EDUCATIONAL SIMULATION · Synthetic patient · Not clinically validated · Not for patient care

ARDS CLINICAL TWIN v1.2

Ventilate a persistent virtual patient

Change the ventilator, advance time, and watch mechanics, gas exchange, and hemodynamics evolve in the same simulated patient.

v1.2 physiology hardening · Deployable education/research preview · Vent mechanics + live reduced HumMod cardiopulmonary physiology

01

Patient

Start with the reference aspiration ARDS teaching phenotype. The patient persists as you change PEEP, FiO₂, tidal volume, respiratory rate, and perform respiratory mechanics maneuvers.

Preloads the reference case and starts the live browser cardiopulmonary model. You can change every exposed ventilator setting afterward.

Cases are synthetic teaching constructs calibrated to published cohort envelopes; they are not deidentified patient records.

Berlin severity

—

Recruitability

—

Executable now?

—

Session readiness

Model provenance and safeguards

Berlin oxygenation severity and mechanical recruitability are independent axes. Vent owns detailed pulmonary mechanics. The default live session uses a reduced, source-aligned HumMod cardiopulmonary core in the browser; it is not the full HumMod executable and is not clinically validated. Research mode can instead replay an externally produced HumMod trajectory.

Scientific transparency: every physiologically meaningful variable, equation, threshold, condition, transition, and scenario value is being assigned machine-readable provenance. The current model is a hybrid of HumMod-exact/adapted physiology, published literature anchors, and explicitly labeled engineering or scenario assumptions.

Ventilator setup

Reference settings are prefilled for the synthetic teaching phenotype. They are simulation inputs, not treatment recommendations.

Research mode · systemic data source

Live mode couples Vent recruitment and airway pressure to a reduced, source-aligned HumMod cardiopulmonary core. It is not the full HumMod executable and is not clinically validated.

Current recruitable fraction is an explicit scenario state; it is not inferred from Berlin severity.

Waiting for explicit inputs.

01

Set up a scenario

Observe the pressure rise during filling, then the pressure drop during the inspiratory pause.

The reference phenotype has no recruitable pool. Normal tissue starts open; consolidated tissue starts closed.

Run settings

Smaller timesteps improve numerical resolution and take longer. Gas exchange is not enabled in this interface.

Ready. Choose your settings and run a simulation.

02

Explore the response

Run a scenario to see the most recent analyzed breath.

Peak pressure

—cmH₂O

Plateau pressure

—cmH₂O

Driving pressure

—cmH₂O

Inspired volume

—mL

Expired volume

—mL

Minute ventilation

—L/min

Plateau and driving pressure are shown only when a suitable volume-control pause is available.

Breath waveforms

All simulated breaths

Time runs left to right. Volume is total modeled compartment volume, not absolute anatomical lung volume.

Airway pressure cmH₂O

Run a simulation to draw waveforms

Flow L/min · positive into lung

Modeled volume mL

Recruitment at the end of the run

Model state

Compartment states will appear here.

Model details and run diagnostics

These outputs describe the mathematical model. They are not measurements from a patient.

No completed run.
What can I learn here?

In volume control, compare the filling pressure with the pressure during a low-flow pause. In pressure control, change the pressure target and observe the volume that emerges. Change one setting at a time.

For recruitment, select an injury phenotype and explicitly choose its starting open fraction. Each run restarts the scenario; changing PEEP does not continue a previous pressure history.

The three-compartment model and mechanical phenotypes are simplified teaching constructs. Gas exchange, spontaneous effort, clinical calibration, and learning outcomes are outside this release. Driving pressure here uses airway PEEP; it does not measure intrinsic PEEP.