Marjon Borgert

168 Chapter 7 Limitations Our study was conducted in a single hospital in a ‘closed-format’ ICU. This limits the external validity of our results. Although the compliance outcomes of one teamwere not shown to the other team, the Hawthorne e ect could have had in uenced our results. This would result in less di erences between groups and thus to an underestimation of the e ect of our intervention. In this study, we did not measured the quality of the transfusion bundle itself. However, even though evidence-based interventions are added to a care bundle, in theory, this could lead to unforeseen consequences. We used bundle checklists to track compliance as recommended by the IHI. 12 There could be a discrepancy between actual delivered care and the reported care.Thismay have given an underestimation of compliance levels. Bundle compliance was self-reported by nurses. We did not perform a double check of how well it was done. It might be possible that self-reporting leads to an overestimation of the results. This could especially be the case in the team that received individual A&F, since these nurses knew they would receive comments on their individual performances. Our results show a di erence in bundle compliance. Reasons for the di erences in compliance might be that barriers exist when changing professional behaviour, a ecting knowledge, attitude and behaviour. 10 We attempted to overcome the barrier of knowledge de cit by educating nurses. To create support, nurses were involved in the bundle design and in developing the educational questionnaire. Nonetheless, we did not attempt to determine nurses’ knowledge or their willingness to change behaviour. Other barriers could exist which we may not have taken into account, such as leadership. 30 However, nursingmanagement were requested not to stimulate implementation to minimize bias. Moreover, there were di erences in the number of nurses who responded to the educational questionnaire. Before nurses answered the questions, they received educational materials by email. Thus, nurses might be educated in the transfusion bundle without lling out the questionnaire. CONCLUSIONS Compared to monthly team A&F alone, providing timely individual A&F plus monthly A&F on team level signi cantly improves the success of implementing a transfusion bundle on the ICU during the active period of implementation, which is expressed in signi cantly better short-term compliance rates. Providing timely individual A&F plus monthly A&F on team level might also be e ective for the implementation of other evidence-based care bundles in healthcare. Future research could elaborate on longer duration of the intervention, the use of information and computer technology to lower costs of the intervention, and to enhance sustainability.