Datasets:

ClarusC64
/

clinical-trajectory-awareness-v0.1

scenario_id string	current_severity string	heart_rate_trend string	resp_rate_trend string	map_trend string	lactate_trend string	urine_output_trend string	oxygen_requirement_trend string	treatment_response string	label int64
train_001	low	stable	stable	stable	stable	stable	stable	improving	0
train_002	low	improving	stable	stable	falling	improving	stable	improving	0
train_003	moderate	improving	improving	stable	falling	improving	falling	improving	0
train_004	moderate	stable	improving	improving	falling	stable	falling	partial	0
train_005	high	improving	improving	stable	falling	improving	falling	improving	0
train_006	low	worsening	worsening	stable	rising	worsening	rising	poor	1
train_007	low	stable	worsening	worsening	rising	worsening	rising	poor	1
train_008	moderate	worsening	worsening	worsening	rising	worsening	rising	poor	1
train_009	moderate	stable	worsening	stable	rising	worsening	rising	none	1
train_010	high	worsening	worsening	worsening	rising	worsening	rising	none	1
train_011	high	improving	stable	stable	falling	stable	falling	partial	0
train_012	moderate	stable	stable	improving	falling	improving	stable	partial	0
train_013	low	worsening	stable	stable	rising	worsening	stable	poor	1
train_014	moderate	worsening	worsening	stable	rising	stable	rising	poor	1
train_015	high	stable	improving	stable	falling	improving	falling	improving	0
train_016	low	stable	worsening	worsening	rising	worsening	rising	none	1
train_017	moderate	improving	improving	improving	falling	stable	falling	improving	0
train_018	high	worsening	stable	worsening	rising	worsening	stable	poor	1
train_019	low	improving	stable	stable	falling	stable	stable	partial	0
train_020	moderate	worsening	worsening	worsening	rising	worsening	rising	none	1

What this dataset does

This dataset tests whether a model can distinguish current severity from future direction.

The task is not to identify which patient looks worse now.

The task is to identify whether the patient trajectory is moving toward stability or deterioration.

Core stability idea

A patient who looks severe may be improving.

A patient who looks mild may be deteriorating.

Trajectory awareness requires separating present-state severity from direction of movement.

Prediction target

The label column is binary.

Label 0 means stable or improving trajectory.

Label 1 means deteriorating trajectory.

Row structure

Each row contains:

scenario_id
current_severity
heart_rate_trend
resp_rate_trend
map_trend
lactate_trend
urine_output_trend
oxygen_requirement_trend
treatment_response
label

current_severity uses:

low
moderate
high

Trend values use:

improving
stable
worsening
rising
falling

treatment_response uses:

improving
partial
poor
none

Evaluation

Submissions must contain:

scenario_id,prediction
test_001,0
test_002,1
test_003,0

Run:

python scorer.py predictions.csv

Optional truth path:

python scorer.py predictions.csv data/test.csv

The scorer reports:

Accuracy
Precision
Recall
F1
Confusion matrix
Structural Note

This dataset tests trajectory awareness under clinical uncertainty.

It is designed to prevent shortcut logic based only on current severity.

The intended reasoning requires attention to direction, trend coherence, treatment response, and hidden deterioration risk.

The dataset does not expose the hidden rationale behind each label.

License

MIT

Downloads last month: 21

Number of rows:

Total file size:

10 kB