The imperfection of ICU admission decisions

ICU clinicians and experts continue to hunt for methods to make intensive care high quality and high value.


The decision about whether to send a patient to the ICU is highly variable, even after years of expert debate and research on which patients benefit most from this expensive resource.

Image by Getty Images
Image by Getty Images

A study published in Critical Care Medicine in October 2018 showed wide disagreement among physicians on whether hypothetical patients should be admitted to the ICU based on illness severity. It also found they were more likely to want to send patients to the ICU if a bed was available or if family was present.

The latter two factors hardly seem like clinical criteria on which to base the decision, but the problem is, there are no such criteria. Experts identify myriad factors that go into decisions about ICU admission, including what the hospital or ED protocols are, how comfortable hospitalists are handling sicker patients, how many beds are available, and who is the ultimate decider.

A set of ICU Admission, Discharge, and Triage Guidelines were published in Critical Care Medicine in August 2016, but the authors noted a limited amount of high-quality evidence on which to base their recommendations.

“This problem just defies standardization, by which I mean there is never going to be such a thing as a set of agreed-upon ICU admission and discharge criteria,” said Jeremy Kahn, MD, MS, a critical care physician at the University of Pittsburgh Medical Center. “It's just too fungible—the notion of who needs an intensive care unit. Obviously, people who are on life support, like a mechanical ventilator, or vasopressor certainly need admission to the ICU, but outside of that, predicting the true need for intensive care has just proven to be impossible.”

“There are so many stakeholders involved, and they have different voices at different institutions,” noted Meeta Prasad Kerlin, MD, MS, a critical care specialist at the University of Pennsylvania in Philadelphia. “The institutional decisions about what resources they are going to have on the floors are going to differ, and they may be based on reimbursement strategies or they may be based on sort of who has the loudest voice when there's a conflict or something like that.”

Despite these factors, ICU clinicians and experts continue to hunt for methods to make ICU care high quality and high value, from measuring patient outcomes to reducing ICU beds to developing new levels of care.

Latest research

Recent research on the effects of ICU care has shown mixed results. A large, randomized study in France found no reduction in six-month mortality after the implementation of a program to systematically admit critically ill elderly patients to the ICU. The results, published in the Oct. 17, 2017, JAMA, refuted the hypothesis that “just being able to put more people in the ICU, even the ones that are sort of on the margins, is going to somehow make a difference,” said Hannah Wunsch, MD, MSc, senior scientist at Sunnybrook Health Sciences Centre in Toronto.

On the other hand, another study published in JAMA in September 2015 looked specifically at Medicare data from older patients with pneumonia and did find a mortality benefit linked with ICU admission, with no difference in hospital costs.

Thomas Valley, MD, MSc, coauthor of that study and a pulmonologist and critical care physician at the University of Michigan in Ann Arbor, said the overall evidence about which patients should and shouldn't be sent to the ICU is fairly scarce. “We really don't have a great way of assessing benefit right now.”

One issue is that mortality risk may not be the best way to measure which patients derive the most benefit, Dr. Valley said. “A lot of the work that's been done before has focused on an individual's risk of dying and really focused on making sure patients with the highest risk of dying go to the ICU. That makes a lot of conceptual sense, but it may miss patients who might have a lower risk of dying but still could benefit greatly from going to the ICU.”

Dr. Valley is not certain how best to measure meaningful improvements in patients' lives after ICU admission, but he expects this will be a topic of research over the next decade.

The overarching criteria by which the medical community judges how well ICUs are being used in the country depends on what definition is being used to decide who should be admitted, he continued. “We're still in the process of actually trying to figure that question out, and until we set that bar, then it's tough to say whether we're overusing or actually underusing the ICU.”

The question of over versus underuse is closely tied to decisions about how many ICU beds a hospital should have. “A big issue in this country is not knowing the right number of ICU beds,” Dr. Kerlin said. “We have an order of magnitude more ICU beds in the U.S. than some other industrialized countries in the world—we have literally almost 10-fold more than the U.K. and we have double Canada.”

The American supply of ICU beds may have dropped in recent years, however, according to a research letter by Dr. Kerlin and colleagues in the Oct. 16, 2018, Annals of Internal Medicine. It showed a decrease in ICU admissions in the United States between 2006 and 2015, albeit with significant geographical variation.

“We're starting to see a plateau finally,” Dr. Kahn said, attributing that to decreased hospital spending on infrastructure and the movement toward health care system consolidation. Another contribution might be strengthened efforts to collect patients' care preferences, he speculated.

“The fewer 85-year-old patients with dementia I admit purely because nobody has had an end-of-life discussion with them, the better,” Dr. Kahn said.

Dr. Kerlin predicts that the number of ICU beds will decrease further as administrators and physicians learn to become more efficient and uniform with admission decisions. “There was an estimate that our ICUs sit at about 60% to 65% occupancy at any given time,” she said. “That means that a third of our ICU beds in the country are not being used at any given time, which feels like a lot of waste in our system. So I think that if we could improve our triage decisions, we could better address that and possibly just be more efficient with our ICU utilization.”

New models of care

One potential solution to the ICU triage dilemma could be so-called intermediate care, providing more than a general ward but less than a full-fledged ICU. There are even fewer data on that, however.

“People are more aware of it as a piece of the model that needs to be addressed, but it's even harder than looking at the benefits of intensive care, because of the complexity of who comes in and out of intermediate care,” Dr. Wunsch said.

There's also the problem that intermediate care has a very heterogeneous definition. “We don't really know at this point whether places that say they're providing intermediate intensive care are providing essentially just telemetry, or whether they're providing organ support such as vasopressors or noninvasive ventilation; we just don't know what intermediate care looks like in different institutions, or even what the nurse-to-patient ratio is in different institutions,” Dr. Wunsch observed.

“There's a lot of interest in trying to understand it better,” Dr. Kerlin said. “We don't know if they help offload ICUs. We don't really know for sure if they decrease costs, which might be one big motivator to open them. We don't know if they change outcomes one way or the other—if they improve outcomes or shorten length of stay or worsen outcomes because patients aren't getting to the ICU when they should be.”

Dr. Kahn has concerns over this model. “Overspecializing hospital beds has risks, and I think that it's just one more judgment decision that hospital-based physicians have to make,” he said.

He suggested that a swing- or flex-bed system in the ICU that enables physicians to designate patients as ready for a stepdown without having to shuffle anything around physically might work out better.

“[This is] a really novel way where you can essentially perfectly triangulate supply with demand as opposed to having to figure, ‘OK, I'm going to open a stepdown unit and maybe the stepdown unit some days is going to be underfilled but some days is going to be completely filled,’” he said. “That to me is a frontier that is just now catching on.”

Next steps

There's some debate among the experts about whether ICU triage decisions ever will, or should, get more protocolized.

In 2013, Dr. Wunsch predicted to ACP Hospitalist that more scientific methods of intensive care triage would be in place within six years (see “Too many or too few: Who belongs in the ICU?” in the February 2013 issue). “We're a little bit better, but not as far as I think most of us would have hoped,” she said now. “It's so heterogeneous in terms of what's going on across the country. Everybody has this optimism that we should be able to predict well and really intervene earlier and make a difference, but it's not so clear that it's making a huge difference yet.”

Dr. Kahn is not optimistic that a system will ever be developed. “We need to once and for all stop trying to develop objective triage criteria because that I believe is just a waste of time. They're never going to work because if there's an ICU bed available, people will override those triage criteria. We see that every day in the hospital.”

The answer, instead, is in the current trend toward cutting ICU beds, he said. “There has to be a general shift towards more value in health care both in hospital medicine at large and in intensive care medicine, including more value-based reimbursement policies.”

On the bright side, he thinks the physician response to these changes will be improvements in the decision making about ICU placement. “Physicians in general are good at triage, and the data would say that when we're forced to triage a scarce resource, we're pretty good at triaging it on an individual, case-by-case basis,” Dr. Kahn said.