Condition Score Cross-Fleet Calibration Index (CSFCI)

GMI-HC-033HealthcareIntegration-Required

Definition

CSFCI measures the statistical consistency of condition health scores across device populations. A score of 85 on a GE ventilator should mean the same thing as 85 on a Dräger ventilator if fleet-level comparisons are to be trusted. CSFCI quantifies the calibration gap: the degree to which scoring algorithms trained on different device populations produce systematically different score distributions for equipment in equivalent condition.

The Measurement Gap

Measurement Gap and Decision

Decision Question: Are condition health scores comparable across device types and manufacturers, or are fleet-level keep/replace decisions being made on scores that mean different things for different equipment?

What Goes Unmeasured Today:

Condition-based maintenance systems assign health scores to medical devices based on equipment telemetry. Run hours, error logs, self-test results, calibration drift, power consumption patterns. Clinical engineering uses these scores for fleet-level decisions: which devices to replace first, which to extend, where to allocate capital. The problem is that scoring algorithms are trained on different device populations with different degradation curves. A health score of 85 on a 5-year-old device from one manufacturer reflects a different actual condition than 85 on a 12-year-old device from another manufacturer. The scoring systems were not designed for cross-fleet comparison. Clinical engineering managers compare scores across their entire fleet, ranking devices by health score to prioritize replacement spending. Devices with generous scoring algorithms get kept too long. Devices with conservative scoring get replaced too early. Capital allocation is systematically biased, but the bias is invisible because everyone assumes the scores are comparable.

Invisible Failure Mode:

Cross-fleet score comparisons drive capital decisions worth millions of dollars annually. If scores are not calibrated across device types, the fleet renewal strategy systematically misinvests: overretaining some devices and prematurely replacing others. No current metric validates cross-fleet score comparability.

Formula

Primary Formula:

CSFCI = 1 − max|μ_type_i − μ_type_j| / Score_Range

Across all pairwise comparisons of device type mean scores for equipment in equivalent condition bands.

Variable Definitions:

Variable Definition Source
μ_type_i Mean health score for device type i within a defined condition band (e.g., devices 3–5 years old with no unplanned failures in past 12 months) Condition monitoring platform
Score_Range Total possible score range (typically 0–100) Condition monitoring platform
Condition bands Groups of devices with equivalent operational characteristics (age, failure history, utilization) used to normalize comparison CMMS/EAM + utilization data

Data Requirements:

Condition health scores by device type and manufacturer. Device age, utilization, and failure history for condition-band normalization. Minimum 20 devices per type per condition band for statistical stability.

Edge Cases and Limitations:

Defining "equivalent condition" across device types requires clinical engineering judgment. Age and failure history are proxies, not guarantees. Device types with fewer than 20 units produce unreliable comparisons. Scoring algorithm updates from vendors may improve or worsen calibration: recalculate CSFCI after any vendor algorithm change.

Proposed formula, pending validation pilots.

At a Glance

Category
Biomedical Equipment Lifecycle Intelligence
← AEM Interval Compliance Alignment Rate (AICAR)CBM Inspection Scope Coverage Rate (CISCR) →

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