ClarusC64/clinical-quad-perfusion-buffer-lag-coupling-trauma-deterioration-v0.6

What this repo does

This repository contains a Clarus v0.6 intervention pathway dataset focused on trauma deterioration dynamics.

The dataset evaluates whether a model can determine if a proposed intervention meaningfully stabilizes a deteriorating trauma system.

The task requires reasoning from:

system state

trajectory toward instability

boundary geometry

recovery geometry

intervention vector

projected trajectory consequence

The model cannot read the answer directly.

It must infer stabilization from the structure of the case.

This shifts the benchmark from simple deterioration detection to intervention reasoning.

Core quad

The trauma deterioration system is represented using four normalized variables.

perfusion_pressure

physiological_buffer

intervention_lag

systemic_coupling

These variables capture the core structural drivers of trauma cascade progression.

Clinical variable mapping

The normalized quad variables correspond to measurable clinical signals.

Quad Variable Clinical Measurements Typical Indicators perfusionpressure Mean arterial pressure (MAP)Systolic blood pressureShock indexCapillary refill time MAP < 65Shock index > 0.9SBP < 90 physiologicalbuffer Physiological reserveHemoglobin levelPlatelet countCoagulation capacity Hb < 8Platelets < 100KCoagulopathy present interventionlag Time to hemorrhage controlTime to transfusionDelay to airway stabilizationDelay to surgical intervention Transfusion delay > 30 minDelayed hemorrhage control systemiccoupling Multi-organ coupling stressShock–respiratory interactionInflammatory cascade activationCoagulopathy–bleeding feedback Rising SOFA scoreLactate increaseSystemic inflammatory activation

These measurements illustrate how normalized values in the dataset map to real trauma physiology.

Prediction target

The target column is:

labeltraumastabilization

This label indicates whether the intervention pathway produces genuine stabilization.

Label logic Default benchmark rule

A row is labeled positive only when both conditions hold:

stabilizationsuccess = 1 trajectoryshift < -0.10

This rule filters out marginal corrections and ensures that positive examples represent meaningful stabilization.

Optional relaxed rule

Positive labels may trigger when:

stabilization_success = 1

This relaxed rule may be used for exploratory builds but is not the default benchmark configuration.

Row structure

Each row contains:

core trauma state

trajectory geometry

perturbation geometry

recovery geometry

intervention vector

projected trajectory consequence

Train rows include:

stabilization_success

labeltraumastabilization

Tester rows exclude these fields.

Why tester rows exclude stabilization_success

The tester file withholds:

stabilization_success

labeltraumastabilization

This prevents answer leakage.

The model must infer stabilization using:

the starting trauma state

drift toward the failure boundary

recovery basin geometry

the intervention vector

the predicted trajectory consequence

This structure forces real intervention reasoning.

Minimal intervention path

minimalinterventionpath encodes the shortest viable stabilization pathway.

Example interpretation:

0 = no viable rescue path 1 = direct stabilization pathway 2 = multi-step stabilization sequence 3 = complex high-risk rescue pathway

The field remains visible because the benchmark evaluates whether the model can combine intervention structure with trajectory consequence to determine stabilization.

Files

data/train.csv — labeled training dataset

data/tester.csv — unlabeled benchmark dataset with withheld stabilization signal

scorer.py — evaluation metrics and confusion matrix computation

cli.py — command-line evaluation wrapper used for benchmark scoring

README.md — dataset card and schema documentation

Evaluation

The scorer reports:

accuracy

precision

recallsuccessfulstabilization

failedrescuerate

f1

confusion_matrix

Primary metric:

recallsuccessfulstabilization

Secondary metric:

failedrescuerate

Interpretation:

recallsuccessfulstabilization measures how reliably the model detects interventions that genuinely stabilize the trauma system.

failedrescuerate measures how often the model fails to recognize a viable stabilization pathway.

These metrics prioritize intervention reasoning rather than generic classification accuracy.

Schema train.csv columns

scenarioid perfusionpressure physiologicalbuffer interventionlag systemic_coupling

driftgradient driftvelocity driftacceleration boundarydistance

perturbationradius collapsetrigger

recoverydistance recoverygradient return_feasibility

deltaperfusionpressure deltaphysiologicalbuffer deltainterventionlag deltasystemiccoupling

trajectoryshift minimalinterventionpath stabilizationsuccess

labeltraumastabilization

tester.csv columns

scenarioid perfusionpressure physiologicalbuffer interventionlag systemic_coupling

driftgradient driftvelocity driftacceleration boundarydistance

perturbationradius collapsetrigger

recoverydistance recoverygradient return_feasibility

deltaperfusionpressure deltaphysiologicalbuffer deltainterventionlag deltasystemiccoupling

trajectoryshift minimalintervention_path

Structural note

The Clarus dataset series evolves through progressively richer representations of cascade dynamics.

Version progression:

v0.1 — cascade state detection v0.2 — trajectory-aware detection v0.3 — dynamic cascade forecasting v0.4 — boundary discovery v0.5 — recovery geometry v0.6 — intervention pathway reasoning

Earlier versions identify when instability is developing.

Version 0.6 evaluates whether a proposed intervention meaningfully alters the trajectory of a system approaching collapse.

This marks the transition from monitoring cascade dynamics to evaluating control pathways.

Production deployment

This dataset structure can support clinical decision environments where trauma deterioration must be detected and corrected before irreversible transition occurs.

Example settings include:

emergency trauma intake

hemorrhagic shock monitoring

surgical escalation planning

ICU trauma surveillance

multi-organ trauma cascade monitoring

Enterprise and research collaboration

This dataset class supports benchmarking for:

intervention-aware clinical AI

trauma cascade modeling

recovery feasibility prediction

false-stability detection

boundary-sensitive decision support systems

Contact

For dataset expansion, custom coherence scorers, or deployment architecture:

team@clarusinvariant.com

Instability is detectable. Governance determines whether it propagates.

License

MIT