Model Boundaries

DT4H separates computational modeling from clinical authority.

DT4H produces infrastructure-level modeling signals: cohort context, reference-human priors, Twin state, calibration confidence, and runtime evidence. These are not clinical diagnoses.

Model boundaries diagram #

MODEL SIGNALSDT4H Outputs

Cohort fit · priors · Twin state · confidence · trajectory

COMPUTATIONStateK

State · transition readiness · confidence posture

EXECUTIONSETPOINT

Protocols · practices · outcomes · feedback

OVERSIGHTClinical / Governance

Validation · interpretation · licensed judgment

Model output boundaries #

DT4H Outputs

Modeling signals

  • Cohort fit and population context.
  • Reference-human priors and variance ranges.
  • Twin initialization and state estimates.
  • Calibration confidence and evidence maturity.
  • Trajectory and transition signals.
Boundary

Not clinical authority

  • Not a diagnosis.
  • Not a prescription.
  • Not autonomous medical advice.
  • Not a substitute for clinician review.
  • Not a replacement for regulated validation.

Safe interpretation chain #

DT4HModels context, priors, confidence, calibration, and Twin state.
StateKComputes current state and transition logic from the calibrated Twin.
SETPOINTTurns state into protocols, practices, outcomes, and feedback loops.
Clinician / GovernanceOwns clinical interpretation, oversight, validation, and regulated use.

Boundary examples #

Cohort fitContext only

A cohort signal provides population context; it is not a diagnosis.

ConfidenceModel maturity

Confidence describes evidence maturity; it is not medical certainty.

TrajectoryRuntime direction

Trajectory reflects modeled direction; it is not deterministic prognosis.

Protocol handoffExecution support

SETPOINT may act on bounded state, but clinical oversight remains separate.

Implementation notes #

Return bounded outputs

Model outputs should clearly identify whether they are cohort context, state estimates, confidence, or calibration signals.

Avoid diagnostic wording

Use language such as signal, state, confidence, trajectory, and readiness instead of diagnosis or prescription.

Require review paths

Boundary-sensitive workflows should include clinician, governance, or operator review paths before clinical reliance.

LayerModel Boundary
StatusActive Draft
SystemDT4H / StateK / SETPOINT
BoundaryInfrastructure, not diagnosis
System lineageDT4HTwinStateKSETPOINTOutcomesRecalibration
Infrastructure boundaryDT4H models cohorts, Twins, calibration, and runtime state. It does not diagnose, prescribe, or replace licensed clinical judgment.
Document statusInfrastructure draft
Last updatedMay 2026
Applies toDT4H.ai / AvatarK.ai ecosystem