James L Hill MD, MBA and Jennifer Dawson MSN, MBA, RN, CPN, NE-BC from University Hospitals worked to create and implement an approach to improve transfusion utilization through an algorithm in HemaLogiX™ dashboard. This algorithm uses clinical criteria to determine the appropriateness at the time of a blood transfusion. Inappropriate transfusions are not only costly, but they also put the patient at risk. This algorithm can help hospitals discover areas for improvement in their transfusion process and ultimately reduce costs and improve blood supply management.
Risk of Blood Transfusions
Inappropriate blood transfusions create both a financial risk and risk for the patient’s health. While reported transfusion reactions include fever, anaphylaxis, and transfusion related lung injury, there are also additional risks including blood borne infection, fluid overload and more. These reactions can occur days, or even weeks after the transfusion takes place. This can make it difficult to attribute these reactions to the transfusion.
Blood transfusions can also be costly, especially when a patient has a reaction from the transfusion. The cost of a transfusion starts with the acquisition cost. Then you need to include the cost for safe handling and test, and clinical costs of supplies, lab transfusion activity. If the patient has a reaction, the intended length of stay and treatment will cost you as well. These can add up to a total of more than $1500 for a transfusion to take place. That is why it is crucial for the benefits of a transfusion to outweigh the risks, and that is exactly what this algorithm in HemaLogiX™ does.
Building the Algorithm
The journey to create this algorithm began in 2016 with a Transfusion High Reliability Medicine Initiative. As part of this initiative, University hospitals examined their practices in relation to transfusions. This included anemia management, cell salvage protocol, massive transfusion protocol, and waste reduction. After examining a variety of areas, they determined that the biggest opportunity for them to make improvements was in the clinical decision making process.
This algorithm needed to be created in a way that could easily fit into this process for physicians. They considered how a physician may evaluate appropriate use of transfusions. After surveying physicians, they found that some of the biggest issues were that data used to make these decisions was put under the wrong doctor or the analysis was irrelevant due to changing patient conditions. With this in mind they set out to create a dashboard that provided data to the person who was responsible to make the decision of transfusion and identify appropriate use.
University Hospitals pulled a panel of thirty specialists together to discuss best practices and develop criteria for giving a patient a transfusion. This information was then codified into an algorithm that provided an appropriateness score. The goal when a physician is giving a transfusion is to have an appropriateness score of 85% or more. This appropriateness data lives on the HemaLogiX™ dashboard with other lab datas. This allows accurate data to be pulled from the time of transfusion and properly attributes it to the correct physician.
This program was first utilized at University Hospitals Cleveland Medical Center. When looking at red blood cell utilization and platelet utilization from 2017, their reference year, to 2020, they found that the implementation of this program led to an increase of appropriateness score and a decrease in utilization for both. In 2017, the average monthly red blood cell transfusions was over 2000 while the appropriateness score was 66%. By 2020, the average monthly transfusions was around 1800 and the appropriateness score was 86%. Similarly, average monthly platelet transfusions in 2017 was over 700 with an appropriateness score of 69%, while in 2020 average monthly platelet transfusions was at around 500 with an appropriateness score of 74%.
They found similar results at Parma Medical Center, a smaller institution, where appropriateness increased and blood supply management improved. They also found that on a smaller scale, it was more manageable to engage with physicians who were high utilizers on a personal level. This created an opportunity to make further improvements.
This algorithm shows great promise in helping physicians with the clinical decision making process of transfusions. When transfusions are appropriately utilized, both patients and providers benefit. Moving forward, they would like to develop this algorithm further to provide real time scores rather than view them after the infusions have already taken place. The hope is to further assist in decision making and increasing appropriateness.