Premium
Hemorrhagic Shock Data Mining Project: Fuzzy Logic Model
Author(s) -
Ward John A,
Sondeen Jill,
Vela Ruben J,
Rivera Stanley C,
Convertino Victor A,
Holcomb John B
Publication year - 2006
Publication title -
the faseb journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.709
H-Index - 277
eISSN - 1530-6860
pISSN - 0892-6638
DOI - 10.1096/fasebj.20.5.a1382-c
We examined mortality as a function of systolic pressure (SysP), heart rate (HR) and shock index (SI = HR / SysP). Archived data from 34 conscious sedated swine that were hemorrhaged from a catheter in the aorta then volume resuscitated to a SysP = 80 mmHg were examined. The rate of blood loss was computer regulated to simulate a severe uncontrolled arterial hemorrhage, 53% of estimated blood volume removed in 5 minutes. Fourteen subjects failed between 0.5 and 23.1 hours (median = 1.3). Twenty survived for 24 hours. A fuzzy logic model was developed to discriminate failures from survivors. Partial membership in low, medium and high input sets was derived from descriptive statistics of SysP, HR and SI collected during the last 30 minutes of observation. Partial membership in triage output sets was converted to a fuzzy centroid value. The centroid discriminated failure from survival with a ROC curve area of 0.978 (95% CI: 0.960 to 0.996). A similar model for the first hour of observation had a ROC curve area of 0.637 (95% CI: 0.589 to 0.685). In its present form, the model is better for assessing the effectiveness of resuscitation than for performing triage within 1 hour of injury.