Red blood cell transfusion practices in two surgical intensive care units: a mixed methods assessment of barriers to evidence-based practice


  • DJM was supported by an institutional training grant from the NIH (T32 HL007534), PJP was supported by a K24 award from the NIH, and APG was supported by a grant from the Agency for Healthcare Research and Quality (K01HS018762).



Despite evidence supporting restrictive red blood cell (RBC) transfusion thresholds and the associated clinical practice guidelines, clinical practice has been slow to change in the intensive care unit (ICU). Our aim was to identify barriers to conservative transfusion practice adherence.

Study Design and Methods

A mixed-methods study involving observation of prescriber (i.e., physicians, physician assistants, nurse practitioners) and bedside nurse daily bedside rounds, provider survey, and medical record abstraction was conducted in one cardiac surgical ICU (CSICU) and one surgical ICU (SICU) in an academic hospital in Baltimore, Maryland.


Of 52 patient encounters observed during bedside rounds, 38 (73%) involved patients without evidence of active bleeding or cardiac ischemia. Surveys were completed by 52 (93%) of the 56 providers participating in rounds. Prescribers in the CSICU and SICU (87 and 90%, respectively) indicated the ideal pretransfusion hemoglobin (Hb) to be not more than 7 g/dL in nonbleeding and/or nonischemic patients compared to a minority of nurses (8% [p = 0.002] and 42% [p = 0.015], respectively). Prescribers and nurses in both ICUs overestimated the typical pretransfusion Hb in their units (CSICU, p < 0.001; SICU, p = 0.019). During rounds, providers infrequently explicitly discussed Hb monitoring or transfusion thresholds (33%) despite most (60%) reporting significant variation in transfusion thresholds between individual prescribers.


Our study identified several provider and system barriers to evidence-based transfusion practices including knowledge differences, overly optimistic estimates of current practice, and heterogeneous transfusion practice in each ICU. Further work is necessary to develop targeted interventions to improve evidence-based RBC transfusion practices.