Statistical analysis
The variables gathered from our maintained database were compared between patients with switching IS drugs and patients without switching IS drugs. Categorical variables were analyzed using a χ2 test and were expressed using numbers and proportion (%). Continuous variables were analyzed using a t-test or a Wilcoxon signed rank test. Continuous normal distribution was expressed as mean ± standard deviation and non-continuous normal distribution was expressed as median and interquartile (IQR). SPSS (IBM, version 26) and R (R Foundation for Statistics Computing) were used to perform the analyses.
Variables that were identified as statistically significantly (p< 0.05) were selected using the univariate logistic regression analysis, and were retained as candidate predictors for prediction modeling. After a stepwise selection process, risk factors were identified in the multivariate logistic analysis. Finally, a nomogram was constructed using these determined risk factors to predict the risk of poor curative effects of those recipients who receive an IS protocol based on TAC, and then may switch TAC to CsA. The established nomogram was further evaluated by using calibration curves. In addition, the discriminative performance of the nomogram was evaluated using the area under receiver operating characteristic (ROC) curve (AUC) and the clinical usefulness of the nomogram was assessed using decision curve analysis (DCA).