Sapio HealthSapio Health
  • Today
  • Patients
  • Inbox
  • Calendar
Insight
  • Population
  • Quality (QOF)
System
  • Models
  • Data feeds
  • Settings
Practice trajectory
-26%
JD
Dr John Dolittle
Ashford Vale Medical Group
All engines
Service 4

Detection Engine

Three-zone CVD classifier
Version
2026.03.4
Retrained
2026-02-18
Status
healthy
Performance

Key metrics

Calibration & discrimination against the engine's published target. Under-target values are highlighted.

AUC
0.877target 0.850
Calibration (ECE)
0.032target 0.050
Sensitivity (positive)
78.0%target 75%
Specificity (negative)
91.0%target 88%
14-day inference volume
1,042 in the last 24h · avg latency 38 ms
Metrics are tracked on held-out evaluation cohorts; values here reflect the last retrain.
Model

Training basis & runtime

Transparent about what the engine was trained on and how it behaves in production.

Model family
Gradient-boosted ensemble · 15 features
Trained on
115,489 patients
Features in basis
15
Positive threshold
0.72
Negative threshold
0.18
Active cohort
2,841
History

Changelog

Each retrain is documented; clinicians can see what changed and when.

  1. 2026.03.42026-02-18
    Retrained with additional gray-zone labels; negative recall +3 pp.
  2. 2025.11.02025-11-02
    Calibrated separately by age strata ≤60 / 60–75 / ≥75.
Safety

Guardrails

Hard limits and runtime checks the engine enforces.

  • Never auto-escalates — returns zone, not action.
  • Gray-zone volume monitored; alerts on >12% drift week-on-week.
  • Every classification surfaces its top 8 contributing features.
Observability

Recent incidents & retrains

Automatic monitoring — warn / critical thresholds on calibration, drift, and latency.

  1. Gray-zone volume drift +9%warnDetection Engine15/04/2026, 22:40
    Week-on-week gray-zone share increased 51→56%. Under automatic monitoring — no action required yet.