Recognition Accuracy Population Variance (RAPV)

GMI-FER-004Fire and EMS ResponseIntegration-RequiredPattern O-22: Fairness Drift

Definition

The spread in high-acuity recognition accuracy across caller strata defined by language, interpreter use, caller age band, and distress level.

The Measurement Gap

There is no metric that tells you how cardiac arrest and high-acuity recognition accuracy varies across caller language, dialect, age, and distress level, which means the populations where recognition performs worst receive slower telecommunicator intervention and the disparity is averaged out of every figure the agency reports.

Formula

RAPV = max_s(Sens_s) - min_s(Sens_s) ; with per-stratum deltas Δ_s = Sens_s - Sens_all

Proposed formula, pending validation pilots.

At a Glance

Unit
Percentage points
Trend Direction
Lower
Category
Call Triage and Dispatch Determination Quality

Decision Supported

Whether model-assisted recognition should be relied on uniformly across caller populations or gated for strata where it underperforms.

← Elicitation-Attributable Error Share (EAES)Aid Boundary Load Externalization (ABLE) →

Assess Your Organization

The AI-GPS Assessment evaluates your AI governance maturity and recommends KPIs like this one for your implementation roadmap.

Take the Free AI-GPS Assessment