Marjon Borgert

181 Implementation of a transfusion bundle RESULTS Patient demographics A total of 386 patients were included in this study, 146 patients during baseline, 112 during implementation and 128 during the post-implementation period. Each patient received a median of 2 units of RBCs during their ICU admission (IQR=1-4). In total, 1128 units of RBCs were transfused. Table 2 shows that the cohorts were similar during the baseline, implementation and post-implementation period with respect to age, gender, typeof admission and ICUmortality. Most admissionsweremedical admissions, followed by cardio-surgical admissions. During the phase of active implementation, less units of RBCs were administered compared to the baseline and to the post-implementation period (di erence baseline 16%, 95% CI: 0.122 to 0.200, P -value < 0.001; di erence post- implementation: -9%, 95% CI: -0.122 to -0.045, P -value < 0.001). Table 2. Patients demographics Baseline period Implementa- tion period Post- implementation P -value Number of patients receiving ≥1 RBC transfusions during ICU admission* 146 112 128 0.03 Total number of transfused RBCs 466/1128 (41) 284/1128 (25) 378/1128 (34) <0.001 RBCs per patient, median IQR 2 (1-4) 2 (1-4) 2 (1-3) 0.49 Age in years, median (IQR 67 (54-75) 62 (52.8-72.0) 62 (52.5-74.0) 0.54 Gender (male), n/N (%)* 78/135 (58) 64/112 (57) 71/128 (56) 0.93 Type of admission* Medical case 88/135 (65) 63/102 (62) 76/106 (72) 0.30 Surgical case 9/135 (7) 11/102 (11) 9/106 (5) 0.53 Cardio-surgical case 38/135 (28) 28/102 (28) 21/106 (20) 0.29 APACHE IV, median (IQR)* 69 (52.5- 96.0) 61 (48.3-82.8) 74.5 (54.5-99.8) 0.005 ICU LOS (days), median (IQR)* 5.02 (2.6-10.1) 6 (3.0-12.0) 5.5 (2.0-14.0) 0.26 ICU mortality, n/N (%)* 36/135 (27) 26/112 (23) 28/128 (22) 0.64 Hospital mortality 40/130(31) 28/101 (28) 33/116 (29) 0.86 RBC= Red Blood Cells, ICU= Intensive Care Unit, APACHE= Acute Physiology and Chronic Health Evaluation; LOS=Length Of Stay; *including ICU readmissions; In some cases the denominator does not correspond fully with the overall number of patients receiving ≥ 1 RBC transfusions during ICU admission. This di erence is due to missing values on some variables.