ops digital twin

Site Start-up + Recruitment Simulation Ops Twin

A2A x MCP in Clinical Ops Digital Twin

Scenario Console

Therapeutic AreaPhase

Countries

Target Sample SizeRecruitment Duration (months)

avg_startup_days75

screen_fail_rate0.35

dropout_rate0.18

competition_index0.5

patient_pool_index0.6

Show agent messages in real-timeDeterministic Seed

Run History

Stored locally (max 5 runs). Each run keeps full context + event log.

A2A Conversation Thread

MCP-like Shared Context

Run a simulation to populate shared context.

Simulation Output

Run a scenario to generate recruitment and startup simulation outputs.

update log

Prototype Change Log

2026-03-01 · v0.2.0
Forecast-Kalibrierung und Confidence-Layer
- - Erfassung beobachteter Ist-Werte zur laufenden Prognosefehler-Verfolgung (predicted vs actual) hinzugefuegt.
- - Automatische Parameter-Rueckkopplung fuer Startup/Screen-Fail/Dropout/Competition/Patient-Pool integriert.
- - Confidence-Panel mit MAPE, Signed Bias, Tracking-Punkten und Parameter-Shift-Diagnostik ergaenzt.
2026-03-01 · v0.1.0
Ops Twin MVP gestartet
- - Dreizonen-UI ergaenzt: Scenario Console, A2A-Timeline und MCP-aehnliches Shared-Context-Panel.
- - Sechs-Agenten-Runtime mit deterministischer Seed-Simulation und geordnetem Handoff umgesetzt.
- - Context-Patch/Event-Log mit Diff-Drawer, JSON-Export, Replay und lokaler Run-Historie hinzugefuegt.