“…Characteristics considered to be potential confounders for the association between use of intrauterine balloon tamponade or use of uterine artery embolization, or characteristics considered to be risk factors for the occurrence of the primary outcome measure alone, were included as covariates in the propensity score model. 17 Characteristics included as covariates that were available at the moment the clinician decided to use intrauterine balloon tamponade or perform uterine artery embolization were: maternal age, gestational age, parity (nulliparity or multiparity), preeclampsia, multiple pregnancy, prior cesarean birth, mode of birth (vaginal birth or cesarean section), cause of hemorrhage (categorized as uterine atony, retained placenta, abnormally invasive placenta and other causes [composite of placenta previa, placental abruption and uterine rupture due to small numbers]), the presence of coagulopathy (defined as a fibrinogen level ≤2 g/L during bleeding), symptoms of shock (defined as at least one measurement of a systolic blood pressure ≤90 mm Hg and/or heart rate ≥120 beats per minute during bleeding), volume of blood loss at time were selected a priori based on the literature 4,5,7,12,13,[18][19][20][21][22][23] and clinical reasoning. Missing variables were imputed using median and logically derived imputation (see Appendix S2 for the rationale behind the imputation method applied per missing variable).…”