Critical Care OptimizatIon of Albumin Ordering

BACKGROUND: Strong evidence suggests that human albumin solutions should not be used for fluid resuscitation except among patients undergoing therapeutic plasmapheresis and select patients with complications of liver cirrhosis (i.e. spontaneous bacterial peritonitis, or large volume ascitic fluid removal). Previous work by the investigators reported overuse of albumin fluid resuscitation as a stakeholder-identified quality improvement opportunity in Alberta ICUs. Audit data from the provincial electronic medical record (EMR) for critical care, eCritical, reinforced stakeholder opinion and demonstrated that within Alberta ICUs in 2016 a considerable amount of albumin use deviated from evidence-based indications. On average, 14% of admissions to adult ICUs received at least one albumin unit during their ICU admission, with several high-use units providing albumin to between 20% and 35% of admissions. In 2017 the investigators began a pilot quality improvement initiative to reduce albumin overuse in six ICUs in Alberta. The intervention was developed according to the Theoretical Domains Framework. Questionnaires completed by clinicians working in the target ICU identified local culture and opinion of clinical leaders as barriers, and education, order sets, and clinical champions as facilitators to reducing albumin use. The intervention thus consisted of establishing a clinical champion, educating clinicians, changing the process for albumin through development of an albumin order sheet, and providing quarterly unit-level audit/feedback data to clinicians on albumin utilization. There was a 41% reduction in albumin utilization, however follow-up data identified problems with sustainability. The current project will build on the pilot initiative to implement a quality improvement intervention to reduce albumin overuse within all adult ICUs in Alberta.

OBJECTIVE: Evaluate albumin reduction within all adult ICUs in Alberta.

METHODS: This will be a registry-based, stepped-wedge quality improvement initiative.

Quality Improvement Initiative: Spread to 16 adult ICUs in Alberta will require understanding and adapting the investigators pilot intervention to local barriers and facilitators. The intervention for this quality improvement initiative will be developed using the Theoretical Domains Framework. The research team will develop working groups responsible for tailoring the intervention and implementation strategy to each participating ICU. This will be conducted as part of the Province-wide quality improvement initiative. The investigators will conduct the quality improvement initiative using a registry-based, stepped-wedge, implementation trial. Owing to the nature of this design, the interventions will be implemented at the level of ICU. Clusters of 2 ICUs will be assigned to receive the intervention every month such that all ICUs in Alberta will receive the intervention by the end of the implementation period.

Registry-based data analysis: To evaluate this quality improvement initiative, the Provincial EMR for critical care, eCritical, will serve as a 'registry' and will be used to capture all clinical and outcome data. Data will be collected until patient death or hospital discharge and censored at 60 days. To optimize efficiency and power, data will be collected 3-months before and after implementation of the interventions.

SETTING: 16 adult ICUs in Alberta's critical care strategic clinical network reflecting teaching/non-teaching and urban/rural hospitals will participate.

STUDY POPULATION: All adult patients (>18 years) admitted to the study ICUs during the study period (November 2019 to January 2021). Study ICUs will include 14 general medical-surgical ICUs, and 2 cardiovascular surgery ICUs.

OUTCOMES: The primary outcome of this study will be the proportion of ICU patients without an evidence-based indication for albumin prescribed at least 1 unit of albumin (any concentration) during admission to ICU. Evidence-based indication will be operationally defined as receipt of therapeutic plasmapheresis OR having a diagnosis of liver cirrhosis and being in receipt of a paracentesis. This latter criterion enables identification of patients with either spontaneous bacterial peritonitis or large volume ascitic fluid removal. This outcome was selected because: 1) there is precedent within current literature; 2) previous experience auditing albumin utilization in Alberta ICUs demonstrate it is available through eCritical; 3) it provides a relevant, broadly understandable estimate of the effect of the implementation science intervention to change albumin overuse; and 4) it permits assessment of the feasibility of implementing the platform to develop and carry out multi-stakeholder implementation science interventions to promote evidence-informed clinical practices and evaluate the intervention.