Skip to main content

Table 4 Logistic Regression Binary model for mortality prediction

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

Variable B coefficient Standard error OR (95% CI) p-value
Age 0.051 0.002 1.05 (1.04–1.06) <  0.001
AHEAD
 0 Reference    
 1 0.091 0.281 1.10 (0.63–1.90) 0.746
 2 0.114 0.202 1.12 (0.75–1.66) 0.573
 3 0.150 0.147 1.16 (0.87–1.55) 0.307
 4 0.911 0.135 2.49 (1.91–3.24) <  0.001
 5 2.011 0.128 7.47 (5.82–9.60) <  0.001
 6 4.394 1.200 80.99 (7.72–849.01) <  0.001
ATHORAX
 0 Reference    
 1 −0.256 0.305 0.77 (0.43–1.41) 0.401
 2 −0.048 0.169 0.95 (0.68–1.33) 0.777
 3 −0.513 0.124 0.60 (0.47–0.76) <  0.001
 4 −0.122 0.140 0.88 (0.67–1.16) 0.381
 5 −0.036 0.195 0.96 (0.66–1.41) 0.855
 6 −1.263 1.199 0.28 (0.03–3.12) 0.303
AUPPEREXT
 0 Reference    
 1 0.107 0.306 1.11 (0.61–2.02) 0.727
 2 −0.315 0.121 0.73 (0.57–0.92) 0.009
 3 −0.544 0.295 0.58 (0.33–1.03) 0.065
 4 −1.17 1.082 0.31 (0.03–2.57) 0.278
ALOWEREXT
 0 Reference    
 1 −0.783 0.380 0.46 (0.22–0.96) 0.039
 2 −0.559 0.169 0.57 (0.41–0.79) 0.001
 3 −0.509 0.158 0.60 (0.44–0.82) 0.001
 4 −0.931 0.251 0.39 (0.24–0.64) <  0.001
 5 0.400 0.228 1.49 (0.95–2.33) 0.079
PointSBP
 4 Reference    
 3 0.316 0.152 1.37 (1.01–1.85) 0.038
 2 0.356 0.148 1.43 (1.07–1.91) 0.016
 1 0.815 0.375 2.26 (1.08–4.71) 0.030
 0 1.325 0.185 3.76 (2.62–5.40) <  0.001
PointGCS
 4 Reference    
 3 0.063 0.142 1.07 (0.81–1.41) 0.658
 2 0.392 0.137 1.48 (1.13–1.94) 0.004
 1 1.072 0.157 2.92 (2.15–3.97) <  0.001
 0 1.810 0.122 6.11 (4.81–7.76) <  0.001
MV 0.858 0.142 2.36 (1.93–2.89) <  0.001
MASSIVEHEM 0.554 0.153 1.74 (1.29–2.35) <  0.001
HEMODINAM 0.818 0.099 2.27 (1.86–2.75) <  0.001
RESPIRATORY 0.541 0.106 1.72 (1.39–2.11) <  0.001
AKIDNEY 0.641 0.098 1.89 (1.57–2.30) <  0.001
COAGULOP 0.610 0.109 1.84 (1.49–2.28) <  0.001
  1. RF Respiratory frequency, SBP Systolic blood pressure, GCS Glasgow coma score. MV Mechanical ventilation, OR Odds Ratio, CI Confidence interval