Skip to main content
. 2020 Sep 18;15(9):e0239082. doi: 10.1371/journal.pone.0239082

Table 3. Model-based scoring system via multivariable logistic regression model for prediction of positive intracranial injuries from CT scan in patients with mild traumatic brain injuries.

Predictors OR 95%CI p-value ß Score
Presence of neurological deficits 18.3 2.6,126.3 0.003 2.9 29
Diffuse headache 5.8 4.0,8.2 <0.001 1.8 18
Signs of skull base fracture 4.8 2.8,8.5 <0.001 1.6 16
GCS <15 (at 2 hours) 4.4 2.5,7.8 <0.001 1.5 15
Dangerous mechanisms 4.1 1.7,10.0 <0.001 1.4 14
Stepping of the skull 3.8 1.8,7.9 <0.001 1.3 13
Significant wound at the scalp 2.7 1.9,4.0 <0.001 1.0 10
Vomiting >2 episodes 2.0 1.2,3.3 0.013 0.7 7
Current anticoagulant use 1.6 0.5,5.5 0.451 0.5 5
Posttraumatic amnesia 1.3 0.9,1.9 0.237 0.2 2
Age ≥60 years 1.3 0.8,2.0 0.357 0.2 2
Posttraumatic seizure 1.1 0.5,2.2 0.902 0.1 1
Loss of consciousness 1.1 0.7,1.5 0.912 0.1 1
Model Intercept 0.03 0.02,0.05 <0.001 -3.48
Total score 133
Mean±SE Min Max
Patients with positive CT findings 32.8±1.0 0 62
Patients with negative CT findings 13.2±0.4 0 79
Patients who required surgical intervention 39.9±2.1 7 79
Patients who did not require surgical intervention 16.1±0.4 0 69

Abbreviations: OR, odds ratio; CI, confidence interval; ß, beta-coefficient; GCS, Glasgow Coma Scale; SE, standard error; Min, minimum; Max, maximum; CT, computed tomography.