Narender Kumar Thota, consultant hemato-oncologist and stem cell/bone marrow transplant specialist at KIMS Hospitals, Secunderabad, said, "Doctors are trained to extract as much relevant history as possible, and in most cases they do. But patients can forget details, and even doctors can occasionally miss something. A structured AI questionnaire can improve the quality of information collected and help explore symptoms in greater depth."
He further noted that AI can make the process more dynamic by tailoring follow-up questions based on previous responses instead of relying on a fixed checklist, thereby saving time while improving documentation.
BandhuCare currently supports eight Indian languages through both text and voice interactions.
It also includes an AI-powered consent agent that explains medical consent forms in simple language and checks whether patients have understood the information before consent is obtained.
He added that the structured digital records generated through the platform could also contribute to research, public health planning and disease surveillance by creating standardised datasets.
The BandhuCare consortium includes CMC Vellore, IIIT Hyderabad, startup Revan AI, AIIMS Guwahati and patient advocacy platform Patients Engage.
The project recently received the IndiaAI-National Cancer Grid CATCH Grant for Cancer 2026.
Following clinical pilots, the team plans to expand the multilingual AI companion to support patients with other cancers as well as chronic conditions such as diabetes, cardiovascular diseases and those requiring dialysis care.
Stay tuned for more such updates on Digital Health News
