A comparison of random survival forest and Cox regression for prediction of mortality in patients with hemorrhagic stroke

Table 1 Multivariate Cox regression analysis results

Characteristics	Standardized β	HR(95% CI)	p value
Marital status(ref:Divorced)
Married	-0.286	0.751 [0.502, 1.123]	0.164
Single	-0.409	0.664 [0.425, 1.038]	0.073
Widowed	0.005	1.005 [0.636, 1.587]	0.984
PVD(ref:no)	-0.478	0.620 [0.418, 0.920]	0.017
Severe liver disease(ref:no)	0.832	2.299 [1.277, 4.138]	0.006
Diabetes without cc(ref:no)	-0.623	0.536 [0.389, 0.738]	< 0.001
Admission age	0.028	1.028 [1.019, 1.038]	< 0.001
Heart rate	0.014	1.014 [1.004, 1.023]	0.004
MBP	-0.010	0.990 [0.977, 1.003]	0.124
Temperature	0.239	1.270 [0.999, 1.615]	0.051
Platelets	-0.002	0.998 [0.996, 0.999]	0.011
WBC	0.022	1.022 [0.993, 1.053]	0.143
Anion gap	0.053	1.054 [1.006, 1.104]	0.027
Creatinine	0.310	1.364 [1.196, 1.554]	< 0.001
Glucose	0.010	1.010 [1.008, 1.013]	< 0.001
Sodium	0.053	1.054 [1.025, 1.085]	< 0.001
Potassium	0.249	1.282 [0.994, 1.655]	0.056
PT	0.047	1.048 [1.014, 1.083]	0.005
GCS	-0.027	0.975 [0.947, 1.003]	0.081
Weight	-0.010	0.990 [0.984, 0.997]	0.003

Notes: HR, hazard ratio; PVD, peripheral vascular disease; MBP, mean blood pressure; WBC, white blood cell; PT, prothrombin time; GCS, glasgow coma scale

ISSN: 1472-6947