Home Clinical Moving the needle on transfusions

Moving the needle on transfusions

Published in the March 2015 issue of Today’s Hospitalist

WHEN GUIDELINES BEGAN ENDORSING a much more restrictive approach to using blood products, physicians and hospitals took notice. Many launched educational campaigns to tackle a culture in which liberal transfusion thresholds were standard practice and patients were routinely “tanked up” around surgeries, procedures and discharge.

But few campaigns kicked off with as much panache as the one at Minnesota’s Allina Health in 2012. As part of a wide-ranging effort, medical staff members wrote and produced a DVD entitled “The Blood Police,” in which an officer issues a warning to a doctor and nurse about to administer two units of red blood cells to a stable patient with a hemoglobin of 8 g/dL.

Mixing a skit with slides containing the latest evidence and recommendations, “The Blood Police” was such a hit that a bigger cast of doctors and nurses returned for an encore production the next year, this time with a professional film crew. Their 2013 “CSI: Frozen Plasma Unit” had a much more elaborate plot “and it became required viewing for Allina’s 4,000-plus nurses. During National Patient Blood Management Awareness Week, which is the first week in November, “CSI” also played on a continuous loop in the doctors’ lounge.

“The videos allowed us to get across two simple messages that apply across all specialties: Don’t order two units in a nonbleeding patient when one will do, and don’t transfuse stable inpatients unless their hemoglobin drops below 7 or 8,” says Lauren Anthony, MD, a pathologist who heads Allina’s blood management program and who had a “CSI” cameo as a cleaning staff member who discovered a depleted plasma bag. The two videos were part of an extensive campaign that “enabled us to drop our red cell transfusion rates by 35%.”

Education is just one strategy that hospitals have rolled out to rein in transfusion rates. Many hospitals “including Allina’s “take a multipronged approach, following education up by hardwiring guidelines and alerts into CPOE systems to interact with clinicians ordering transfusions.

But there’s more that can be done, thanks to CPOE systems that now let hospitals identify specific outliers, those services and doctors who still adhere to transfusion thresholds of 9 or even 10. While most hospitals haven’t gone the next step to actually engage those physicians one-on-one, some health systems have found that giving physicians feedback about how their transfusion orders compare to their colleagues’ has been very effective.

Changing “transfusion culture”
At the 25-bed Gifford Medical Center in Randolph, Vt., hospitalist director Martin Johns, MD, says the medical staff there has had no problem implementing the new, more restrictive guidelines.

Surgeons at Gifford now routinely employ techniques to minimize blood loss during surgery. And with postop transfusions now viewed as a surgical complication, Dr. Johns points out that surgeons are willing to “try just about anything” to not transfuse.

As for postop care and blood use among medicine patients, the hospitalist group “with two doctors who together share one FTE slot and four full-time NP/PA hospitalists “controls “virtually 100%” of the transfusions ordered in the hospital.

“This is a major advantage of being in a small facility and not a big center: You can create a process based on changing guidelines and implement it quickly,” says Dr. Johns. The hospitalist group was able to overhaul its entire transfusion practice “by having our lab director talk at one group meeting.”

Larger community hospitals and academic centers, however, have many more clinical minds that need to be changed. And what Howard Corwin, MD, the director of critical care medicine at the University of Arkansas for Medical Sciences in Little Rock refers to as a hospital’s “transfusion culture” can vary from hospital to hospital.

“More senior physicians often tend to do what they’ve done for a long time,” says Dr. Corwin. “If the culture is a particular transfusion threshold, either consciously or subconsciously, that’s the way people practice.”

National recommendations, local efforts
At Allina, system leaders in 2012 and 2013 wanted to raise not only awareness about recently-issued guidelines, but launch a wide-ranging blood management program. While that involved a comprehensive effort, the leaders of the initiative stuck with this key tenet: You need clinicians to change other clinicians’ minds. That became the guiding principle of the health system’s educational efforts, even down to casting its two videos.

“One of my colleagues wanted to cast the CEO in a role, but I said, ‘No, that’s an administrator,’ ” says Dr. Anthony. “Our message is that blood management is good medicine, not some administrative thing to save money.” Instead, that cameo “playing a bartender in the second video “went to the system’s CMO.

Dr. Corwin authored a study that appeared in the December 2014 issue of the Journal of Hospital Medicine. That research looked at the cumulative effect of combining an educational campaign on appropriate transfusing with an in-house guideline and order set. “There’s an advantage to having internal people within a health system develop their own guidelines based on recommendations from professional societies,” he points out.

The Crozer-Keystone Health System in southeastern Pennsylvania likewise used new evidence and guidelines as an opportunity to “modernize our standards,” says Thomas Bader, MD, chief medical officer for four of the system’s five hospitals. Based on national recommendations, the system’s medical executive committee crafted its own, which were then disseminated through several grand rounds and onsite CME.

That local provenance certainly helped get out the word, Dr. Bader reports. But “just because you pass something at the medical executive committee doesn’t mean that it gets done.” What really helped turn Crozer-Keystone’s transfusion culture around were the cardiologists and trauma surgeons who championed the new transfusion thresholds.

“Unless someone had told me in advance that cardiologists and trauma surgeons are the folks you need to target to really make a difference, I wouldn’t have known to do that,” Dr. Bader points out. “But they ended up being the de facto leaders. When not even the cardiologists think a patient needs blood, that has a halo effect.”

The impact of best-practice alerts
At Stanford University Medical Center in Stanford, Calif., leaders also rolled out an educational campaign, then hardwired new guidelines into the center’s CPOE, with a best-practice alert (BPA) firing whenever clinicians order a transfusion outside of guideline recommendations. The alert also asks providers to enter a free-text reason for that order.

“The educational campaign was not nearly as effective as the BPA,” says hospitalist Lisa Shieh, MD, PhD, medical director of quality within the department of medicine. Education, she points out, helped cut the percentage of transfusions being ordered at the center for patients whose hemoglobin was greater than 8 from 57% down to 52%.

After the best-practice alert was introduced, however, that percentage fell below 30%. That was the case even though the best-practice alert resulted in clinicians aborting a specific order only 2% of the time, according to results of a study that Dr. Shieh coauthored in the January issue of the Journal of Hospital Medicine. In the rest of the orders where alerts fired, the ordering clinicians overrode the alerts.

“Clinical decision support obviously does not solve a problem completely,” says Dr. Shieh, “but I think if it’s designed correctly, it can be very high-impact.” As for why transfusion rates fell so sharply when an alert changed clinicians’ minds on the spot only rarely, “our study captured only those clinicians who decided to proceed with an order for a transfusion.” The data suggest, she adds, that BPAs yield an indirect effect over time.

As Dr. Shieh explains, once users have been interrupted by the BPA several times, “they will be discouraged from even trying to order inappropriate blood transfusions in the future. That’s why, over time, the best-practice alerts have been so effective.”

Mapping individual practice
One thing that Allina, Crozer-Keystone, University of Arkansas and Stanford all have in common: Their CPOE systems allow them to track how many transfusions (and at what hemoglobin levels) are being ordered by individual clinicians.

They know who the outliers are “or they would know if they chose to mine their own data, which most have so far decided not to do. Instead, they’re not giving individual doctors feedback and letting culture change take its course.

But at the University of California, San Francisco (UCSF), the hospitalist who is also principal data scientist for UCSF’s Center of Digital Health decided to feed that data back, using the center’s CPOE to map out the transfusion practices of both departments and individual physicians.

The resulting tree graph was updated every month between 2013 and 2014 while the academic center and the hospitalist group worked to change their practices around transfusions.

As Alvin Rajkomar, MD, who created the graphs, explains: “The size of the larger boxes is proportional to the number of transfusions ordered by the labeled department. The smaller boxes represent transfusion orders by individual physicians within that department. The size of each of those boxes is again proportional to the number of transfusions ordered by that physician.”

In addition, “the color of each box is directly proportional to the median hemoglobin threshold those attendings use,” Dr. Rajkomar notes. “Red boxes mean your patients are literally full of blood even as they are getting more.” Boxes that shade to gray or white, however, indicate transfusing behavior that generally falls within guidelines and adheres to restrictive hemoglobin levels.

Dr. Rajkomar says that he and the hospital medicine division fed back updated tree graphs monthly to the hospitalist group in 2013 and 2014. “These are meant,” he says, “to elicit a visceral reaction from physicians.”

Anonymous data
As for identifying which doctor owns which box, “We’re still grappling with how to best present this in the public forum and have purposefully kept these anonymized,” Dr. Rajkomar says. “When people ask for their individual data, we’re happy to give it to them. But when we’re presenting these graphs in a large group, we tend to not name names, at least not yet.”

Still, it’s clear from comparing an early tree graph to one generated a year later that the tool “along with all the discussions it generated “has been very effective. UCSF has since also started using best-practice alerts in its CPOE system when doctors try to order transfusions outside of the recommendations.

Given how successful UCSF’s initiative was in driving down transfusion rates, Dr. Rajkomar says he’s no longer generating this particular series of tree graphs.

He admits that creating such a graph initially entails “a fair amount” of programming, some of which was “nontraditional” and “more advanced.”

“In terms of the raw amount of time,” he says, “it took me about a weekend, but I’ve had some extensive training.” And hospitals certainly don’t need physicians to program such graphs. “As more hospitals adopt electronic health records, the availability of these data for any type of quality improvement will become ubiquitous.”

Future efforts
To further reduce transfusion rates, Dr. Rajkomar says that researchers at UCSF are studying the outcomes of patients treated by doctors who maintain more liberal transfusion thresholds. Stanford’s Dr. Shieh points out that when clinicians explained why they were ordering a transfusion outside the guidelines, 12% cited “symptomatic anemia.”

That may mean, she says, that “symptomatic anemia” should be added to the appropriate indications for transfusion, so doctors placing such an order wouldn’t warrant a best-practice alert. But a symptomatic anemia finding “can be very subjective,” Dr. Shieh adds. “We may look at that more closely and consider getting back to those services and individual physicians.”

For Dr. Corwin at the University of Arkansas, the next areas to target to further reduce transfusion rates are the system’s outpatient transfusion suites and the OR. “The other thing we’re going to look at,” he says, “is how preoperative anemia is identified and treated in elective surgeries.”

Tackling the issue of preoperative anemia is also the next frontier for blood management at Allina “and one that presents a lot of challenges.

For one, curbing transfusions among people found to be anemic during preoperative evaluations “just hasn’t risen to the top of the primary care to-do list,” says Dr. Anthony. “Patients aren’t really brought in soon enough before an elective surgery to really address anemia.”

Then there’s this problem, explains Joshua Martini, MD, an anesthesiologist who’s been engaged with the Allina blood-product initiatives since their inception (and is the star of both blood-related videos): Patients’ insurance “or lack of “determines where they are evaluated preoperatively. “How they’re referred affects where their preop evaluation comes from.”

Patients evaluated in some outpatient clinics routinely avoid unnecessary transfusions, he says. But patients evaluated in other primary care settings may not.

“It’s incredibly difficult to educate the primary care base,” says Dr. Martini, adding that he believes inappropriate transfusing around preoperative anemia will decrease, once preop evaluations can be standardized. “In the future, I believe we will realize the benefit of managing preoperative evaluations internally by perioperative physicians and physician extenders. A properly structured preoperative clinic can deliver standardized approaches to surgical optimization.”

In the meantime, “I suspect that we still have 10% or 20% of transfusions that we could safely do away with,” says Dr. Corwin of the University of Arkansas. “There’s still blood out there.”

Phyllis Maguire is Executive Editor of Today’s Hospitalist.

Blood management: making a splash

WHEN CLINICIANS at Minnesota’s Allina Health decided to launch a system-wide blood management program and educate their staff about new blood-product guidelines, they wanted to go big.

That meant hiring an outside consultant to help “and opting for what Lauren Anthony, MD, a pathologist who helped launch the initiative and blood management program, calls “an educational splash,” including high-profile presentations and videos.

“Having that outside influence and educational splash generated energy and interest,” says Dr. Anthony. “It also flushed out people who were already interested in reducing transfusions and already incorporating new guidelines in their practice.”

But the first step “what Dr. Anthony calls the “pre-splash” “was calling in a national MD-expert to talk to clinicians about blood management. “We first wanted somebody with no skin in the game just talking about what good medicine blood management is,” she explains.

For the next phase, the system did hire a consultant, Strategic Healthcare Group LLC, which is based in Indianapolis. Those consultants did a great job pointing out best practices and benchmarking “to show our physicians that we were over-transfusing,” Dr. Anthony explains.

What an outside consultant can’t do, however, is help health systems operationalize performance improvement. To do so, she notes, you need to know the political ins and outs of your own health system and listen to your doctors.

“We wanted the consultant to do a big presentation on benchmarks and guidelines, but one physician said, ‘You’re not going to get a lot of physicians to show up unless you schedule it two months in advance,’ ” says Dr. Anthony. While the consultant was disappointed with that timeline, “we didn’t have enough seats in the room” when the presentation finally took place.

But even with outside help and planning, the initiatives would not have succeeded without some significant muscle on the part of administration.

“Hospitals are fatigued with all these initiatives,” she points out. Meeting with staff in all the hospitals across Allina, Dr. Anthony says she kept hearing, “We already have a lot on our plate.” And when she wanted to set up a system council with representation from each hospital, a few facilities claimed they had no one on staff who wanted to serve.

“The chief medical officer made it clear that no one could opt out,” says Dr. Anthony. “Having the chief medical officer sign off on everything ensured that we got an audience at our meetings and time and resources carved out.”