To the Editor:
Frank et al.1 describe development of a maximum surgical blood ordering schedule. We are pleased that they used our findings2 regarding choice of whether to perform Type and Screen preoperatively. From the authors:1
“Using previously proposed criteria, we developed an algorithm … to determine the appropriate preoperative blood order for each procedure category. These criteria included: 5% or more of patients transfused with erythrocytes2 ; median estimated blood loss (EBL) more than 50 ml2 ; and a transfusion index 0.3 or more.”
Although the reliability and validity of the first two of the criteria refer to our article,1,2 our table 1 summary of our Results was different:
“Select a threshold for ‘minimal EBL’ (e.g., 50 ml) by using the smallest median EBL with many scheduled procedures and cases for which the lower 95% confidence limit for the incidence of erythrocyte transfusion was more than 5.0% … For each of the scheduled procedures with median EBL … less than [this] threshold … calculate the lower 95% confidence limit for the incidence of transfusion … For each of the scheduled procedures for which the calculated value … is less than 5.0% and for which there are 19 or more cases, set the MSBOS to indicate no type and screen.”
Thus, the value of 50 ml was to be determined statistically for each hospital; our criterion was less than 50 ml not larger than 50 ml; and we did not use 5% but the lower confidence limit of 5%. The criterion of less than 50 ml versus larger than 50 ml had a substantive effect at our studied hospital because the EBL often were reported using rounded values (e.g., not 49 ml but 50 ml).2 If the authors1 apply the criteria that we published, are any of their hospital’s maximum surgical blood ordering schedule recommendations changed?