Research Organisations
For Research Organisations
Pharma, clinical, and computational-biology teams see a gap between the richness of real-time monitoring data and the ability to obtain it in a reproducible, auditable form. ICU episode data is collected for clinical purposes, not for research: it lacks the provenance guarantees a trial protocol requires, and system state at the moment of a decision is rarely preserved.
The problem we solve
Research teams need:
1 · A reproducible runtime
Replaying the same clinical episode must produce the same result.
2 · A data-provenance chain
Preserved under non-repudiable audit.
3 · Visualisations
Correlating signals, decisions, and outcomes over time.
Commercial systems do not provide any of these as standard.
Our answer
HealthOS ICU provides three capabilities relevant to research:
Deterministic replay
The Compute core (the bedside execution module) is replay-safe: load any historical waveform segment with the version of the compiled bundle (rule version) active at that time, and you get the exact same output the live system produced. This is the foundation for both retrospective analysis and prospective controlled experiments.
Evidence-of-fact receipt
Every output is backed by an evidence-of-fact receipt (proof of provenance): a cryptographic chain from the approved knowledge graph through SHACL validation, the knowledge compiler, and runtime execution to the specific recommendation. This receipt is publishable as a traceable citation.
Retrospective episode review
The diagnostic contour (the closed episode-review cycle) is logged in the audit event log and visualisable via DICOMweb segment retrieval. Research teams can replay an episode, annotate decisions, and export structured data for statistical analysis.
Collaboration model
Joint study
The research organisation provides the study protocol; the platform provides a deterministic runtime, evidence-of-fact receipts, and structured data export.
Data partnership
Retrospective episode cohorts are provided through a governed export from the audit log; de-identification is part of the governance framework.
Methodological co-authorship
The closed knowledge-distillation cycle methodology is citable; co-authorship on methodology papers is possible.
Documents and training
Reference architecture
Whitepaper.
Full concept
The complete document.
Supporting research
Basis and literature.
Research Program
Closed knowledge-distillation cycle, forensic audit cycle, reproducible determinism, diagnostic contour.
Research aggregation delivery
Delivery options for research.
Research delivery options, including P-RES-AGG and P-RES-OPS, are entitlement-bound add-ons. The Program Certificate remains evidence of training; fulfillment goes through a separate admission review.
Discuss a joint study or a data partnership.