Reduced Red Blood Cell Transfusion in Cardiothoracic Surgery after Implementation of a Novel Clinical Decision Support Tool

Abstract

Packed red blood cell (PRBC) transfusion can increase short- and long-term adverse outcomes and health care costs. We compared the transfusion practices in cardiothoracic surgery before and after implementation of a novel clinical decision support (CDS) tool. The transfusion CDS tool was implemented within computerized provider order entry of a multi-institutional urban hospital system in September 2012. Data were queried for 12 months pre-intervention and for another 12 months post-intervention to compare transfusion practices for all adult patients having isolated coronary artery bypass grafting (CABG) or isolated surgical aortic valve replacement (SAVR). The total number of patients undergoing either isolated CABG or isolated SAVR was 744 pre-intervention and 765 post-intervention (p= 0.84). There was no significant difference in age (64 ± 11.4 years vs 64.5 ± 11.2 years, p= 0.37) or sex (30.2% vs 32.2% female, p= 0.42) between the 2 groups. The number of postoperative transfusions (374 [50.3%] vs 312 [40.8%], p < 0.001), postoperative PRBC units given (1.59 ± 2.9 vs 1.25 ± 2.5, p= 0.01), pre-transfusion hemoglobin level (8.09 ± 1.5 g/dL vs 7.65 ± 1.4 g/dL, p < 0.001), and incidence of surgical site infection (3.1% vs 1.1%; p= 0.005) were significantly reduced after implementation of the transfusion CDS tool. There were no significant differences in intraoperative transfusions (206 [27.7%] vs 180 [23.5%], p= 0.06), intraoperative PRBC units given (0.73 ± 1.5 vs 0.65 ± 1.4, p= 0.28), ICU length of stay (3.29 ± 3.9 days vs 3.37± 4.8 days, p= 0.74), or in-hospital mortality (1.3% vs 1.4%, p= 0.87). Implementation of a transfusion CDS tool was associated with lower pre-transfusion hemoglobin levels, fewer transfusions, decreased infection rates, and decreased health care costs, without an increase in short-term mortality.