Statistical analysis

We used the SPSS 22.0 software (IBM SPSS Statistics, Version 22.0. Armonk, NY: IBM Corp) to accomplish the following statistical analyses.

The continuous data was presented as the mean ± standard deviation (SD) and the median with interquartile rage (IQR) (25th–75th percentile). The categorical data was presented as the patient number with percentage.

In the univariable analysis, we utilized the Pearson’s chi-squared test with Yates correction or Fisher’s exact-test, as appropriate, to compare the categorical variables, and the Student’s t-test to compare the continuous variables. The effects of EIBL on the length of stay and chest tube duration were assessed by a Kaplan-Meier analysis using log-rank test.

The ROC analysis was conducted to evaluate the discriminative power of EIBL with regard to PCCs. The area under curve (AUC) with its 95% confidence interval (CI) was then extrapolated.

Finally, a multivariable binary logistic-regression model applying the hosmer-Lemeshow test for goodness-of-fit and the C-statistic for discrimination was established based on all clinicopathological variables with univariable P-value < 0.05 to identify the independent risk factors for PCCs. The odds ratio (OR) with 95% CI was then obtained.

The statistical significance would be revealed by P-value< 0.05 in both univariable and multivariable analyses [24].