A novel model for predicting the outcome of intracerebral hemorrhage: Based on 1186 Patients

Abstract

ObjectiveTo establish a model for predicting the outcome according to the clinical and computed tomography(CT) image data of patients with intracerebral hemorrhage(ICH). MethodsThe clinical and CT image data of the patients with ICH in Qinghai Provincial People's Hospital and Xuzhou Central Hospital were collected. The risk factors related to the poor outcome of the patients were determined by univariate and multivariate logistic regression analysis. To determine the effect of factors related to poor outcome, the nomogram model was made by software of R 3.5.2 and the support vector machine operation was completed by software of SPSS Modelor. ResultsA total of 8265 patients were collected and 1186 patients met the criteria of the study. Age, hospitalization days, blend sign, intraventricular extension, subarachnoid hemorrhage, midline shift, diabetes and baseline hematoma volume were independent predictors of poor outcome. Among these factors, baseline hematoma volume๥20ml (odds ratio:13.706, 95% confidence interval:9.070-20.709, p < 0.001) was the most significant factor for poor outcome, followed by the volume among 10ml-20ml (odds ratio:11.834, 95% confidence interval:7.909-17.707, p < 0.001). It was concluded that the highest percentage of weight in outcome was baseline hematoma volume (25.0%), followed by intraventricular hemorrhage (23.0%). ConclusionThis predictive model might accurately predict the outcome of patients with ICH. It might have a wide range of application prospects in clinical.