Variation between hospitals in rates of reported transfusion reactions: is a high reporting rate an indicator of safer transfusion?

Abstract

Background and objectives: It has been suggested that the rate of reported transfusion reactions is positively correlated with safety of the transfusion chain in a hospital. We evaluated this assumption in the Transfusion Reactions in Patients Dutch National Hemovigilance Office database taking reported incorrect blood component transfused as a proxy for unsafe transfusion.

Methods: Reports from 2006 to 2010 and annual numbers of transfused blood components from the 103 hospitals were analysed. The rate of transfusion reactions per 1000 blood components was calculated per hospital. Logistic regression analysis was performed between reporting of at least one incorrect blood component and tertile of transfusion reaction rate.

Results: Out of the 103 hospitals, 101 had complete data in some and 93 in all 5years. In all, 72 had reported at least one incorrect blood component transfused; this was associated with blood use level and also with rate of reported transfusion reactions: odds ratio 4·2 (95% confidence interval, 1·3-13·7) in the highest vs. the lowest tertile after adjustment for blood use level.

Conclusion: Hospitals in the Netherlands which report more transfusion reactions per 1000 units are also more likely to have reported incorrect blood component transfused. The data do not support that hospitals with a higher rate of transfusion reaction reports are safer.