A practice guideline and decision aid for blood transfusion

Abstract

An attempt was made to reduce exposure of patients to blood products by using a point-of-ordering decision support system and strict adherence to a practice guideline, by observing physician behavior in the multidiscipiinary intensive-care unit (ICU) of a tertiary-care medicai center. Hemoglobin (Hg) level at the time of transfusion, total units of red blood cells (RBCs) per admission, units per patient per ICU day. fraction of patients receiving no transfusions, and incidence of single-unit transfusions covering 628 patients were measured. In Phase 1, RBC transfusion behavior was observed without intervention. In Phase 2. a special order form for RBCs that suggested a transfusion threshold of 8.6 g/dl, of Hg was introduced. In Phase 3, the suggested threshold was lowered to 7.0 g/dL and required all transfusions that did not meet the n e w guideline to be prospectively reviewed by a transfusion medicine physician. The Hg level at transfusion fell from 8.5 g/dL to 8.2 g/dL (p = 0.008). The use of single-unit transfusions fell from 32 percent to 17 percent (p = 0.001), but there was no change in the number of patients receiving any blood, the total units per admission, or units per patient per clay. In this setting, a practice guideline with a point-of-decision support system did not influence blood usage. Intermediate outcomes (such as the level of anemia at transiiision or compliance to a guideline) should not substitute for more substantial outcomes (reduction in blood use or virus exposure) in assessing medical practices.