Marjon Borgert

178 Chapter 8 METHODS Design A before and after study was conducted from January to December 2014. We primarily assessed the e ect of the transfusion bundle on the percentage of appropriate transfusions. To objectively assess this e ect, we focused on the number of appropriate transfusions. Appropriate transfusions were de ned as transfusions for which the last recorded pre-transfusion Hb level was lower than the patients’ individual preset Hb threshold as registered in the electronic patient le by the ICU physicians. A secondary outcome was the likelihood of appropriate transfusions, controlling for other variables. We implemented the transfusion bundle fromMay to August 2014. This implementation period denotes the transition period. The post-implementation, from September to December 2014, refers to the period in which the intervention is considered fully implemented as intended. Setting The study was conducted in a 28-bedmixedmedical-surgical ICU of a university hospital in Amsterdam, the Netherlands. The ICU is a‘closed format’department for adult patients (≥ 18 years) with four units in which patients are under the direct care of the ICU team. The patient-nurse ratio is 1:1 or 1:2, depending on the patients’ severity of illness. Study population The transfusion bundle was applied by nurses to every eligible ICU patient who received at least one unit of RBCs. Transfusion in patients for whom therapeutic haemapheresis was indicated or patients who were massively bleeding were excluded due the urgency of the situation. The massive blood transfusion protocol was activated in case of the presentation of the following signs or symptoms: (i) rapid decrease in blood pressure (systolic < 90mmHg; and (ii) not responding to uid therapy; and (iii) existence of a high suspicion for bleeding. Furthermore, blood products other than RBCs were excluded. Transfusion bundle We have developed the transfusion bundle by using a prospective risk analysis, i.e. the Bow-Tie analysis model. 17 A multidisciplinary expert team of two ICU nurses, three intensivists, one haemovigilance o cer and one laboratory analyst from the blood bank joined the Bow-Tie analysis session. The results of the Bow-Tie analysis were used to identify the potential interventions to include in the transfusion care bundle. The expert team selected the nal set of ve key interventions through discussion until consensus was achieved. InTable 1, the interventions are shown. The interventions were