Development of a risk‐adjusted blood utilization metric

Abstract

Blood utilization has become an important outcome measure for surgical patients because of the recognized risks and costs associated with transfusion. However, comparisons of blood utilization between providers or institutions are difficult, because there is no standard method for risk adjustment when assessing surgical blood requirements. We examined whether accepted diagnosis-related group (DRG) case mix indexes can be used for this purpose. We retrospectively analyzed electronic medical record data from 37,403 surgical inpatients to assess the relationship between intraoperative blood component transfusion requirements and the case mix indexes: weighted Medicare severity DRG and weighted all-patient refined DRG. Thirty-one surgeons from the general surgery service were compared to determine correlations between blood component utilization and case mix index in both a risk unadjusted and an adjusted fashion. Case mix indexes and transfusion requirements were directly correlated for red blood cells (RBCs), plasma, and platelet (PLT) transfusions (p < 0.0001 for all three blood components, for both indexes). Surgeons with greater case mix index values had greater transfusion requirements, and adjustment for case mix index resulted in less variation among surgeons (p < 0.0001, p = 0.0003, and p < 0.0001 for unadjusted vs. adjusted utilization of RBCs, plasma, and PLTs, respectively). The standard DRG-based case mix indexes used to determine hospital reimbursement were strongly correlated with intraoperative transfusion requirements. We propose that these methods can be used as a risk-adjusted blood utilization metric for surgical patients.