These are fully synthetic patients. Each was authored as a coherent multi-year kidney-care journey — demographics, family, a care team, and a chain of encounters, dialysis visits, and hospitalizations — so the same person can be re-used across every framework in this project.
Now with generated Labcorp-style lab reports, imaging/pathology studies, and red-team aberrant data integrated at each patient's correct dates.
They exist to let us probe one question: can a system reason honestly about a real clinical record — knowing when it can determine something and when it cannot?