04 / RESEARCH
The research foundry
REAL COHORTS.
Synthetic futures. Prospective learning.
CardioFoundry is being built as a research environment as well as a patient and clinical product: learn from real research cohorts, generate synthetic cardiovascular states, test engines before deployment, and eventually collect structured prospective outcomes into the ontology.
128,809real research participants / source universe underlying current modeling work
FILM / RESEARCH
A cardiovascular research loop
EVIDENCE SHOULD
compound.
Follow the research path from real-world data to state construction, synthetic cohorts, engine testing, and prospective evidence.
CURRENT / QUALIFIEDPLANNED / PROJECTED
REAL-WORLD RESEARCH UNIVERSE128,809participants / records
CARDIOVASCULAR STATE
PressureAtherogenicMetabolicAdipositySmokingRenalInflammatory
PHENOTYPE → STATESYNTHETIC PATIENT GENERATORPROJECTEDsimulate states · build cohorts · explore edge cases
ENGINE LABTEST BEFORE DEPLOYMENTcohort behavior · edge cases · validation
PROSPECTIVE ONTOLOGYPROJECTEDobservations · interventions · outcomes · provenance
REAL EVIDENCE→STATE→SYNTHETIC→TEST→PROSPECTIVE→ONTOLOGY
04.1 / QUALIFICATION
A qualification snapshot
THE SCIENCE HAS
to survive its own tests.
CardioFoundry’s foundational Heart Score architecture was qualified on held-out data before becoming the frozen T6 foundation.
Qualification metrics describe internal technical agreement for the frozen T6 implementation. They are not claims of clinical outcome benefit.
04.2 / RESEARCH MODES
Three complementary modes
OBSERVE. SIMULATE.
Then learn prospectively.
Qualified research datasets with preserved phenotype definitions, missingness, covariates, model lineage, and validation.
Planned simulation layer for stress-testing engines, exploring edge states, and designing research without substituting synthetic data for clinical truth.
Planned structured capture of observations, interventions, outcomes, time, and provenance so evidence can accumulate longitudinally.