Reducing transfusions in critically injured patients using a restricted-criteria order set.

Abstract

We sought to examine the effect on blood usage of a new electronic order set restricting transfusion orders to specific evidence-based criteria for each unit (U) of red blood cells (RBC), plasma, and platelets. Prospectively collected transfusion data for Trauma ICU patients were compared for the 12 months before (PRE) and 8 months after (POST) order set implementation. Criteria for RBC transfusion were 1 U only for hemoglobin <7 g/dL in stable patients or <8 g/dL with angina, myocardial infarction, or cardiogenic shock; 2 U for hemoglobin <5 g/dL; and multiple U in the presence of shock, hypotension, or bleeding. Restrictive ordering criteria were also applied to plasma and platelets. Massive transfusion patients were excluded. Differences in demographics and outcomes were assessed with Wilcoxon-Mann-Whitney test or Wilcoxon rank sum test for continuous variables, and χ test for categorical variables. The percentage of patients receiving transfusions over time was compared with trend tests. Severity of illness (SOI) was graded from 1 (minor) to 4 (extreme). Of 1,038 Trauma ICU patients (583 PRE, 455 POST), 228 (22%) were transfused. Median SOI [IQR] (4 [3-4] vs. 4 [3-4]) and mortality (24.3% vs. 22.5%, p = 0.757) were similar for PRE and POST transfused patients, respectively. The percentage of patients getting transfused decreased for all transfusions (25.4% vs. 17.6%, p = 0.003), RBC (19.9% vs. 11.2%, p < 0.001), and plasma (9.3% vs. 5.9%, p = 0.047) in PRE and POST, respectively. After adjusting for age, Injury Severity Score, admission hypotension, and other variables, there was a lower odds of receiving any transfusion (OR 0.67 [0.49-0.92], p = 0.015), and RBCs specifically (OR 0.60 [0.41-0.86], p = 0.006), in the POST period. The frequency of pre-transfusion hemoglobin ≥7 g/dL decreased by 27%, and mean direct costs of transfusion decreased by approximately 28%, after the intervention. A significant reduction in transfusions was achieved after introduction of an order set restricting orders to predefined evidence-based criteria. Therapeutic study, level III.