Statistical analysis
First, the inter-group comparisons were analysed to confirm the validity
of the simulation-based training at Kirkpatrick level 3. Second, the
relationship between programme attendance and massive transfusion was
analysed to confirm the validity of the programme at Kirkpatrick level
4. Receiver operating characteristic (ROC) curve analysis of massive
blood transfusion and the amount of blood loss at the referring
hospitals was performed to determine the cut-off level of massive blood
transfusion and to analyse the relationship between massive blood
transfusion and programme attendance. Inter-group comparisons were
performed using the Student t-test and nonparametric test for continuous
variables (maternal age, body mass index at delivery, parity, SI, amount
of blood loss, time from delivery to maternal transfer, body
temperature, and Hb level on arrival) or the Fisher’s exact test for
nominal variables (mode of conception, mode of delivery, oxygen
administration, existence of an SI record, use of more than two IV
routes with lines ≥20 gauge, unconsciousness, massive transfusion,
hysterectomy, IVR use, and maternal death). Relationships between
massive blood transfusion and other variables were analysed using
logistic regression analysis. A p-value of <0.05 was
considered statistically significant. All statistical analyses were
performed using EZR (version 1.54, Saitama Medical
Center).22