The battle for high-value care

Experts give their advice on value projects and the evidence on stress tests and clot prophylaxis.


Being a hospitalist who provides high-value care is almost as hard as being a Jedi, as Vineet Arora, MD, MAPP, MACP, explained to attendees at the Midwest Hospital Medicine Conference, held in Chicago in October.

Image by Getty Images
Image by Getty Images

“You are the ones on the front lines. You're the ones who are stewards of all these resources. . . . Balancing the pressures of the ‘evil CEO Vader’ from our ‘Star Wars' story and all of the fee-for-service incentives in our system while also promoting value—this can be a very tough challenge,” said Dr. Arora, a hospitalist and tenured professor of medicine at the University of Chicago.

To help hospitalists win this war, she offered her wisdom on improving value generally in hospital medicine (which involved a “Star Wars”-themed acronym), while another meeting speaker, Anthony C. Breu, MD, focused on two particular battlegrounds—stress tests and venous thromboembolism (VTE) prophylaxis.

A force for value

As a hospitalist, it's very difficult to know the costs of the tests and treatments you order for patients, Dr. Arora said, citing a study published in the Journal of Hospital Medicine in 2010 in which researchers made hospitalists guess the costs of aspects of inpatient care. “They said, ‘Do you know how much these things charge?’ The estimates were so far off that they actually [had] to graph this . . . in a [logarithmic] scale,” she reported.

Photo courtesy of Dr Arora
Photo courtesy of Dr. Arora

Combating the problem of overtreatment does not require knowledge of precise costs. It does, however, require a champion, Dr. Arora said. And not just anyone will do. The first tip in her acronym—Find the right champion—recognizes that the appropriate person will vary based on the specifics of the value-improvement project.

“You have to really think [about] who is the target for your work and who's going to be the adequate champion,” she said. For example, one project at her hospital was to reduce overuse of proton-pump inhibitor (PPI) infusions in patients with upper gastrointestinal bleeding (“‘Skip the Drips' Reduced PPI Overuse,” February 2018 ACP Hospitalist).

The project team found that the support of one particular gastroenterologist was required to get broader engagement with the effort. “A lot of the folks said, ‘Well, if this guy doesn't buy in—he's the fellowship director—none of us are going to do it,’” said Dr. Arora. He did, and the project was a success.

A well-placed champion is necessary, but not always sufficient, as shown by another story Dr. Arora shared. “Not everything is going to stick, and we can still learn from what doesn't stick,” she said.

Several years ago, she and her colleagues tried to push their hospital to switch from testing creatine kinase-MB (CK-MB) to troponin-only testing in patients with suspected acute myocardial infarction (MI). Despite having two subspecialty champions, the project faced major obstacles. One was the inclusion of CK-MB in a bundled cardiac panel of tests. “We also had anti-champions—people who did not believe in this—and they were like, ‘No, if you do this on my patients, I'm not going to play ball,’” said Dr. Arora.

This challenge led to her second recommendation: Observe and understand cultural norms. “Even if you have a champion, it's important to get a flavor of that culture. Is your organization ripe for this?” she said.

Broader data on the shift from CK-MB to troponin testing show that hospitals varied widely in their ripeness for this change after the ABIM Foundation's Choosing Wisely campaign recommended it in 2015. Dr. Arora and colleagues published an analysis of the rate of troponin-only testing at 91 teaching hospitals in the December 2017 Journal of Hospital Medicine.

“What's interesting is that you did see big jumps in certain hospitals that either went from the lowest tertile to the highest or to the middle,” she said. “There was something going on in those hospitals such that when the recommendation came out, their culture was ready to adopt.” Frontline clinicians and patients are some of the best judges of a hospital culture, Dr. Arora said. “It's not what your leaders say. It's what people do.”

In some cases, people may be fully supportive of your goal but lack the tools to achieve it. That's when you implement the next tip: Redesign the system. As an example, Dr. Arora talked about her initiative to reduce interruptions in inpatient sleep (described in the January Journal of Hospital Medicine and “Sleep Heals,” July ACP Hospitalist).

“We asked people . . . ‘What is it that right now is impairing you from doing this?’ And actually all of them said, ‘Epic. . . . I don't even know how to change the vital signs off of q4. All of the labs default to 4 a.m.,’” she said. In response, the project leaders changed the prompts and options in the electronic health record (EHR) and saw a significant improvement in the sleep-friendliness of orders.

Of course, before you can make such a fix, you have to notice the problem. That relates to Dr. Arora's penultimate tip: Cultivate situational awareness. To test trainees' awareness of low-value care practices in their environment, she and her colleagues created a fictional patient scenario called the horror room. “It's not as bad as it sounds. Basically, we embed a bunch of safety hazards and low-value care hazards into this room,” Dr. Arora said.

Almost everyone noticed poor hand hygiene, but very few noticed that the patient had an unnecessary Foley catheter. This same problem has been identified in research on real patients, Dr. Arora noted. “When we ask clinicians, attendings, residents, ‘Does your patient have a Foley?’ a lot of them say they don't. But then, when you go the bedside, you see [a catheter bag]. And so this phenomenon is dubbed immaculate catheterization by Dr. Sanjay Saint,” she said to laughs from the audience.

The effective solution to this was another EHR change, “literally telling you that your patient has a Foley on the patient list,” said Dr. Arora. “If we see that they have a Foley on there, we might ask why.” The hospitalist who led this improvement project also created a similar notification about the use of telemetry.

“This actually led to a major reduction in both catheter and telemetry use in our hospital. And just telemetry by itself is $53 a day,” Dr. Arora said. Additional benefits include reduced alarm fatigue from telemetry and urinary tract infections from catheters.

“So within a few short months, this pays for itself and is a very, very minor intervention that you can think about in terms of a nudge to actually get clinicians to do the right thing,” she said. That was the final tip in her acronym: Embed nudges to promote the right thing.

Dr. Arora's fellow “Star Wars” fans may have already noticed what her high-value advice spells. “I love acronyms, and so I hope you will use the FORCE,” she concluded.

Two things, no reasons

As an editor of the Journal of Hospital Medicine's “Things We Do For No Reason” section, Dr. Breu battles all kinds of unnecessary care. But he focused his conference talk on two particular interventions. The first was stress testing in patients with low-risk chest pain, which he defined as those with a nonischemic electrocardiogram (EKG) and negative serial troponin tests.

Photo courtesy of Dr Breu
Photo courtesy of Dr. Breu

There are a couple of reasons why hospitalists might think they should order a stress test for such patients, including recommendations from the American College of Cardiology and American Heart Association (ACC/AHA), noted Dr. Breu, who also a hospitalist in the VA Boston Healthcare System and an assistant professor of medicine at Harvard Medical School.

“As of right now—October 2019—it is still the guideline from the ACC/AHA to get a stress test within 72 hours. They say it's actually preferable within 24 hours or during the hospitalization, but if you're so cavalier as to discharge the patient without it, you've got to get it within 72 hours,” he said. Not surprisingly, then, the test is also the standard of care in many hospitals.

But both of these factors are likely to change, according to Dr. Breu. “The lead author for the guidelines published in 2010 has suggested they are likely going to update the ACC/AHA guidelines to remove this,” he said.

In the interim, he encouraged hospitalists to consider what they are trying to accomplish with a stress test. “For 10-plus years of ordering stress tests, I actually never even thought about what I was trying to do,” said Dr. Breu. “It's worth thinking for yourselves.”

One possible rationale is to diagnose unstable angina, but this is not a likely outcome these days. “Our troponin assays are becoming so good that we're really seeing less and less unstable angina,” he said. “There are some who argue that unstable angina doesn't exist—these troponin assays pick up every acute coronary syndrome. I actually don't think that's true. I think many of us have probably seen a patient who truly does have unstable angina, acute plaque rupture, but a totally negative, even high-sensitivity, troponin. But that sliver is tiny. These patients are really, really rare.”

Unstable angina should be diagnosed based on history, anyway, he added. “You've got to use your clinical judgment to make the diagnosis, not the stress test.”

Another rationale for stress testing might be to identify patients with increased short-term risk of MI, but that turns out to be unlikely, too. Dr. Breu reviewed four studies that looked at 30-day risk of MI in patients with negative EKG, troponins, and history. “The event rates are consistently less than 1%, and in two of the studies, it's 0%,” he said.

Additionally, in one of the studies, published in the Journal of Nuclear Cardiology in 2013, half of the MIs in the stress test arm were in patients with negative stress tests. “Stress tests cannot, unfortunately, diagnosis or predict or find all the patients who are short-term risk for acute MI,” said Dr. Breu. “Stress tests don't find that vulnerable plaque, and they often give us and our patients a false sense of reassurance when maybe what they need, frankly, is optimal medical management.”

That leads to the third, and final, possible rationale for a stress test in a low-risk patient: diagnosing stable coronary artery disease (CAD). “I think diagnosing stable CAD is a good thing for internists to do, but I don't know that it needs to be diagnosed right now, immediately, especially in a patient who's had one episode of chest pain,” said Dr. Breu.

Stable disease is more appropriately handled after discharge, he recommended. “I do think these patients need appropriate follow-up. Some of them probably do have CAD and need optimal medical therapy, but I don't think they need a stress test, and I don't think they need an angiogram,” Dr. Breu said. “Ordering these stress tests on low-risk patients costs us about $3 billion a year, and we're not reducing events that we care about at all, unfortunately.”

Pulmonary embolism (PE) is another event that hospitalists care a lot about. “Many of us in the room have likely been in a situation where one of our patients has an acute PE in the hospital. And I can tell you from first-person experience, the first question that is asked almost invariably is, ‘Were they on prophylaxis?’” said Dr. Breu.

The fear of having to answer “no” may motivate a lot of VTE prophylaxis in low-risk medical patients, but the data don't support this practice, according to Dr. Breu. “The evidence is weak, weaker than you might realize,” he said. (His recommendations don't apply to surgical or critical care patients or those who have had an MI or stroke, he noted.)

Three randomized trials provide the main evidence for VTE prophylaxis in hospitalized medical patients: MEDENOX, published in the New England Journal of Medicine in 1999; PREVENT, published in Circulation in 2004; and ARTEMIS, published by The BMJ in 2006.

These studies all found a reduction in VTE associated with prophylaxis, but there was an important caveat: Asymptomatic deep venous thrombosis (DVT) was included in the composite outcome. “They went and did venography or ultrasound, looking for outcomes,” Dr. Breu said. “The composite outcome really was driven almost exclusively by all these asymptomatic events.”

Another study, published in CHEST in 2000 and focused on patients who did have a VTE during or after hospitalization, raised additional questions about the effectiveness of prophylaxis. “They found 13 fatal PEs; 12 were getting prophylaxis,” said Dr. Breu. “It's most often a failure of the prophylaxis to work, not a failure to prophylax.”

Still, when you put all the historical data together, as a 2014 Cochrane Review did, you find a trend towards a reduction in PE with prophylaxis. “But the event rates are low: 0.23% PE in the prophylaxis arm, 0.36% in the placebo arm. That's a number needed to treat of 769,” he said. “When I talk to residents, I say, ‘Sorry to tell you, you will complete internal medicine training without ever preventing a PE.’”

Compare that number to the numbers needed to harm, found in the same review: 323 for major bleeding and 100 for minor bleeding. “I would say that preventing one PE at 769 may be worth a major bleed, but it's not a slam dunk,” said Dr. Breu.

There are also the less dramatic downsides of prophylaxis to consider, including patients' pain, nurses' time, and cost. “If you assume a number needed to treat of 769, and a length of stay of five days, it takes 3,800 injections of enoxaparin or over 11,000 injections of unfractionated heparin to prevent one PE. That is a lot of jabs,” he said.

It seems like a strong case against VTE prophylaxis in low-risk patients. “And here's the thing: This whole time I've been talking about high-risk patients,” said Dr. Breu. “If the data for high-risk patients are not as strong as you might otherwise believe, it doesn't exist at all for low-risk patients.”

So instead of routinely ordering prophylaxis, do a risk assessment, which, by the way, makes you compliant with The Joint Commission's core measure, he noted. “It doesn't mandate that you give to everybody. It just mandates that we assess risk and give it if they're high risk or, if we don't give it, say why not.”

And then, perhaps the hardest part of this high-value tip: Remember that if a patient does get a VTE, it was probably going to happen no matter what. “This is a really important message I want all of you guys to internalize,” said Dr. Breu. “They get a PE and then we blame ourselves. I think a lot of it is unfounded.”