Marjon Borgert

166 Chapter 7 Table 4. Univariate logistic regression covariates for transfusion bundle compliance Compliance Covariates OR (95% CI) P- value a Type of A&F (Team A and Team B) 4.19 (3.01 to 5.82) < 0.001 Time of intervention (impl. and post-impl.) 0.60 (0.44 to .82) 0.002 Nurses’ age 0.99 (0.98 to 1.01) 0.59 Nurses’ gender 1.38 (0.98 to 1.96) 0.07 Nurses’ years of work experience 0.99 (0.98 to 1.01) 0.54 Response to educational questionnaire 2.17 (1.48 to 3.22) < 0.001 Patient: Apache IV score 0.99 (0.99 to 0.99) 0.03 Patient: died in ICU 1.21 (0.87 to 1.69) 0.26 a signi cant when P -value is ≤ 0.20. DISCUSSION Our implementation study has shown that during the active period of implementation the combination of monthly A&F on team level with timely individual A&F, signi cantly improves short-term bundle implementation, compared to monthly A&F on team level alone. This resulted in signi cantly higher compliance levels in Team B. Even though signi cantly more transfusions were given in the team that received monthly A&F, our results also indicates that when using the combined A&F strategy nurses are four times more likely to be compliant to the bundle than whenmonthly teamA&F was used alone. When we stopped the A&F intervention during the post-implementation period, compliance dropped in both teams. However, even though in both teams compliance reduced in the four months after implementation, compliance levels were still signi cantly better in in the team that was exposed to the combined A&F strategy. These ndings are consistent with Zajonc. 19 He showed that individual knowledge about team performance combined with knowledge on performance on an individual level enhances team performance. 19 We have shown low compliance levels in the team where monthly A&F was given. This is in contrast to Lawrence and Fulbrook 25 who implemented the ventilator bundle. 25 They reported compliance levels of 68% by using monthly A&F. However, they provided A&F over a longer period of time, i.e. six months. The di erence in our compliance data between the nursing teams could be explained by the time span in which feedback was delivered as well as the level on which the data was aggregated, i.e. team or individual