Clinical Outcome Prediction in Aneurysmal Subarachnoid Hemorrhage Using Bayesian Neural Networks with Fuzzy Logic Inferences

Table 1

Multiple linear regression demonstrates significant predictor variables for neurologic outcome in aneurysmal subarachnoid hemorrhage.

Independent variable	𝑃 value	𝛽 coefficients	Collinearity diagnostics
		(95% confidence interval)	(tolerance, VIF)
Normal motor response	<0.001	−0.329 (−0.496, −0.161)	(0.27, 3.71)
Cerebral infarction	<0.001	0.790 (0.695, 0.885)	(0.86, 1.16)
History of myocardial infarction	0.009	0.386 (0.097, 0.675)	(0.92, 1.09)
Cerebral edema	<0.001	0.322 (0.190, 0.453)	(0.96, 1.05)
History of diabetes mellitus	0.028	0.239 (0.026, 0.452)	(0.98, 1.03)
Day-8 fever	<0.001	0.231 (0.150, 0.311)	(0.93, 1.08)
Prior subarachnoid hemorrhage	0.004	0.197 (0.063, 0.332)	(0.98, 1.02)
Admission angiographic vasospasm	0.015	0.175 (0.035, 0.315)	(0.93, 1.08)
Neurological grade	<0.001	0.167 (0.093, 0.242)	(0.16, 6.43)
Intraventricular hemorrhage	0.001	0.142 (0.056, 0.229)	(0.80, 1.25)
Ruptured aneurysm size	0.001	0.130 (0.053, 0.206)	(0.97, 1.03)
History of hypertension	0.009	0.119 (0.030, 0.208)	(0.85, 1.18)
Vasospasm day	0.05	0.112 (0.001, 0.225)	(0.20, 5.11)
Age	<0.001	0.018 (0.015, 0.021)	(0.86, 1.17)
Mean arterial pressure	0.012	0.003 (0.001, 0.006)	(0.91, 1.10)

VIF: variance inflation factor.