Skip to main content
2024 AABB Annual Meeting Main banner
Icon Legend
This session is not in your favorites.
This session is in your favorites. Click again to remove it.
Presentation Icons
Live Stream Sessions
Live Stream Sessions

Abstract

Transfusion Service

Oral Abstract - Patient Blood Management

OA4-AM24-MN-08 - Red Blood Cell Transfusion During 814,564 Admissions to 24 hospitals: Wide Variability in Transfusion Practice Across Hospitals, Physicians, and Patient Diagnoses

Monday, October 21, 2024
8:15 AM - 9:15 AM  
Location: 372
CE: Zero
Background/Case Studies: Despite a strong evidence base to guide red cell transfusion for hospitalized patients, there remains substantial variability in clinical practice. In this study, we sought to characterize current transfusion practice in a multicenter study of 24 hospitals.

Study

Design/Methods: Retrospective analysis of patients admitted at hospitals captured in the GEMINI database between December 1, 2016 and June 30, 2022. GEMINI captures admissions to general internal medicine, subspecialty wards, and critical care areas, representing nearly 50% of admissions to these services across the province of Ontario (population of 14.6 million). Descriptive analyses examined variability in pre-transfusion hemoglobin across hospitals, diagnoses, subspecialties, and individual physicians. Additional analyses explored the prevalence of 2-unit-transfusions and change transfusion thresholds over the study period.

Results/Findings:

The available data captured 814,564 unique hospital visits at 24 hospitals and attended by 2,370 physicians over 7,475,447 inpatient days. Of these visits, 92,402 (11.2%) involved a red cell transfusion with a total of 294,684 red cell transfusions over the study period. Compared with non-transfused patients, transfused patients had a higher proportion of Charlson Comorbidity scores above 2 (34.3% vs 23.3%, p< 0.001), longer length-of-stay (18.4 vs 8.0, p< 0.001), and higher in-patient mortality (15.7% vs 6.4%, p< 0.001). Pre-transfusion hemoglobin was similar across hospital sites (median = 6.9 g/L, IQR = 6.5-7.4 g/dL) with median values ranging from 6.7 g/dL to 7.3 g/dL. The three most commonly transfused diagnoses were gastrointestinal bleeding, congestive heart failure, and coronary artery disease (Figure 1). Pre-transfusion hemoglobin thresholds were similar across hospitals with a decrease in average hemoglobin at 18 of 24 hospitals over the study period (mean change of -0.3 g/dL [range -1.0 to +0.3 g/dL]). There was substantial variability among hospitals in the prevalence of two-unit transfusions as a proportion of all transfusion orders ranging from 2.1% to 16.5%. Among individual physicians with at least 100 transfusions, median pre-transfusion hemoglobin was homogenous (median 6.9 g/dL IQR = 6.8-6.9 g/dL) and showed a rightward skew with 3% of physicians demonstrating a median pre-transfusion hemoglobin above 7.5 g/dL. Compared with male physicians, female physicians tended to transfuse at a lower hemoglobin (7.09 vs. 6.94 g/dL, p < 0.001).

Conclusions: This multicenter retrospective study found an overall decrease in pre-transfusion hemoglobin with substantial residual variability on quality indicators of red cell transfusion practice. These findings may provide target tracking for large-scale quality improvement campaigns (e.g. Choosing Wisely indicators), inform clinical trials of transfusion strategies, and facilitate audit-and-feedback interventions to improve transfusion practice.