During the past five years, randomized, controlled trials and clinical guidelines provided data for making evidence-based decisions about blood transfusion. In light of this new information, we monitored transfusion practices/variations among hospitals in our region. We sought to determine whether hospitals actively involved in addressing transfusion utilization and best practices in 1999 (hospitals 1–3) experienced changes when re-surveyed in 2003. As a “control” group, we determined transfusion patterns at 9 hospitals in 2003 (hospitals 4–12), that did not participate in the 1999 survey. We compared autologous transfusion practices involving orthopedic surgery patients at 3 hospitals participating in both surveys and at 8 hospitals contrasting 2003 survey data and 1999 hospital data (no data was available for comparison from one hospital). Data comparisons involved transfusions between Oct.–Dec.1999 vs. Oct–Dec. 2003. In addition we compared transfusion practices involving cardiovascular surgery patients (DRG’s 104–107, 109–111) at 2 hospitals participating in the 1999 and 2003 surveys and 5 hospitals joining the survey in 2003. Among the hospitals participating in 1999, the number of patients donating autologous units for orthopedic surgery declined 68%(88 to 28 patients), the number of patients transfused decreased by 58% (55 to 23patients), and the number of units collected was 68% lower (151 to 49 units). In contrast, at hospitals participating in the survey for the first time in 2003, there was only a 7% decline in the number of autologous units collected (274 to 255); the number of patients donating was unchanged (156 and 157). Of note, hospitals 10 and 12 were affiliated with hospital 3 and followed the same transfusion protocols. The average number of units transfused per patient in 1999 and 2003 in both groups (~1.5 and 1.8) was essentially unchanged. The relatively low number of units infused further questions the need for autologous donations in these patients.

At 2 hospitals participating in both the 1999 and 2003 surveys, the percentage of cardiovascular surgery patients transfused with red blood cells decreased (40% to 29% and 60% to 44%). This compares to a higher percentage of patients receiving transfusions (range 46–80%) at hospitals participating only in 2003. A similar difference appears in the average number of units infused per transfused patient (2.2–3.6 vs. 2.9–4.7). No differences in patient outcomes were attributed to changes in transfusion practices. Blood utilization patterns improved at hospitals and their affiliates committed to transfusion practice benchmarking compared to hospitals not utilizing this monitoring modality. We conclude that utilization variation analysis remains an effective methods for influencing transfusion practice.

Author notes

Corresponding author