Emergency providers determine whether or not to admit patients to the hospital every day, but a new study suggests that while many of these decisions are consistent and clear-cut regardless of region or hospital, for certain common, low-mortality conditions, some physicians are as much as six times more likely to admit patients than others.1
Such findings may not raise many eyebrows at first glance, but the potential savings that could be achieved by reducing this variation in decision making are truly eye-popping. Study authors estimate that the health care system could save as much as $5 billion a year.
The potential savings come into sharper focus when you consider the statistics highlighted in the research: EDs are the main source of hospitalizations in this country, and emergency providers make this important decision about whether to admit a patient as often as 350,000 times a day.2 Further, the result of these decisions is close to 20 million admissions per year.3
In light of these numbers, you would think that more attention would be focused on this area, according to Keith Kocher, MD, one of the study authors and an assistant professor of emergency medicine at the University of Michigan in Ann Arbor, MI. "I am also a practicing emergency physician, and from my perspective on my shifts, [the decision to admit a patient] is by far the decision that comes with the most potential costs and downstream consequences," he explains.
Kocher acknowledges that emergency providers make many decisions that come with cost and quality implications. For instance, whether to get a computed tomography (CT) scan on someone has been much discussed in recent years. However, in terms of impact, Kocher suggests that this kind of decision pales in comparison to the admission decision.
Evidence is lacking
To establish where variations in decisions about admission were most prominent, and what impact these variations had on mortality, researchers analyzed national data on more than 28 million emergency visits to 961 hospitals in 2010. These involved ED visits by adults with 15 of the most commonly admitted medical and surgical conditions.
"What we found is there really aren’t any differences or variations with respect to those high-risk conditions where pretty much everybody feels the standard of care is to hospitalize," says Kocher.
For example, the researchers report that there was little variation between hospitals or physicians on decisions to admit for patients with heart attacks, acute kidney failure, or sepsis.
However, there was considerable variation in decision making for patients who presented to the ED with chest pain but were not experiencing a heart attack. The researchers report that patients at the hospitals with the highest rates of admission were as much as 6.55 times more likely to be admitted than patients who were treated at hospitals with the lowest admission rates.
There was also variation in admission decisions for patients who presented with soft-tissue infections and asthma-related difficulties. Investigators say that some hospitals were three times more likely to admit these types of patients than others. Similarly, patients with chronic obstructive pulmonary disease (COPD) or urinary tract infections were twice as likely to be admitted at some hospitals compared to others.
What accounts for this variation? Kocher suggests that for many of these conditions, there is uncertainty in the literature about what the optimal admission decision should be. "There is just not a lot of evidence to suggest what the right decision is," he says. However, Kocher adds that there are also non-clinical factors that can influence how admissions decisions are made.
"A lot of times, there are important social, family, and hospital resource factors that really have a big part in admission decisions," says Kocher. "For example, if you’ve got someone who is borderline for hospitalization, one alternative may be to arrange for expedited outpatient follow-up care. However, whether or not you can arrange that from your ED depends on the kind of resources in your community or what kind of health care system you work in."
Kocher adds that in making an admission decision, physicians consider such factors as whether there are pathways to accommodate expedited follow-up or mechanisms for delivering some of the needed care in the home setting. Further, while there is plenty of evidence that providers working in the same work setting can make different decisions regarding admission, it is also clear that culture plays a role in determining how things are done in the ED.
"If you come in as a new provider, you quickly sort of adapt to what everyone else is doing," says Kocher. "There are just so many factors that can influence this, and I think that is why you see a lot of this variation. It is in those conditions where there is a lot of gray."
Find better resources, tools
Despite the uncertainty involved with making admissions in borderline cases, the researchers note that in all five of the conditions for which they found variation in the admission decisions, the in-hospital mortality rate was very low. Further, the estimated in-hospital charges associated with these types of admissions were in excess of $52 billion, although payments to hospitals for these charges were probably closer to $16 billion, according to researchers. Nonetheless, the researchers note that there is still a clear opportunity for potential savings.
Amber Sabbatini, MD, MPH, the lead author of the study and an instructor of emergency medicine at the University of Washington in Seattle, WA, says that the findings underscore the need to find better ways of determining which patients with less serious medical conditions are at the highest risk of having serious complications and, therefore, need to be hospitalized.
"The important thing is that sometimes the type of care you get depends on where you show up, so the same patient is treated differently at different places," says Sabbatini. "The fact that hospitals are so different from each other and that entire groups of emergency physicians are, in some cases, behaving so differently would suggest that there are opportunities to reduce the variation, and that may actually mean creating better resources and tools."
In cases in which admission decisions are appropriately variable due to a scarcity of resources in the community for follow-up care, then solutions would involve creating lower-cost, more efficient alternatives for patients, notes Sabbatini. Further, she adds that this is an area that is ripe for improvement, but too often overlooked.
"While there has been increasing interest in the formation of ACOs [accountable care organizations], the creation of bundled payments for inpatient care … and moving care to medical homes, the ED has sort of been lost in all of that," says Sabbatini. "Part of what my co-authors and I were trying to do [with this research] was highlight the importance of the ED as a center for care coordination, and to think strategically about how we could use the ED to help create care that is more coordinated and efficient."
Potential solutions include the more effective leveraging of case managers and social workers to help with transitions and a more responsive outpatient care environment, says Kocher. "You really need to have those [resources] in place to allow these types of decisions to happen."
Payment reforms would help
Sabbatini and Kocher agree that payment reforms are needed to ensure that patients are not admitted for financial reasons. Kocher notes that there have been some high-profile cases in which emergency providers were improperly incentivized to admit patients because they brought in revenue for hospitals. However, he observes that the opposite scenario is also possible for hospitals that are located in areas that serve large numbers of uninsured patients.
It is clear that under new shared-risk models, it is going to be increasingly less profitable for hospitals to admit patients, observes Sabbatini. "At the same time, hospitals have to find ways to stay profitable, and this might actually be driving the development of more cost-efficient alternatives," she explains.
At the institutional level, emergency medicine leaders need to be actively involved in devising these solutions, adds Sabbatini. "That means creating local resources for patients, improving efficiency, and also assisting physicians with appropriate decision-making," she says. "I’ve seen interdisciplinary protocols that are created for the ED to guide practice be very effective at standardizing practice."
Kocher agrees that solutions need to come at the group or department level, as well as at the hospital and system levels. "As hospitals become more integrated with outpatient care and with things like the ACO structure, these types of changes become more feasible because they really require coordination between the ED, which rests in this gray area between the inpatient and the outpatient, and the outpatient part of delivery system," he says. "If you start thinking about how to create alternatives to hospitalization, that really requires resources in place to facilitate these alternatives."
Since it can take time to facilitate alternative plans, policy makers may need to revisit the metrics that are used to assess EDs and emergency providers, says Kocher. "There is certainly a lot of pressure on emergency physicians to make quick and timely dispositions of their patients, but it may be important to become more flexible about this, and allow for potentially longer ED stays if it allows for these alternative plans to develop," he explains. "Unfortunately, this runs counter to a lot of how emergency medicine works where some of the administrative quality metrics are all about decreasing length of stay (LOS) and shortening time to discharge."
Kocher acknowledges that such metrics are important markers of patient satisfaction, but he suggests that added flexibility on these measures could potentially improve the patient experience in many instances. "We are talking about enhancing that experience with what might be the sacrifice of a longer LOS," he says.
If emergency providers don’t fully engage on these issues, then other voices will surely fill the void, according to Kocher. "We need to recognize that, particularly in this era of increasing cost consciousness, we need to be proactive within our own backyard," he says. "Increasingly the federal government and other payers are going to be looking at these types of issues for emergency medicine. We need to be ready to respond to those pressures."
Sabbatini agrees, noting that there is a lot of responsibility on the practicing physician to advocate for the care of their patient. "That means we don’t take a back seat," she says. "We proactively work together to create a united front."
- Sabbatini A, Nallamothu B, Kocher K. Reducing variation in hospital admissions from the emergency department for low-mortality conditions may produce savings. Health Affairs 2014;33:1655-1663.
- National Center for Health Statistics. National Ambulatory Medical Care Survey: 2010 emergency department summary tables. Hyattsville, MD: NCHS; 2010.
- Kocher K, Dimick J, Nallamothu B. Changes in the source of unscheduled hospitalizations in the United States. Med Care 2013; 51:689-698.
- Keith Kocher, MD, Professor of Emergency Medicine, University of Michigan, Ann Arbor, MI. E-mail: firstname.lastname@example.org.
- Amber Sabbatini, MD, MPH, Instructor of Emergency Medicine, University of Washington, Seattle, WA. E-mail: email@example.com.