Spss 22.0 windows software

Kolmogorov-Smirnov test was used to confirm the normality of variables. Normal distribution variables were expressed as mean ± standard deviation whereas non-normal distribution variables were expressed as median (interquartile range). Categorical variables were expressed as numbers (percentage). We performed Student's t test and Mann-Whitney U test to compare the difference of normal distribution variables and non-normal distribution variables, respectively. And χ² test was used to compare the difference of categorical variables.
We performed multivariate logistic regression analyses to find risk factors of in-hospital mortality. Spearman rank correlation test was utilized to analyze the relationship between two variables. The predictive value of factors and models was evaluated by drawing receiver operating characteristic curves (ROC). Z test was performed to compare the predictive value of different factors and models.
A P value < .05 was considered to be statistically significant. SPSS 22.0 Windows software (SPSS, Inc, Chicago, IL) was used for all statistical analyses.

Kolmogorov–Smirnov tests were performed to confirm the normality of included variables. Non-normally distributed variables and normally distributed variables were shown in the form of median (interquartile range) and mean±standard deviation, respectively. And categorical variables were presented as numbers (percentage). We compared the difference of non-normally distributed variables and normally distributed variables between two subgroups by using Mann–Whitney U-test and Student’s t-test, respectively. Difference of categorical variables were analyzed by χ² test or Fisher test. Univariate logistic regression analysis was firstly utilized to explore the potential factors for mortality of included TBI patients. Then, potential significant factors in univariate regression were selected by conducting stepwise forward multivariate logistic regression. Significant risk factors in multivariate analysis were finally utilized to construct a prognostic model using logistic regression. We draw receiver operating characteristic (ROC) curves of single factors and the constructed prognostic model. Z-tests were conducted to compare the area under the ROC curve (AUC) difference of these factors and the model.
We considered that a two-sided p-value<0.05 was of statistical significance. SPSS 22.0 Windows software (SPSS, Inc, Chicago, IL) was used for all statistical analyses and figures drawing.