Hospitals do not operate on generic metrics; they must manage complex **resource planning and financial outcomes**. This competition forces a shift in perspective by focusing directly on the *economic impact* of clinical predictions, recognizing that every patient readmission prediction carries a highly specific financial consequence.

## The Asymmetric Cost of Readmission Prediction

In hospital resource planning, all prediction errors are not equal.

* **Successfully identifying a readmission risk** via preventive intervention (a True Positive) yields a net savings of **$4,200**. (Avoids costly readmission, saves an avg. ICU day + CMS penalty fees).
* **Wasting intervention resources** on a patient who will not be readmitted (a False Positive) results in a **$380** loss. (Unnecessary follow-up, care coordinator time, and phone triage).
* **Completely missing a patient who WILL be readmitted** (a False Negative) inflicts a massive **$9,500** blow due to unplanned emergency readmission, full re-admission costs, CMS HRRP penalties, and an average 4.9-day hospital stay.

## Your Mission

Your objective is to move beyond standard binary classification and build a **cost-sensitive 30-day readmission probability model**.