42%

It is 3am in the ICU. A patient is admitted. The monitoring system calculates their mortality probability and displays a single number on your screen.

You are the doctor. You have 30 seconds to decide whether to escalate treatment, maintain it, or withdraw care. But the algorithm does not explain itself. The sensors can fail silently. And the data carries the weight of history.

This simulation explores how predictive algorithms in ICUs transform mortality from a biological event into an engineered probability, and what that means for the doctors, patients, and systems caught in between.