The purpose of this quality improvement (QI) initiative was to develop, implement, and evaluate the effectiveness of an evidence-based algorithm designed to reduce the number of redefine the standard of care to reduce the number of packed red blood cell (PRBC) transfusions in adult critical care patients while maintaining a stable hematological status and maintain current length of stay outcomes.
Description:
Priorities in the management of the adult critical care client are to ensure adequate ventilation and oxygenation and restore or maintain tissue perfusion at the cellular level. To meet this physiological requirement the hematological status of the client must be adequate to support cellular respiration needs. Transfusion is rarely indicated when the hemoglobin (hgb) is greater than 10 gm/dl and is typically indicated when it is less than 6 gm/dl. For intermediate hgb concentrations between 6-10 gm/dl evaluation of PRBC transfusion should include a relative risk assessment based on the client’s needs. Generally, hematocrit (hct) levels between 25-30% are adequate for optimal tissue oxygenation.
Design:
This QI project utilized a pre-post design. It included an educational component to assess physician/nurse knowledge of current transfusion protocols that currently existed at the initiation of this project. A transfusion algorithm was then developed and implemented that focused exclusively on the client with an intermediate hgb level between 6-10 gm/dl. The transfusion algorithm was developed and implemented to redefine the standard of care for PRBC transfusion for adult clients in the ICU setting.
Development of the Algorithm
Prior to the implementation of this QI project, current practice in our facility was based on the “10/30 rule,” where transfusion was considered routine when hgb levels were below 10 gm/dl and hct was less than 30%. The medical director of the critical care units, working with a CNS-led team of interprofessional clinicians, were recruited to review and revise the current transfusion algorithm using a systematic approach that provided evidence for the development and implementation of a PRBC transfusion protocol with guidelines utilizing various professional society guidelines published since 2012. A 90-day retrospective systematic review of 214 charts in persons with an intermediate hgb level finding and who had received at least one PRBC transfusion was conducted to examine pre-transfusion blood pressure, heart rate, respiratory rate, O2 saturation, level of orientation (if warranted), and diagnosis or procedure-related condition. A risk-based, stratified, decision outcome algorithm was designed in which both objective and subjective findings were included, as was the reason for hospitalization and procedural interventions related to potential hematological factors.
Implementation:
Use of the revised transfusion algorithm began the first day of the month following an educational intervention aimed at improving clinician knowledge related to the changes in PRBC transfusion in the revised transfusion protocol. The CNS Project Lead collected data beginning one month following implementation of the revised algorithm and documented the total PRBC transfused units based on the aforementioned parameters.
Data Analysis and Results:
The statistical software package SPSS (V25.0) was used for data analysis, with a statistical significance established at .05. Continuous variables were reported using mean and SD. The mean difference between pre-intervention and post-intervention transfusions was compared using a 2-sample independent t-test. Categorical variables were reported using frequency distribution and percentages. The sample was primarily male (78%), with a mean age of 68, admitted for either a surgical procedure, or who has undergone a surgical procedure and was admitted to the critical care units. Changes related to pre/post data revealed that clients with hgb levels between 6-7 g/dl were likely to receive a transfusion (p<.04), whereas persons with hgb levels between 7-9 g/dl were less likely to receive a transfusion (p< .02). Procedural variances were noted related to orthopedic and cardiac surgical procedures with those populations more likely to receive a transfusion (p<.02). When the decision to transfuse was compared to pre-post objective clinical findings, i.e., heart and respiratory rate and level of orientation, persons with an intermediate hgb were less likely to receive a transfusion, but not significantly (p<.09). When clients were categorized by DRG differences in length of stay (LOS) no significant change was found (p< .04), and the total number of pre-intervention transfusion units (n=253) compared to post-intervention (n=197) was significantly less (p=<.04).
Limitations:
Two limitations were identified in this project. Nursing staff turnover between the pre and post-project period was 17.6%. A second limitation was noted with regards to the attending physician background, with critical care hospitalist being less likely than surgeons to order a PRBC transfusion. A further issue was identified when anesthesia orders related to a specific procedure were compared with the revised algorithm and were grounded on hgb/hct measures versus physiological findings related to patient status.
Implications and Conclusions:
The evidence from this project demonstrates that a risk-based, stratified, decision outcome algorithm may reduce the number of transfusion events in the critical care setting, while not impacting patient status or LOS. By reducing the number of transfusion events the likelihood of a transfusion complication is decreased. The implications of this evidence for patient care guidelines and the education of healthcare providers demonstrates similar outcomes with fewer transfusion events than in the pre-intervention phase. Additionally, this project demonstrates the need to conduct additional translational research projects to reduce risk and achieve similar or improved outcomes related to transfusion practices in the critically ill adult population.