🧪 Health Diagnosis Simulation

A hidden patient is assigned one of four possible chest-pain diagnoses — acute myocardial infarction, GERD, costochondritis, or pulmonary embolism — each starting from a clinically plausible prior probability. Ordering a test (ECG, troponin, D-dimer, or CT pulmonary angiogram) draws a positive or negative result from that hidden condition's real per-disease positive rate, and the simulator applies Bayes' rule live across all four candidates at once, renormalising their probabilities and animating a 3D bar chart that shows the leading diagnosis pulling ahead as evidence accumulates. A cost and radiation tracker keeps the trade-off between more information and more expense visible, and a confidence-threshold slider marks the point at which a real clinician would feel safe committing to a diagnosis — confirm your answer against the hidden true condition to see whether the evidence actually supported it.