Too many or too few: Who belongs in the ICU?
Little guidance exists for transfer, admission decisions
By Stacey Butterfield
There are almost 100,000 critical care beds in U.S. hospitals, according to a 2009 study in Critical Care Medicine.
Some experts are starting to question how these many beds are being used. One reason for concern, of course, is the cost.
Photo by Thinkstock.
“In the United States, intensive care unit (ICU) costs are a substantial portion of hospital costs, which in turn are a substantial portion of health care costs, so it stands to reason that we can constrain health care costs by reducing ICU admission rates, particularly if there are patients who are being admitted to the ICU who might not benefit,” said Jeremy Kahn, MD, an associate professor of critical care, medicine and health policy and management at the University of Pittsburgh.
The possibility that ICU treatment doesn't benefit some of the patients who receive it has been substantiated by recent research. A study in one Canadian hospital found that when the ICU was full, patients were less likely to be transferred to intensive care, but not any more likely to die, according to results published in the March 26, 2012 Archives of Internal Medicine.
Then, a larger study, of more than 30,000 patients at 118 Veterans Affairs hospitals, found that the severity of illness of patients admitted to ICUs varied very widely among hospitals, reflecting “a lack of consensus about which patients most benefit from ICU admission,” the authors concluded, in the Sept. 10, 2012 Archives.
“One would think that for a patient of the same severity generally similar types of treatment would be given,” said lead author Lena M. Chen, MD, ACP Member, an assistant professor in internal medicine at the University of Michigan in Ann Arbor.
But when it comes to ICU admission, no such uniformity exists right now in the U.S., the experts agree. However, they have ideas about how everyone from researchers and policymakers to administrators and hospitalists can help correct this situation.
The tricky ones
Two kinds of patients are responsible for most of the variation in intensive care unit admission practices—the too-healthy and the too-sick.
“They may be very healthy but the doctor says, ‘Put them in the ICU.’ Or they may be at the other extreme—almost dead, end of life—and nobody wants to deal with putting or keeping them on the ward and involving a palliative care team in a drawn-out process,” said Neil Halpern, MD, FACP, chief of critical care at Memorial Sloan-Kettering Cancer Center in New York City.
The latter patients are more uncommon, but also more costly, noted Jack Zimmerman, MD, an emeritus professor of critical care at George Washington University in Washington, D.C. These patients are too sick to be saved by the ICU, but they end up there anyway, sometimes for a very long time. “That is a problem that physicians are facing in general—how to better care for patients at the end of life,” he said. “[One cause] is unrealistic expectations on the part of families.”
These ICU admissions can also result from physicians' failure to communicate well enough and soon enough with patients. “We generally do a poor job of soliciting patient preferences for intensive care at the end of life. If we did a better job, we might identify patients who not only would not benefit, but also wouldn't want intensive care,” said Dr. Kahn.
Physicians do elicit these preferences when a bed shortage necessitates it, according to the Canadian study. Although the studied patients' mortality risk didn't change when the ICU was full, their goals of care did—in the direction of less intensity. “These were discussions that were forced to occur, because there weren't as many beds,” said Hannah Wunsch, MD, assistant professor of epidemiology and anesthesiology at Columbia University in New York City.
That's good news, since it means overuse of ICUs by too-sick patients could potentially be remedied by the conversations about end-of-life decisions that physicians are already supposed to be having with their patients.
The issue of too-healthy patients is more challenging and more prevalent. “In our studies, they probably amount to 40% or 50% of the patients on the [intensive care] unit. They're not receiving any life support at all. They're patients who are receiving some technological monitoring and concentrated nursing care,” said Dr. Zimmerman.
Multiple studies have shown that whether these patients are admitted to the ward or the ICU varies widely by hospital. “We've looked at diabetic ketoacidosis. At hospitals across New York State, they admit anywhere from zero [diabetic ketoacidosis cases] to the ICU up to 100%, depending on what hospital you're in,” said Dr. Wunsch.
Post-surgical patients are also major contributors to the variation in ICU admission patterns. “There are a few procedures, like carotid endarterectomies, where it used to be everybody went to the ICU,” said Dr. Wunsch. Now, some hospitals' postsurgical patients are sent to the ward and some are sent to the ICU.
In some cases, the variation may result from differences in the capabilities of regular hospital wards. “If the patient's had vascular surgery, the patient may just need the nurse to check pulses somewhere. The ICU may be the only place in the hospital where the nurses have the training and experience and are not overwhelmed with other things that they can actually do that,” said Dr. Zimmerman.
Nurse workload affects nonsurgical patients as well. “If you have a very low nurse-to-patient ratio on the hospital inpatient wards, then a less sick person will have to go to the ICU,” said Dr. Halpern.
And, despite their suspicions and observations, researchers haven't definitively proven that the high-ICU-use hospitals are making the wrong choice. “What we don't have are good clinical studies indicating that we can systematically identify these low-severity patients and safely admit them to the ward,” said Dr. Kahn.
“Right now it's a lot of gestalt,” agreed Dr. Chen. Experience generally gives physicians more skill and confidence in making this choice, but it's still tricky, the experts agreed.
In this setting of uncertainty, hospitals tend to develop their own cultures about ICU admissions. The culture may be shaped by administrators, physician leaders or even other clinicians. “ICU nursing may be welcoming for patients, or may push back, saying, ‘That patient is not sick enough,’” said Dr. Halpern.
One potential remedy for this variability is to take admission decisions out of the hands of individual clinicians. Dr. Wunsch's research has shown that the United Kingdom has many fewer ICU beds per capita than the U.S, but similar outcomes. Their triage system might have something to do with that.
“In most other countries, almost all the ICUs are closed. In ours [in the U.S.], most of them are open. That might be a starting point,” she said. “Then it's less on any individual physician to have to make the decision for their own patients.”
In addition to reducing the conflict between what's best for the patient versus what's best for the system, closing the ICU brings more expertise to bear on admission decisions, noted Dr. Halpern. “If the ICUs are managed with careful triage, then you have a much better chance of managing those ICU beds appropriately,” he said.
There aren't enough critical care specialists to staff all the nation's ICUs, so there's an opening for hospitalists to gain the expertise to fill this role, possibly through intense one-year critical care fellowships, for example. “If I were running a community hospital, and I had an excellent cadre of hospitalists and I wanted them to work in my ICU, I would say to them we will financially support you during your critical care training on the condition that you come back here to staff our ICUs upon completion,” said Dr. Halpern.
Additional training for regular floor nurses could be another solution to overuse. As Dr. Wunsch's study showed, many hospitals find it feasible to have diabetic ketoacidosis patients cared for on the medical ward, for example. “That requires training people to be comfortable with protocols involving insulin infusions on the floor,” said Dr. Wunsch.
Simple or not, such change won't happen on its own. This could be another opportunity for hospitalists to step up. “Sometimes it just takes a courageous individual to start questioning whether or not this needs to happen all the time,” Dr. Wunsch said.
In addition to a leader, changing these patterns or protocols requires buy-in from all the stakeholders. “There's a collaboration between physicians, administrators and nurses to identify alternatives to care for those patients,” said Dr. Zimmerman. “Sometimes it means just an up-training and sometimes some increased staffing on a floor. Sometimes it means putting together a stepdown unit.”
Stepdown, or intermediate care, units are one of his preferred solutions to the overuse of ICU beds. “There exist patients in the hospital who don't need intensive care but more than they can receive on the floor, and intermediate care is the place for them,” he said. “Over the years, I've known of many hospitals that have been able to save on costs and avoid having to build more ICU beds. We added intermediate care beds and used our ICU more for life support.”
Stepdown units pose disadvantages of their own, however, according to Dr. Kahn. “As you create more and more severity silos within the hospital, you end up making triage decisions even more difficult,” he said. It's also uncertain how these units work out financially. “Their cost structures aren't necessarily lower than intensive care units. There's a strong conceptual rationale that stepdown units might be useful, but there's less empiric data,” he said.
The cost structure of ICUs affects any plans to change admission patterns, the experts noted. Admission of borderline patients may be necessitated by cost-cutting elsewhere in the hospital. “One way to cut back on floor staffing and costs is to stick patients into the intensive care unit to make up for poor staffing on floors,” said Dr. Zimmerman.
The ICU itself could also be an important source of income for the hospital. “Hospitals have an incentive to have high-severity, high-paying DRGs, and admitting a patient to the ICU makes it easier to code that patient as a higher-paying DRG. Additionally, ICUs are used to support profit-generating service lines, such as cardiothoracic surgery, oncology and acute myocardial infarction,” said Dr. Kahn.
With their high technology and staffing, ICUs are expensive to run, and most of the cost is fixed. That means reducing the number of patients admitted to the ICU will cut hospital revenue more than it does expenses—a proposition that's likely to be unappealing to hospital administrators.
It's why Dr. Kahn favors a more drastic solution to ICU overuse. “The only way to reduce the cost of hospital care is to reduce fixed costs, and the way to do that over the long term is by closing ICU beds,” he said.
Dr. Halpern agreed, based on his studies of the ICU bed supply in the U.S. and current utilization. “Looking at it from the big picture, just the numbers, we have an impression that in the U.S., there are more than enough ICU beds, and that no more need to be opened. This may be blasphemous, but ICU beds could be reduced,” he said.
He might not call it blasphemy, but one expert definitely disagrees. “If anything, we're going to be needing more ICU beds. Our population is aging. The patients who come into the hospital are going to be sicker; they already are,” said Gabriel J. Escobar, MD, regional director for hospital operations research for Kaiser Permanente's Northern California Region and Division of Research in Oakland, Calif.
“The best thing is to optimize the care that we give, rather than decrease the number of beds,” he said. He pointed out that one group of patients—those who are transferred to intensive care from general medical-surgical wards or stepdown units—has disproportionately high mortality, suggesting that clinicians' ability to judge who needs intensive care may not be as good as it could be.
“If this patient has a probability of mortality of 0.2% and this patient has a probability of 20%, most [physicians] are pretty good at making that distinction. But when you have a patient who is in between, that level of discernment is very difficult to do just by eyeballing,” he said.
Of course, everyone would like to make that choice easier and more accurate, but there's uncertainty about how to do that. “A physician is always going to feel that the safest place for his or her patient is the ICU,” said Dr. Kahn. “If they had the choice, who wouldn't want to be in an intensive care unit?”
Maybe lots of patients, given the infections, disruptions and restrictions associated with intensive care. Dr. Wunsch offered a counterexample. “You've got Granny who's 90 years old, who has come out of surgery and doesn't really need to be in the ICU next to that patient with multidrug-resistant Klebsiella and across the hall from the patient who's coding all night,” she said. The immobility associated with typical ICU care can also slow recovery, she noted.
It adds up to a lot of contradictions and confusion for the individual hospitalist making ICU admission decisions. Admitting a patient to the ICU could help, harm or make no difference in his recovery. It could help the patient but harm the health care system. Or reduce overall costs, but not for your hospital. Creating an intermediate unit could improve care and cost or further confuse your triage decisions.
Even the advocates of reducing ICU usage recommend admitting borderline patients on an individual basis. “I would err on the side of putting the patient in the ICU,” said Dr. Zimmerman. (Taking them back out is always an option, he noted. “You can save almost as many resources by shortening the length of stay of these patients as you can from not admitting them to begin with,” he said.)
The experts acknowledge the dilemma they're posing. “I would be hard pressed to stand up in front of a group of physicians and say here's the information you're lacking on how to better triage,” said Dr. Wunsch.
But they hope to soon have more answers. “Severity of illness scoring is now feasible outside the ICU. You have for the first time the ability to come up with probability models that will be quite accurate,” said Dr. Escobar. Objective severity of illness scores available in real time could help rationalize ICU triage, he added.
“I think we'll get there in four, five, six years. We'll have a better idea how to triage,” predicted Dr. Wunsch.
As for the cost issue, Dr. Kahn suspects that payers and administrators will resolve that dilemma soon. “As reimbursement rates go down, hospitals are going to make a concerted effort to reduce the fixed costs of care, and they will do that by reducing the number of ICU beds,” he said.
That would probably mean more work for hospitalists. “Physicians will be called upon then to provide the same quality care without the number of ICU beds they currently have. I believe that's possible,” Dr. Kahn said. “Physicians should be prepared.”
Are you involved in hospital medicine? Then you should be getting ACP Hospitalist and ACP HospitalistWeekly. Subscribe now.
From the July 29, 2015 edition
- Score validated for predicting bleeding risk in inpatients
- Readmissions common with severe sepsis and mild or moderate AKI, study finds
ACP Career Connection
Looking for a new hospitalist position?
ACP Career Connection can help you find your next job in hospital medicine. Search hospitalist positions nationwide that suit your criteria and preferences. Jobs are posted about two weeks before print publication of Annals of Internal Medicine, ACP Internist, and ACP Hospitalist. Exclusive “Online Direct” opportunities are updated weekly. Check us out online.
ABIM Maintenance of Certification for Hospitalists
Hospital-based internists have the option of maintaining their certification in either Internal Medicine or Internal Medicine with a Focused Practice in Hospital Medicine. Learn more about resources from ACP to complete both MOC programs.
Internal Medicine Meeting 2015 Digital Presentations
Choose from over 170 recorded Scientific Program Sessions and Pre-Courses. Available in a variety of packages and formats so you can choose the combination that works best for you.