Datasets:

ClarusC64
/

clinical-quad-cardiac-load-reserve-lag-coupling-heart-failure-transition-v0.8

scenario_id string	cardiac_load float64	reserve_capacity float64	perfusion_lag float64	organ_coupling float64	drift_gradient float64	drift_velocity float64	drift_acceleration float64	boundary_distance float64	secondary_boundary_distance float64	boundary_competition_ratio float64	boundary_uncertainty float64	trajectory_uncertainty float64	regime_confidence float64	regime_transition_score float64	transition_direction string	regime_separation_margin float64	transition_uncertainty float64	transition_velocity float64	perturbation_radius float64	collapse_trigger string	recovery_distance float64	recovery_gradient float64	return_feasibility float64	delta_cardiac_load float64	delta_reserve_capacity float64	delta_perfusion_lag float64	delta_organ_coupling float64	trajectory_shift float64	minimal_intervention_path string	stabilization_success int64	label_heart_failure_transition int64
CHFTR001	0.82	0.28	0.71	0.76	0.44	0.39	0.21	0.18	0.23	1.28	0.14	0.19	0.62	0.58	congestion_to_low_output	0.12	0.18	0.22	0.31	pulmonary_edema	0.27	-0.16	0.54	-0.18	0.14	-0.12	-0.1	-0.17	reduce preload; augment contractility; improve oxygenation	1	1
CHFTR002	0.88	0.22	0.79	0.83	0.61	0.54	0.33	0.11	0.13	1.18	0.19	0.24	0.49	0.74	congestion_to_cardiogenic_shock	0.08	0.23	0.35	0.36	hypotension_spiral	0.34	-0.05	0.28	-0.06	0.03	-0.04	-0.02	-0.04	vasopressor support; inotrope escalation; invasive monitoring	0	0
CHFTR003	0.76	0.36	0.64	0.69	0.29	0.24	0.11	0.26	0.41	1.58	0.1	0.13	0.71	0.33	congestion_to_low_output	0.19	0.12	0.15	0.24	renal_hypoperfusion	0.22	-0.14	0.66	-0.12	0.1	-0.09	-0.08	-0.15	diuresis; afterload reduction; monitor urine output	1	1
CHFTR004	0.91	0.18	0.84	0.87	0.73	0.61	0.41	0.09	0.1	1.11	0.22	0.27	0.42	0.81	congestion_to_cardiogenic_shock	0.06	0.26	0.42	0.39	multi_organ_hypoperfusion	0.39	-0.03	0.21	-0.03	0.01	-0.01	0	-0.02	mechanical support evaluation; vasopressor; ICU escalation	0	0
CHFTR005	0.69	0.43	0.58	0.61	0.21	0.18	0.07	0.33	0.52	1.58	0.08	0.1	0.79	0.27	stable_congestion	0.24	0.09	0.1	0.19	volume_overload	0.18	-0.11	0.73	-0.09	0.08	-0.07	-0.06	-0.13	gentle diuresis; sodium restriction; reassess filling pressure	1	1
CHFTR006	0.86	0.25	0.77	0.81	0.57	0.49	0.28	0.14	0.15	1.07	0.17	0.22	0.53	0.77	congestion_to_cardiogenic_shock	0.07	0.21	0.31	0.34	shock_onset	0.32	-0.04	0.32	-0.05	0.02	-0.03	-0.01	-0.06	optimize MAP; inotrope trial; urgent echo reassessment	0	0
CHFTR007	0.74	0.39	0.6	0.65	0.26	0.22	0.09	0.29	0.34	1.17	0.11	0.12	0.75	0.49	congestion_to_low_output	0.14	0.13	0.18	0.23	pulmonary_congestion	0.2	-0.15	0.68	-0.11	0.09	-0.08	-0.07	-0.16	loop diuretic; nitrates; noninvasive ventilation	1	1
CHFTR008	0.93	0.16	0.88	0.9	0.79	0.66	0.46	0.07	0.08	1.14	0.24	0.29	0.38	0.86	congestion_to_cardiogenic_shock	0.05	0.28	0.45	0.41	refractory_low_output	0.42	-0.02	0.18	-0.02	0	-0.01	0.01	-0.01	MCS pathway; vasopressor-inotrope combination; transfer	0	0
CHFTR009	0.71	0.41	0.56	0.59	0.19	0.16	0.05	0.35	0.49	1.4	0.07	0.09	0.81	0.22	stable_congestion	0.22	0.08	0.09	0.17	fluid_retention	0.16	-0.12	0.77	-0.08	0.07	-0.06	-0.05	-0.14	oral diuretic optimization; daily weights; BNP follow-up	1	1

What this repo does

This repository provides a Clarus v0.8 clinical quad dataset for detecting and reasoning about heart failure regime transitions.

The dataset models situations where a patient state is no longer contained within a single heart-failure basin but is shifting between competing regimes such as:

congestion-dominant decompensation
low-output failure
cardiogenic shock transition

This is the conceptual upgrade introduced in Clarus v0.8.

Earlier ladder versions detect instability, forecast deterioration, estimate collapse boundaries, model recovery geometry, and reason about intervention.

v0.8 introduces regime transition geometry.

The system now measures not only distance to the nearest failure boundary but also distance to the nearest competing regime boundary.

This allows models to detect:

regime switching
competing failure modes
unstable regime identity
transition-aware intervention reasoning

Core quad

The four core variables in this dataset are:

cardiac_load
reserve_capacity
perfusion_lag
organ_coupling

Operational definitions:

cardiac_load

Hemodynamic and metabolic burden placed on the failing heart.

reserve_capacity

Remaining physiologic tolerance and adaptive reserve.

perfusion_lag

Delay or impairment in effective tissue perfusion.

organ_coupling

Degree to which cardiac deterioration is propagating into multi-organ dysfunction.

Clinical variable mapping

Quad Variable	Clinical Measurements	Typical Indicators
cardiac_load	Cardiac index, BNP, filling pressures	CI < 2.2, BNP > 500, PCWP > 18
reserve_capacity	Ejection fraction, functional reserve, lactate clearance	EF < 35%, rising lactate
perfusion_lag	Time to MAP response, urine output delay	UO < 0.5 ml/kg/hr
organ_coupling	Creatinine rise, hepatic congestion, SOFA score	Rising creatinine, SOFA > 6

These mappings are illustrative rather than prescriptive.

The dataset encodes structural system state, not a specific clinical protocol.

Why the second boundary matters

Most instability detectors estimate only one quantity:

distance to the nearest collapse boundary.

This is insufficient for real clinical deterioration.

Heart-failure progression often involves movement between competing basins.

A patient may appear to be worsening within a congestion regime while the deeper reality is a transition toward cardiogenic shock.

v0.8 introduces:

secondary_boundary_distance

The model now measures:

distance to the current regime boundary
distance to the competing regime boundary

This enables detection of regime competition.

When the two distances converge, regime identity becomes unstable.

The derived signal

boundary_competition_ratio

secondary_boundary_distance / boundary_distance

acts as a marker of this instability.

Values approaching 1 indicate a system positioned between regimes.

Structural Note

This dataset is part of the Clarus ladder.

The ladder reconstructs instability geometry step by step:

v0.1 cascade detection
v0.2 trajectory awareness
v0.3 cascade forecasting
v0.4 boundary discovery
v0.5 recovery geometry
v0.6 intervention reasoning
v0.7 uncertainty-aware intervention
v0.8 regime transition geometry

Each rung increases resolution of the system state space.

v0.8 is the first rung where competition between instability basins is explicitly modeled.

That upgrade transforms Clarus from a collapse detector into a phase-transition instrument for complex systems.

Example regime competition row

A typical v0.8 regime transition scenario looks like this.

Field	Example
cardiac_load	0.88
reserve_capacity	0.22
perfusion_lag	0.79
organ_coupling	0.83
boundary_distance	0.11
secondary_boundary_distance	0.13
boundary_competition_ratio	1.18
regime_transition_score	0.74
transition_direction	congestion_to_cardiogenic_shock
regime_separation_margin	0.08
transition_uncertainty	0.23
transition_velocity	0.35
collapse_trigger	hypotension_spiral

Interpretation:

The patient is close to the current regime boundary while a competing regime boundary is also nearby.

The system is transitioning from congestion-dominant failure toward cardiogenic shock.

Row structure

Each dataset row contains layered signals.

Core state

cardiac_load
reserve_capacity
perfusion_lag
organ_coupling

Trajectory signals

drift_gradient
drift_velocity
drift_acceleration

Boundary geometry

boundary_distance
secondary_boundary_distance
boundary_competition_ratio

Uncertainty layer

boundary_uncertainty
trajectory_uncertainty
regime_confidence

Regime transition layer

regime_transition_score
transition_direction
regime_separation_margin
transition_uncertainty
transition_velocity

Perturbation and collapse markers

perturbation_radius
collapse_trigger

Recovery geometry

recovery_distance
recovery_gradient
return_feasibility

Intervention vector

delta_cardiac_load
delta_reserve_capacity
delta_perfusion_lag
delta_organ_coupling

Outcome fields

trajectory_shift
minimal_intervention_path
stabilization_success
label_heart_failure_transition (train only)

tester.csv excludes:

stabilization_success
label_heart_failure_transition

Files

data/train.csv

Training rows including outcome labels.

data/tester.csv

Evaluation rows with labels removed.

scorer.py

Reference v0.8 evaluation script.

Returns:

Classification metrics

accuracy
precision
recall
f1
confusion matrix

Regime transition diagnostics

recall_regime_transition_detection
false_stable_regime_rate
transition_direction_accuracy
high_uncertainty_transition_miss_rate
transition_detection_accuracy
false_transition_rate
missed_transition_rate
boundary_competition_error
misidentified_primary_regime_rate

Additional information

support counts
threshold transparency
dynamic label column discovery

Evaluation

Primary metrics

recall_regime_transition_detection

false_stable_regime_rate

Transition diagnostics

transition_direction_accuracy
high_uncertainty_transition_miss_rate
missed_transition_rate
boundary_competition_error
misidentified_primary_regime_rate

Additional reference diagnostics

transition_detection_accuracy
false_transition_rate

The goal of evaluation is not only to test whether a model predicts the correct label.

It tests whether the model correctly detects regime transitions and their geometry.

Note on transition_velocity

transition_velocity represents the rate at which the system moves toward the competing basin.

The variable is included in the dataset because it can help models reason about escalation speed.

The reference scorer does not yet evaluate this field directly.

Future versions may introduce velocity-aware diagnostics.

Production deployment

This dataset class functions best as a transition-aware monitoring layer rather than a standalone classifier.

Typical deployments include:

ICU deterioration monitoring

heart failure escalation systems

simulation benchmarks for intervention timing

digital twin modeling of clinical deterioration

AI model auditing for failure-mode recognition

The value lies in representing where the patient sits between competing basins.

Enterprise and research collaboration

This dataset class is suitable for collaboration with:

hospital analytics teams

critical care research groups

heart failure pathway teams

clinical AI safety researchers

digital twin developers

early-warning and escalation platforms

The framework provides explicit representation of:

which regime a patient is leaving

which regime is emerging

how uncertain the transition is

whether stabilization remains structurally feasible

License

MIT

Downloads last month: 18

Number of rows:

Total file size:

21.4 kB