2.5 Data analysis
Data were described using mean ± Standard Deviation (SD), frequency, its percentage and 95% Confidence Intervals (CIs). Demographic and clinical information of patients were compared based on herb-drug interactions by bivariable logistic regression models. Every variable with a P-value<0.2 in bivariable models was selected and entered into the multivariable logistic regression model using backward elimination. In every step of multivariable models, variables with great P-value were continuously removed from the model set until all remaining variables in the model were significant (P-value<0.05). All descriptive and analytical analyses were performed by using Stata software version 16.