Patient Twin

Patient Twin is a training simulator on synthetic patients. A trainee works one patient through a 12-hour admission. It will serve trainees, medical doctors, nurses, to identify high risk diagnosis and treat them appropriately, failure to do so will be seen in the patient course as bad decisions would be shown during the bedside nursing updates. About 795,000 Americans die or are permanently disabled by diagnostic error each year, many present with sepsis or heart failure and a different treatment is provided causing high risk for complications and loss of revenue to the system. Workflow: 1. Trainee interviews the patient on a live video call. 2. They pull a week of smartwatch and home-device data and order tests in their own words. 3. They commit to a diagnosis and plan. 4. They answer the night nurse at 3:40 AM and the day nurse at 9:40 AM to receive updates and act on the updates 5. They write the note, from scratch or from an ambient AI tool like the ones currently in use. 6. They get an honest score with critical-error caps.