Machine learning techniques for mortality prediction in critical traumatic patients: anatomic and physiologic variables from the RETRAUCI study

Table 1 Risk factors associated with mortality. Description and attribute evaluation

Variable abbreviation	Type	Group	Description	Attribute evaluation (weight)
Age	N	Patient	Age in years	0.04767
Sex	C	Patient	Male / Female	0.00168
AHEAD	S	AIS	AIS scale for Traumatic brain injury (0–6)	0.08977
ANECK	S	AIS	AIS scale for neck injury (0–5)	0
AFACE	S	AIS	AIS scale for face injury (0–4)	0.00120
ATHORAX	S	AIS	AIS scale for thorax injury (0–6)	0.00949
AABDOM	S	AIS	AIS scale for abdomen injury (0–6)	0.00373
ASPINE	S	AIS	AIS scale for spine injury (0–6)	0.00380
AUPPEREXT	S	AIS	AIS scale for upper extremity injury (0–4)	0.00507
ALOWEREXT	S	AIS	AIS scale for lower extremity injury (0–5)	0.01186
AEXTERNAL	S	AIS	AIS scale for external and thermal injury (0–6)	0.00245
PointRF	S	T-RTS	Points of Respiratory Frequency (4–0)	0.02438
PointSBP	S	T-RTS	Points of Systolic Blood Pressure (4–0)	0.02855
PointGCS	S	T-RTS	Points of Glasgow Coma Score (4–0)	0.09633
MV	C	Status	Mechanical Ventilation (Yes/No)	0.05836
MASSIVEHEM	C	Status	Massive Haemorrhage (Yes/No)	0.01503
HEMODINAM	C	Failure	Hemodynamic failure (Yes/No)	0.04002
RESPIRATORY	C	Failure	Respiratory failure (Yes/No)	0.02438
AKIDNEY	C	Failure	Kidney failure (Yes/No)	0.02234
COAGULOP	C	Failure	Coagulopathy (Yes/No)	0.02234

N Numerical, C Categorical, S Scale, T-RTS Triage-Revised Trauma Score, AIS Abbreviated Injury Scale. Attribute evaluation (weight): Ranking of attribute with respect to Information Gain Attribute Evaluation method (in bold the most important)

ISSN: 1471-2288