Spss statistical software v23

Discrimination (i.e., the degree to which a model differentiates between those who died and survived) was calculated with concordance (c-index or statistic), equivalent to the area under the ROC curve. A value of 0.5 indicates no predictive ability, 0.8 is considered good, while 1 is perfect discrimination. Hosmer and Lemeshow goodness of fit statistic and Nagelkerke r² were calculated for assessing overall model performance. To assess the calibration of the model, (i.e., agreement between predicted and observed risk of mortality), calibration plots were used. Perfect calibration is implied by a 45° diagonal line (calibration slope = 1 and a calibration intercept = 0). Deviations above or below the line reflects underprediction and overprediction by the model. We assessed the predictive accuracy of the score in the validation cohort with discrimination and calibration as mentioned above. We did all analysis with SPSS statistical software v23. Calibration plots were constructed Stata/IC v16 (trial version). The present study is reported in compliance with standard TRIPOD guidelines for prediction models (S1 TRIPOD Checklist).

Normally distributed data are expressed as mean and standard deviation, while non-normally distributed data are expressed as median and interquartile range. Pearson χ2 test was used to compare sex ratios at baseline, whereas the differences in patient body mass index (BMI), length of AR course, and visual analog scale (VAS) scores between the AR and AR-asthma groups were verified using the Mann-Whitney U test. The expression levels of critical DEGs measured using quantitative PCR were compared between the two groups using t-tests. Statistical significance was set at P < 0.05. All statistical analyses were performed using SPSS statistical software (v23; SPSS Inc., Chicago, IL, USA).

Continuous variables were expressed as median and interquartile range (IQR) and categorical variables as a percentage. t-test and Mann–Whitney testing were used to compare continuous variables and Fisher’s exact test was used for categorical data. All reported p-values were two-sided, and a p-value less than 0.05 was considered statistically significant. All statistical tests were two-sided, and a p value < 0.05 was considered statistically significant. Statistical analysis was performed using the SPSS v23 statistical software (IBM Crop., Armonk, NY, USA). A power analysis was performed to estimate the optimal sample size required in order to detect a statistically significant difference in vedolizumab trough levels between patients responding/not responding to the next intervention/undergoing a surgery or not. For an alpha error defined as 0.05, with a power of 80%, a minimal sample of N = 16 is required for each group to detect a minimal difference of 25% in trough levels.