A novel measure to summarize blood transfusion practice during critical illness

Nicholas A. Bosch,Anica C. Law,Zhan Shi,S. Reza Jafarzadeh,Allan J. Walkey

DOI: https://doi.org/10.1111/trf.17814

2024-03-23

Transfusion

Abstract:Background Large‐scale observational studies have summarized transfusion practice using traditional measures of central tendency (e.g., the mean hemoglobin concentration at the time of transfusion). However, the mean hemoglobin concentration fails to identify specific hemoglobin concentration thresholds that drive practice. In the following brief report, we propose a novel measure of "practice discontinuity" that identifies specific practice‐defining hemoglobin thresholds. Study Design and Methods We used the PINC AI Database (2016–2022) to identify adult patients admitted to an intensive care unit with at least one hemoglobin concentration measurement. For each day that hemoglobin was measured, we identified whether the patient received a red blood cell transfusion using hospital charge codes. We defined the "practice discontinuity" measure as the hemoglobin concentration at which there was the largest increase in transfusion use going from a higher to an incrementally lower hemoglobin concentration. We also calculated the mean and median pretransfusion hemoglobin concentrations. Results We identified 1,298,367 patients and 4,905,839 patient‐days for inclusion. RBC transfusion occurred in a total of 530,654 (10.8%) patient‐days. The overall pre‐transfusion mean and median hemoglobin concentrations were 8.4 and 8.0 g/dL, respectively. The practice discontinuity measure identified 7.0 g/dL as the hemoglobin concentration at which transfusion use increased the most, from 46.6% of patient‐days at a concentration of 7.0 g/dL to 74.8% of patient‐days at a concentration of 6.9 g/dL. Discussion We propose that future studies of red blood cell transfusion practice consider inclusion of the practice discontinuity measure to more fully summarize clinical practice.

hematology