About

DT4H exists to make longitudinal health modeling infrastructure real.

Health systems have records, visits, labs, devices, and workflows. DT4H adds the missing infrastructure layer: cohort-aware Twins that can initialize, calibrate, compute state, and learn from outcomes over time.

Mission #

PROBLEMFragmented History

EHRs, labs, notes, devices, and behaviors remain disconnected.

DT4HModel Infrastructure

Cohorts · reference humans · Twins · calibration

STATEKState Computation

Confidence · readiness · trajectory · transition logic

SETPOINTExecution

Protocols · practices · outcomes · recalibration

Why DT4H exists #

01

EHRs are transactional

They preserve encounters, but not always adaptive interpretation across time.

02

Clinician knowledge is longitudinal

Experienced clinicians carry mental models built from years of pattern recognition.

03

Wearables are signal sources

They add data, but do not automatically create calibrated person models.

04

DT4H models state

It connects context, priors, confidence, trajectory, and feedback.

05

SETPOINT executes

Bounded state becomes practices, protocols, outcomes, and recalibration loops.

06

Governance stays explicit

DT4H is infrastructure, not autonomous diagnosis or treatment.

Infrastructure philosophy #

ModelBefore recommendation

Initialize and calibrate the person model before making execution decisions.

ConfidenceBefore certainty

Expose model maturity and evidence strength instead of hiding uncertainty.

TrajectoryBefore snapshot

Model what is changing, not only what was measured once.

GovernanceBefore claims

Separate infrastructure, research, and clinical authority.

LayerGovernance Framework
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