CEO SUMMARY: Two years ago, the rate of hemolysis in blood drawn in the Cleveland Clinic’s Emergency Department was about nine times higher than the ASCP recommended rate of 2%. With a two-year cooperative agreement and funding from the federal Centers for Disease Control and Prevention, the ED and clinical lab staff developed a quality improvement program to identify the causes of hemolysis and train staff to reduce those rates. Their efforts were successful in driving the rate down to the recommended 2% and sustaining it at that level.
IN HOSPITALS THROUGHOUT THE NATION, labs continually find themselves dealing with the ongoing problem of hemolyzed specimens. Not only can hemolyzed specimens have a significant negative effect on patient care, but each hemolyzed specimen increases the cost of care, an important issue when hospital finances are being squeezed.
This was true at the emergency department at the Cleveland Clinic Main Campus, because staff was mostly unaware and unconcerned about hemolysis in blood samples. As in any busy ED, the staff was concentrating on providing timely and appropriate care to a steady stream of patients, most of them seeking immediate attention.
However, a big change happened at the Cleveland Clinic after it was awarded a grant from the CDC to study the problem of hemolysis and develop ways to reduce the number of hemolyzed specimens that originated in the ED. Hemolyzed specimens averaged about 13% at the start of a two-year process improvement project. At the end of 24 months, that rate fell to 2%, thus generating better patient outcomes, significant cost savings, and improved patient satisfaction.
“At the national level, baseline hemolysis from ED samples is an unquantified concern,” stated Michael P. Phelan, MD, an emergency medicine physician and an associate professor at the Learner College of Medicine at Case Western Reserve University, who was one of the leaders of this study. “There has also been very little historical progress in addressing this problem.”
To address the lack of evidence, the Laboratory Medicine Best Practices Program of the federal Centers for Disease Control and Prevention gave the Cleveland Clinic and Phelan, as principal investigator, funding to study the issue in 2014 and 2015.
“The goal of our study was specifically to gather evidence and develop best practices in how to prevent hemolyzed lab specimens,” noted Phelan.
“In the literature, the hospital ED was identified as a major source of hemolyzed samples, and the leading cause of unsuitable specimens was hemolysis,” Phelan said. “What’s more, the hemolysis rate in the ED is significantly elevated compared with that of other departments in the typical hospital.
“The literature shows that the rate of hemolysis in the ED runs anywhere from 6.8% to 30%!” he added. “The American Society of Clinical Pathology says the baseline should be at about 2%.
“When we started this program, our rate in the ED for moderately hemolyzed results for one week was 18.5% of chem-lab tests,” noted Phelan. “This compared to a rate of 2.3% for all other draws at the main campus. We defined moderately hemolyzed as having a hemolysis index of greater than 80.
“Our main campus hospital is an acute care, tertiary referral center with a 60-bed ED that gets 70,000 visits per year,” noted Phelan. During a session at THE DARK REPORT’S Executive War College in New Orleans in April, Phelan and pathologist Edmunds Z. Reineks, MD, PhD, described the efforts of their collaborative program between the ED and the clinical laboratory staff at the Cleveland Clinic to identify the causes and costs of hemolysis and reduce the rate of this pre-analytic problem. Reineks was part of the project and is the Laboratory Director of the Automated Chemistry Laboratory at the main campus.
“In addition to having a much higher rate of moderately hemolyzed specimens, the ED at the main campus also had a rate of grossly hemolyzed specimens of 4.3% versus 0.8% for all other locations at the main cam- pus,” Phelan explained. “We defined grossly hemolyzed specimens as those having a hemolysis index of 300 or more.”
Clearly, the ED at the main campus had a significant issue with hemolysis that Phelan and Reineks addressed in a stepwise fashion. First, they quantified the problem and estimated the cost of the hemolyzed specimens that originated in the ED. Second, they assembled a team of clinicians to educate nurses and medics about ED hemolysis and some of the most effective methods to minimize hemolysis.
Third, they conducted a performance improvement project centered around “plan-do-study-act” (PDCA) cycles to improve their methods for collecting, handling, and transporting lab specimens.
Patient Satisfaction Issues
Reineks explained the problem from the lab’s perspective. “When blood samples are hemolyzed, there’s interference in some 39-odd different lab tests, most importantly to potassium results,” he said. “Unreliable lab tests are a particular concern for potassium results because falsely elevated potassium may indicate a life-threatening abnormality, and low potassium also is critical for the ED.
“Another concern is the delay in care from the time you recognize a problem with a hemolyzed sample to the time you get a final result back,” he added. “Imagine the time it takes to have a blood sample drawn, resulted, and realize that result is abnormal—then add another sample to be re-run that requires significant rework by both nursing in the ED and lab technologists.
“There is the additional problem of low patient satisfaction because a hemolyzed sample causes delays in care,” he said. “There is also a cost, which we estimated at about $73,000 for the Main Campus hospital or about $1.13 per patient. If each of our 10 hospitals has costs in that range, then the potential savings are more than $700,000.”
Step 1: Identify the Problem
“In our first year, we wanted to quantify and define the problem by gathering data and establishing a baseline hemolysis rate,” he said. “Then, in the second year, we planned to introduce the interventions designed to improve our hemolysis rates.”
After Phelan established the need for the program, Reineks described the role the laboratory staff played in defining and measuring hemolysis and how the lab could contribute to reducing the rate.
“The fundamental goal is to avoid hemolysis because the lab wants a quality specimen that will produce a good result, which, in turn, optimizes the care process for the patient,” observed Reineks. “So in that way our lab’s goal aligns well with that of the ED.”
As this project began, the team needed to gather data. “The first step in our improvement programs was to define a level of hemolysis in the specimens originating in the ED,” noted Reineks. “Therefore, each time the lab received a specimen with a hemolysis index of 80, we added a cautionary comment to the result. Anytime a specimen’s hemolysis index was above 300, the lab rejected the sample and didn’t report that result.
“With a hemolysis index of 300, it’s impossible to get a normal potassium on a sample,” he said. “That is why the lab rejects those samples.
“One important role for the laboratory was data collection and interpretation. At about this time we started using a business intelligence system called Altosoft,” Reineks said. “We use it because it downloads data from our Sunquest laboratory information system and reorganizes it so that we can search it, create dashboards, and export that data to spreadsheets—all without doing any coding.
A Need for High-Value Data
“When we got this system, it updated the data every six hours,” said Reineks. “Refinements to this software now deliver critical management data in real time. We generate dashboards and report that data in a few minutes to anyone who needs it. It’s been a great help in both managing the lab and in conducting research.”
Having good data allowed Reineks and Phelan to understand how hemolysis affected two aspects of throughput: lab test results and ED patients’ length of stay.
Using Smaller Tubes for Blood Draws Contributes To Reduction of Hemolysis Rates to 2% in ED
AS PART OF A TWO-YEAR COOPERATIVE AGREEMENT WITH THE CDC, a team at the Cleveland Clinic worked to quantify the rate of hemolyzed samples originating in the Emergency Department, then identified ways to reduce the rate of hemolyzed samples. Two physicians and one nurse took a leading role in this effort. One was Michael P. Phelan, MD, an emergency medicine physician. The other was Edmunds Z. Reineks, MD, PhD, a pathologist. Annmarie Kovach, MSN, RN Nurse Manager of ED Main Campus, was the leader from nursing.
Change in Tube Size
One factor that made a significant contribution to the reduction of hemolyzed samples from 13% down to 2% was the use of a rainbow draw protocol and adoption of a smaller tube size.
- At Cleveland Clinic Main Campus ED, the rainbow draw includes 2×6 ml lithium heparin tubes, along with other (smaller) tubes.
- Heyer and colleagues had suggestive data regarding use of small tubes, but the data were not definitive.
- Switched to drawing 3×2 mL lithium heparin tubes.
During the term of the study, the improvement team twice introduced the use of a smaller tube for sample collection. Each time, the ED’s measured rate of hemolyzed specimens, originally at 13%, fell to 2%. That 2% rate was sustained following the second introduction of smaller tubes.
“I can’t tell you how important it was to get data from the LIS because one of our most significant struggles was getting data out of our EHR,” Phelan said. “We wanted to know the effect on ED patient throughput from hemolyzed samples and we wanted to know how those samples affected time in the department.
“Data from the lab told us that any discharged patient who had a hemolyzed specimen spent an additional 49 minutes longer in the department,” he stated. “Any admitted patient who had a hemolyzed specimen was spending, on average, 23 minutes longer in the department.
“Thus, reducing hemolysis improves patient throughput because it decreases the need for re-work, which is a key factor for ED nurses,” continued Phelan. “When you reduce rework you also increase patient satisfaction because patients won’t have to wait around while we do the redraw and run more tests.
“Determining the cost of hemolysis can be difficult because most of the costs are due to staff labor and we did not have very good models,” he observed. “There are some increased costs for lab analytes, but the highest costs consist of labor for re-drawing samples and re-running them, both in the ED and in the lab.
“As explained earlier, our model showed that we could save approximately $73,000 at the main campus hospital if we could cut our hemolysis rate to about 2%,” Phelan added. “If one hospital can save about $70,000 per year and your health system has 10 or 12 hospitals, then the costs can add up quickly.”
Step 2: Identify Opportunities
With data now available, the next step was to assess existing workflow.“In doing our work on this project, we found that there was a wide variation of practice among those who draw blood and the training they receive,” explained Phelan. “Nurses and medics do our blood draws because we do not have phlebotomists in the ED. Some hospitals in our system have phlebotomists, but it is not at all typical.
“There was no standardized approach or road map for sample collection, which meant that our practices were done according to personal preferences,” he added. “You could ask anyone and the method of blood draw would be different because it would be influenced by training and job description. Also, we learned there were a lot of myths about blood draws.
“Another problem we identified was the lack of documentation about how blood samples were collected at our facility,” noted Phelan. “We thus built into our EHR a lab documentation module for our nurses and medics to better capture methods and try and correlate those methods with our hemolysis results.
Assembling the Team
“Once we had the data we needed, we put a team together that included the ED professional staff, front line nurses, and nursing leaders,” he said. “We also had data analysts, biostatisticians, and experts in continuous quality improvement as well as our vendor (BD) to help focus our attention on the data and to get front line workers’ input on possible solutions.
“Over time, we learned that the most important members of the team were those in nursing,” emphasized Phelan. “We explained to them that this effort had the potential to affect patients significantly, and that was an important selling point for them.
“Once we had the team in place, we educated them about the causes of hemolysis and identified possible interventions,” he said. “For example, use of a straight stick resulted in a significant reduction in hemolysis. But, it also would change the culture significantly since at our main campus ED we draw the majority of our labs through an IV. The reason we do this is to avoid a potential second stick if blood and an IV are required (which may not be apparent when the patient is triaged).
“Although there was limited and conflicting data, the team wanted to try using the smaller volume/vacuum tubes as a first improvement step and because it was simply a change in equipment,” he said.
Step 3: Implement and Learn
“So, to start, we didn’t settle the straight-stick question because the ED staff was concerned about the issue of two needle sticks if the patient later needed an IV,” recalled Phelan. “But it was agreed to use the smaller 2ml tubes with heparin. BD helped us make that switch to drawing blood into the 2ml tubes with heparin. The result was amazing: Immediately following the switch to smaller tubes, our hemolysis rate dropped to 2%!
“To accomplish this, we had to change every tube in the main campus ED,” he continued. “Of course, there were some problems associated with this change. The lab staff could not get some of the readings they needed, labeling and filling issues cropped up, and so after one week, we went back to using the larger tubes. At that point, the hemolysis rate shot right back up. (See chart above.)
“We next instituted two other projects,” stated Phelan. “For one, we asked a select group of medics to change their practice and increase use of straight sticks to obtain blood samples first. This change followed the literature and it showed a statistically significant reduction in hemolysis. However, this was not clinically significant. That may be due to our limited number of patients getting the direct stick, since the vast majority of samples were still obtained from an IV.
Other Process Improvements
“For the other change, we attempted ED-wide hemolysis online education for our nurses and medics,” he stated. “While we saw statistically significant improvement of at least 1% to 2%, we did not see clinically significant changes as we did with the small tubes.
“Having tried these other approaches, toward the end of the second year of our agreement with the CDC, we finally replaced the larger tubes with small tubes,” Phelan stated. “Once again, we saw a clinically significant reduction of hemolysis back to around 2%.
“So, what did we learn?” Phelan asked. “We learned that when we gave frontline staff the information and education about best practices, they were empowered to assimilate the knowledge and decide which best practices to implement.
“We learned that when the ED staff implemented the smaller volume/vacuum tubes, we saw a significant drop in hemolysis rates,” he continued. “There was some reduction in hemolysis that resulted from the increased number of straight sticks for obtaining blood samples, but nothing like the clinically significant hemolysis reduction we saw with the use of smaller volume/vacuum tubes.
“We also learned from our analysis that straight stick and IV antecubital locations reduced hemolysis significantly,” observed Phelan. “In addition we found that shorter tourniquet time and using a larger gauge needle were associated with lower hemolysis. We found no association between syringe vs. vacuum tubes and lower hemolysis rates.
Sticking Once or Twice?
“Now, getting everyone to follow recommendations, such as using straight sticks for ED samples consistently, is probably not feasible. Here’s why: You may want to use a straight stick for every patient but there are other factors to consider, especially in those cases where ED patients are difficult blood draws. They are hemodialysis patients, or transplant patients, or they are morbidly obese and do not want to be stuck twice,” he said.
Phelan further observed that, “We do what works for our culture, and that would be my recommendation for other hospitals as well.”
For his part, Reineks summed up by saying, “We also didn’t anticipate that trying to change practices in a busy environment with scarce resources would be such a challenge. Having said that, we worked through those challenges and produced meaningful results nonetheless.”
Contact Michael Phelan, MD, at PhelanM@CCF.org; Edmunds Reineks, MD, at ReinekE@CCF.org.