OpenAI powers 10BedICU's push to fix India's critical care gap

India has roughly one oncologist for every 2,000 cancer patients. The United States has one for every 100. That brutal math is what 10BedICU has been fighting since the Delta wave pushed the country’s healthcare infrastructure past its breaking point in 2021. With only 2.3 ICU beds per 100,000 people, the crisis exposed a gap that founder Srikanth Nadhamuni couldn’t ignore. His initiative, part of the eGov Foundation, started by wiring up government hospitals with medical tech and a telehealth platform called CARE. Now it’s tapping OpenAI’s models to stretch what overworked clinicians can actually accomplish in a day.

10BedICU isn’t a scrappy pilot anymore. The network spans over 200 hospitals across nine states, all connected through CARE. The model is a deliberate public-private handshake: 10BedICU puts cutting-edge equipment into facilities that serve patients who can’t afford private care, and the government absorbs the ongoing operating costs. Dr. Subranshu Bhattacharya, a senior medical officer in Assam, points to a tangible shift in capability. Cardiac, cerebrovascular, and respiratory failure cases that once required referrals to distant specialists can now be managed on-site.

The real bottleneck, however, isn’t just beds or ventilators. It’s specialists. Even telehealth can’t fully close a gap where major cities themselves are short-staffed. That’s where the new suite of OpenAI-powered tools comes in. CARE Scribe uses Whisper and GPT-4 to transcribe doctor-patient conversations in English, Hindi, Malayalam, and Bengali, automatically converting them into structured EMR entries. Early testing in a Kerala palliative care center suggests it cuts data-entry time by over 50% while also improving record quality, simply because nurses can speak in their native language.

Two other tools are in the pipeline, both built on GPT-4. CARE Device Connect uses GPT-4 Vision to pull readings from older, otherwise incompatible hospital monitors via high-resolution cameras, piping real-time data into the platform for continuous monitoring. The State of Karnataka is piloting it across 43 hospitals with an eye toward turning existing ICUs into “SmartICUs.” Meanwhile, CARE Discharge Summary automates the summarization of patient records. Nadhamuni insists on a slow, iterative rollout, testing features in partnership with state governments and keeping patient data—stripped of personally identifying information—locked down under local privacy laws. The ambition isn’t subtle: use AI not to replace clinicians, but to carve out enough administrative slack so they can actually practice medicine.