DT4H separates source data, reference intelligence, cohort context, Twin state, runtime computation, execution feedback, and clinical interpretation boundaries.
EHR · wearables · ArenaK · clinician · SETPOINT
Cohorts · reference priors · Twin state · calibration
State · confidence · readiness · provenance
Access · privacy · clinical oversight · auditability
EHR records, wearable streams, ArenaK events, clinician notes, and SETPOINT outcomes.
Cleaned and structured evidence used for modeling and runtime interpretation.
Population priors and reference-human assumptions from research and pilot cohorts.
Individual longitudinal state, confidence, trajectory, and calibration history.
StateK confidence, readiness, transition logic, and provenance surfaces.
SETPOINT practices, adherence, check-ins, outcomes, and recalibration evidence.
Only collect data needed for modeling, calibration, validation, and execution context.
Separate raw inputs, Twin state, research views, operator views, and clinical views.
Track evidence source, freshness, confidence, assumptions, and recalibration history.
Boundary-sensitive outputs require human review and explicit governance context.
DT4H does not convert model state into diagnosis.
SETPOINT execution remains bounded and does not replace clinical treatment decisions.
Confidence must remain visible and should not be presented as clinical certainty.