*2.14. Statistical Analysis*

As this is an observational non-interventional study, there may be confounding factors between treatment allocation and outcomes in analyses of comparative response rates and samples may be included of patients that do not represent real-world clinical practices. Accordingly, all analyses in this study will be considered to be of an exploratory nature. Statistical analysis will be performed with SPSS v.22 software version 9.2 or later.

Population characteristics (including demographic characteristics, medical conditions, disease duration, types of treatment used at the start of the study and other variables collected in the data collection logbook) and all primary and secondary endpoints will be summarized by indicating the mean, standard deviation, minimum value, maximum value, median, 25th and 75th percentile and 95% confidence interval (CI) of the mean for continuous variables, and the absolute and relative frequencies, with a 95% CI of the proportion for categorical data. In the bivariate analysis, the chi-square test will be used to determine di fferences in proportions and Student's *t*-test for paired data before and after the administration of anti-TNF treatment.

The statistical analyses conducted throughout this study on bioinformatics data managemen<sup>t</sup> to determine significant di fferences between established groups, their corresponding graphs and the study of ROC curves (AUC values) will be performed using GraphPad Prism software version 8.2.0. STAMP (Statistical Analysis of Metagenomic Profiles), another bioinformatics tool based on Python, will also be used. STAMP was specifically designed for statistical processing and creation of plots from large amounts of biological data [38].
