Hemovigilance Hero

The University of Texas MD Anderson Cancer Center, USA, is world renowned for groundbreaking research and innovative clinical care. Less well known is its hemovigilance unit, where dedicated specialists work tirelessly – at the bedside and behind the scenes – to ensure patient and donor safety in the center’s internal, self-operated blood bank.

Kimberly Klein, associate professor in Laboratory Medicine Administration at MD Anderson, is one such scientist whose groundbreaking work really needs shouting about. As leader of the hemovigilance unit, Klein is at the forefront of transforming transfusion safety. The unit’s approach combines cutting-edge technology and expert nursing care to enable real-time monitoring of transfusions. By integrating informatics, the team aims to predict and prevent transfusion reactions, improving patient outcomes.

Here, Klein tells us more about the impact of this important work.

Could you introduce the MD Anderson Cancer Center’s Hemovigilance Unit, including its approach and goals?
 

Our’s is the first real-time hemovigilance unit in the country. Using an approach invented by  my predecessor, James Kelly, we proactively screen vital sign changes in patients during their transfusion. Our nurses monitor each patient’s dashboard during the procedure, looking for any vital sign changes that could be reflective of a transfusion reaction. If they see signs, they deploy one of our advanced care providers to the bedside to perform the physical exam, diagnostic workup, and treatment.

We then address any questions that might come up during the assessment and treatment of these reactions. Finally, after reviewing the data from the diagnostic tests, we make the diagnosis and sign out the case.

One to three percent of all transfusions result in a transfusion reaction. We aim to proactively capture these reactions from the onset. In that way, we can stop the transfusion, mitigate the symptoms faster, and improve the overall patient care experience.

What is your role at the unit?
 

As medical director, I oversee continuous process assessment and improvements. For example, at one point we weren't capturing certain patients at risk because the probe we used to monitor their temperature was not communicating with our dashboard. I worked with our staff to ensure that our algorithms capture that information. 

I also oversee the physicians that rotate through our service to make sure any reactions they pick up are signed out in a timely manner.

What work have you been doing to address transfusion safety issues?
 

We are converting all of our documentation into digital patient notes that we're entering into EPIC, our electronic patient record system. EPIC contains discrete data fields that we can then cross reference using informatics. That allows us to answer important questions, such as, “Is a blood culture needed for a transfusion-associated circulatory overload (TACO) reaction as opposed to a febrile reaction?”

This approach leads to cost savings and quality improvements. For example, we could assess the frequency of febrile transfusion reactions and whether they occur more or less frequently with plasma-reduced products, platelets stored in a platelet additive solution (PAS), or platelets treated with a pathogen reduction technology (PRT). Then we could assess the difference product modifications make in the quality of care for our patients, and in preventing subsequent reactions down the line. 

We can also use the system to investigate a specific reaction, like a transfusion-related acute lung injury (TRALI), that might link back to our donor operations. If we notice a pattern with donations from a repetitive donor, we can prevent any subsequent donations in a more timely manner. 

All these little projects are adding up to improve the overall transfusion experience for the patients.

What has been the impact of this work on the unit’s procedures and patient outcomes?
 

Studying our transfusion reactions in real time allowed us to pick up an anomaly that occurred when we had to change from a time-based duration for blood transfusions to a rate-based duration. The volumetric pumps that we used at the bedside to push the blood through the tubing and into the patient would not automatically stop at four hours, which is the FDA-prescribed limit for transfusion duration. 

We found that this was happening in around 50 cases out of 8,000 transfusions per month. This prompted us to make some changes to the nurses’ workflow, which has reduced the occurrence of overrunning transfusions to around seven cases per month. We continue to work with nursing to further bridge the gap. 

Another change we made based on our research was to discontinue our use of random donor platelets. That was after we observed the transfusion reaction rates were higher with random donor platelets than with single donor platelets. 

We've also looked into some data on the effects of dividing our red blood cell products into two products for patients that have TACO reactions. We want to see if that has an impact on the overall incidence of TACO on our patient population. That’s an ongoing project.

Because we now have a digital footprint, we've been able to create other IT projects – such as our electronic maximal blood ordering schedule. From a list of all the procedures performed in the hospital, according to CPT code, it averages out how many blood transfusions occur during each type of surgery over the last year. This gives managers an idea of how many blood products to order preemptively in each surgical unit. It also gives the hemovigilance unit more oversight of the procedures on the surgery list each day, so we can better anticipate blood product demand.

How are you able to monitor the impact of this work on patient outcomes?
 

Right now, the process is very manual, requiring a lot of retrospective chart reviews, which can be time consuming. When we did our comparison of single versus random donor platelets on the rate of transfusion reactions, we had to review almost 1,600 charts to make our determinations. 

Now that we have access to discrete data fields, we will be able to use data mining to pull out that type of information in the future. Eventually, when we have enough data points, we might be able to use AI to predict transfusion reactions.

How are you spreading the word about this transformational work?
 

As well as publishing our research in the journals Transfusion and Laboratory Medicine, we’ve also presented several abstracts at conferences, both national and international. 

These presentations have resulted in a few inquiries from other institutions seeking advice on setting up similar systems. Some have asked if we plan to license the system, which might be a consideration for the future.

What are the future directions for your research?
 

The next big project is to study the impact of different dosages on TACO reactions. We also plan to investigate the effects of newer products such as PAS or PRT platelets on the overall incidence of transfusion reactions.