Skip to main content
. 2022 Aug 9;9:967341. doi: 10.3389/fcvm.2022.967341

Figure 3.

Figure 3

Comparison of procedural outcomes between software versions and over time. Procedure duration (A), fluoroscopy duration (B) and dose area product (C) were gaussian-transformed and analysis was performed via ANOVA. ANCOVA sensitivity analysis was additionally performed with adjustment for procedure date and whether PV-angiography was performed or not. Left censored variable usage of contrast dye (D) was analyzed via Tobit-regression. Significance is indicated for ANOVA post-hoc testing with Turkey-correction for procedure duration (A), fluoroscopy duration (B) and dose area product (C). For usage of contrast dye (D), significant predictors in the unadjusted Tobit-model are indicated.