Rstudio ggplot2

Drug-specific initiation and switching prevalence were plotted for each calendar year. To assess determinants of the initial choice of P2Y12 inhibitor, we used multivariable logistic regression considering all baseline covariates and comorbidities. To assess determinants of deescalation, we used a Cox proportional hazards regression considering all baseline covariates and bleeding events (nonmajor or major) during follow-up as a time-varying covariate with a 30-day risk (ie, the influence of a bleeding event as a determinant was considered for 30 days after occurrence). Major bleeding risk was analyzed as a composite per ACR-HBR criteria in 1 model (as a binary variable), and individual ACR-HBR components were analyzed individually in another model. Each model also considered the full set of patient baseline characteristics. We provide descriptive statistics of these variables across the study groups. Since the likelihood of missing values was low (<1%), we used complete case analyses.
Between February and May 2022, all analyses were performed using SAS, version 9.4 (SAS Institute Inc) statistical software, and visualization was performed using RStudio ggplot2 (RStudio). A 2-sided P < .05 was set as the threshold of significance.

P values were calculated by using two-tailed Student’s t test, two-tailed Chi square test, two-tailed Mann-Whitney U test or two tailed Mann-Whitney Wilcoxon test as indicated. Box plots represent the distribution of spatial distances with medians or means as indicated (middle line), 25th and 75th percentile (box), 10th and 90th percentile (whiskers), and outliers (single points). Box and whisker plots were generated using GraphPad or RStudio ggplot2.