Infection Control Surveillance

The problem

Infection-control teams review hundreds of facility metrics at once, so naive comparisons produce a flood of false alarms.

Approach

1. 01

   Combined CDC and CMS hospital-acquired infection datasets into a consistent facility-measure panel.

2. 02

   Used funnel-plot control charts to flag outliers relative to expected variation at each volume level.

3. 03

   Applied FDR correction across the full set of tests to control false discoveries during multiple comparisons.

4. 04

   Added a predictive layer to highlight facilities trending toward outlier status.

Outcome

• A signal-first watchlist rather than an unmanageable list of statistically noisy flags.

• Transparent statistics — every flag shows its expected range and correction context.

• Early-warning view for facilities drifting before they cross a threshold.