Limited utility of algorithms predicting blood transfusions

Area of Science:

• Transfusion Medicine
• Surgical Outcomes Research
• Health Services Research

Background:

• Pre-operative autologous blood donation (PABD) aims to reduce wastage and unnecessary allogeneic transfusions.
• Published algorithms exist to predict transfusion needs, but their clinical utility requires analysis.

Purpose of the Study:

• To analyze the clinical utility of two published algorithms (A and B) in predicting transfusion needs in patients undergoing total hip arthroplasty.
• To assess the effectiveness of these algorithms in reducing transfusion wastage and unnecessary cross-matching.

Main Methods:

• A cohort of 195 patients undergoing total hip arthroplasty with PABD was studied.
• Two algorithms (A and B) were used to predict transfusion needs and compared against actual transfusions.
• Algorithm assumptions and formulas were modified to enhance prognostic accuracy.

Main Results:

• The optimal variation of algorithm A reduced allogeneic transfusions to 27.3% and wastage to 73.8%.
• Algorithm B's original version resulted in 78.8% transfusions and 46.2% wastage, improved to 69.7% with variation.
• Algorithm A demonstrated significantly better reduction in overall transfusion risk (P=0.001) but similar resource consumption (P=0.09) compared to algorithm B.

Conclusions:

• Algorithm A significantly improved prediction of transfusion risk but neither algorithm meaningfully identified patients benefiting from PABD or cross-matching.
• The algorithms did not increase PABD or allogeneic transfusion rates, nor did they reduce transfusion risk or resource consumption effectively.