Home Feature Does that pneumonia patient really need to be hospitalized?

Does that pneumonia patient really need to be hospitalized?

April 2005

While a new study examining the best treatment setting for patients with community-acquired pneumonia (CAP) will probably not change your day-to-day practice, it highlights some interesting opportunities for hospitalists to take the lead on streamlining throughput for a potentially large group of inpatients.

The study, which appeared in the Feb. 1, 2005, Annals of Internal Medicine, offers perhaps the most solid evidence to date that low-risk CAP patients can be safely and effectively treated outside the hospital. The article builds on years of research pointing out that there is little benefit to treating a number of low-risk CAP patients in the hospital.

Most hospitalists are already aware of the prognostic tool used in the study, the pneumonia severity index (PSI), also known as the PORT or Fine score. The research, however, may focus new attention on the tool, which is largely unused by many physicians.

Even physicians who use the PSI to predict pneumonia patients’ risk of mortality acknowledge that it can be unwieldy. Its complexity makes it difficult to remember, and like all risk stratification tools, it doesn’t offer the final verdict on many patients, particularly complex individuals with multiple co-morbidities.

lung infectionInpatients: New recommendations on treating lung infections

Experts, however, say there is another reason that the PSI has failed to catch on with physicians, particularly hospitalists: It is widely viewed as a triage tool to be used by the emergency department when pneumonia patients are admitted.

But because the PSI is just as unpopular among ER physicians as it is among hospitalists, no one is doing a good job of calculating the mortality risk of CAP patients. As a result, experts say, a significant number of low-risk CAP patients who could be safely and effectively treated at home are instead being hospitalized.

The number of these patients varies from institution to institution, but it is believed to be significant. One Boston teaching hospital found in a chart review that up to half of its patients admitted with CAP fell into a low-risk category.

While no one knows exactly how many low-risk pneumonia patients who could be treated at home are being admitted, many believe the issue will hit the radar screens of hospitals that are desperately looking for ways to free up beds.

Opportunity for hospitalists
Scott Flanders, MD, an authority on CAP who is associate professor of medicine at the University of Michigan and director of the University of Michigan’s hospitalist program, says a recent review of admission data at his facility found that a significant number of low-risk CAP patients are admitted to the hospital. Because research indicates that his hospital is not alone, he thinks the recent Annals of Internal Medicine study shines a light on one way that hospitalists can help their institutions improve throughput and bed availability.


“This study shows that there are opportunities for hospitalists to find capacity in their system and open up beds by avoiding unnecessary admissions,” Dr. Flanders explains. “I see the PSI as a useful tool for hospitalists to work with their institutions and their emergency rooms to see if they can identify a population of patients that would be safely and more appropriately managed elsewhere. When you talk about freeing up beds, that’s where a hospitalist group could capitalize on the results of this study.”

Dr. Flanders notes that the University of Michigan is exploring the idea of creating an observation unit to identify low-risk patients who may not need to be admitted. And while hospital officials are still defining exactly how an observation unit would work and what types of patients it would see, Dr. Flanders says that pneumonia patients are an essential part of the conversation.

“We specifically discussed the pneumonia population as a group that we could target to avoid unnecessary admissions to the hospital,” he explains. “In that sense, having an algorithm that is easy for ER doctors and hospitalists to apply, and one that has been prospectively validated in a randomized trial, really gives them a lot of power to say that unless there is a good reason, this patient should not be admitted.”

While a number of hospitals are reportedly looking into observation units as a way to care for low-risk patients who won’t benefit from a hospital stay, there are questions about whether low-risk CAP patients are a good match.

Alpesh Amin, MD, an expert in CAP and executive director of the hospitalist program at the University of California, Irvine, says that in his experience, short-stay units may help treat some of the low-risk pneumonia patients cited in the Annals study.

In general, Dr. Amin says that short-stay units offer the most help to patients who are suffering from problems that can be resolved in a finite time frame. Ideal patients for these units include individuals who are dehydrated and will perk up after four or five hours of IV fluids, patients who can benefit from mild diuresis, individuals who require short-term nebulizer therapy, patients with soft tissue infections such as mild to moderate cellulitis, and electrolyte disturbances.

When it comes to low-risk pneumonia patients, Dr. Amin says, short-term units will likely offer the most help treating individuals who are also suffering from dehydration or nausea. Once the dehydration and nausea has been resolved, he says, the patient can be sent home on oral antibiotics.

Working with the ED
Observation units are not the only approach that hospitalists can spearhead to help keep low-risk pneumonia patients out of the hospital. Even as the University of Michigan develops an observation unit, Dr. Flanders says its hospitalists may be able to make inroads into treating low-risk CAP patients on an outpatient basis by working more closely with ER physicians.

Because the hospitalist program has beefed up its staffing levels, in part to provide 24/7 coverage, Dr. Flanders says he has given his hospitalists a simple message: “You are more than able to see a patient who has been admitted to you by the ED and discharge them from the ER if you do not feel they should be admitted.”

While many hospitalists have little to no feedback in making admission decisions, it is a model that is being embraced at some hospitals. Johns Hopkins Bayview Medical Center in Baltimore, for example, has developed a system where hospitalists and ER physicians work together to decide which patients need to be admitted.

The hospital created an “admission pager” “a dedicated wireless phone system “that ER physicians use to consult with hospitalists before every admission. (For more information on this program, see the March 2004 issue of Today’s Hospitalist online at www.todayshospitalist.com for more information.)

Eric Howell, MD, associate director of Bayview’s hospitalist service, says that in general, treating low-risk pneumonia patients at home is not a top priority at the hospital. While payers in the area are quick to deny payment for what they view as unnecessary care for patients with chest pains, he explains, low-risk CAP patients are not high on their list. As a result, Bayview hasn’t given low-risk CAP patients the same priority as patients who present with a complaint like chest pain.

He notes that when beds are in short supply, however, that changes. Hospitalists manage bed availability in part by keeping a close eye on which CAP patients ER doctors want to admit.

“If I have a 35-year-old patient with community-acquired pneumonia and a pulse ox of 98 percent,” Dr. Howell says, “I’ll consider sending him home. I’d rather take the 65-year-old woman with hypoxia, renal failure and liver disease, all of the things that the PORT prediction rules says add up to increased morbidity and mortality.”

Protection from liability
Dr. Howell, however, estimates that only about one-third of the hospitalists in his service use the PSI to make that type of decision. How many practicing hospitalists, after all, need a complicated formula like the PSI to know they need to admit the 65-year-old woman in the above example?

Critics like to point out that the PSI doesn’t take into account key patient characteristics like social status. If your pneumonia patient is an alcoholic or homeless, for example, you’re not going to discharge them even if they have a low PSI score.

While your clinical judgment gives you enough information to treat many of your CAP patients, physicians who specialize in the disease say the tool can complement your instincts. If you’re going to work with ER physicians to send low-risk patients home for treatment, for example, the PSI can provide some back-up to protect you from legal liability.

Some hospitalists interviewed for this article said they worry that sending a patient home from the emergency department when the ER physician recommends an admission could expose them to liability if the patient develops problems. Hospitalists familiar with the PSI, however, say that because the Annals of Internal Medicine study used a randomized, controlled trial to prove its validity, the tool goes a long way toward addressing those concerns.

“The bottom line is that you need to know if patients are in a low-risk PSI class so you have the backing that they probably don’t need to be admitted,” Dr. Flanders says. “The Annals study tells us that low-risk patients do just as well with oral therapy as they do with IV therapy, so you clearly have supporting evidence that a low-risk patient who is hospitalized can be treated with oral therapy.”

Dr. Flanders says that support becomes even more important if you decide on outpatient treatment for a high-risk pneumonia patient, someone with a PSI score of 4 or 5. He notes that recent guidelines from organizations like the Infectious Diseases Society of America call for those individuals to be hospitalized.

While data show that up to 90 percent of higher-risk CAP patients are being hospitalized, a small percent are being sent home, possibly putting some physicians at risk if something goes wrong.

Follow-up care
If you do make the call to send a low-risk pneumonia patient home from the ER or discharge her from the ward early, another important consideration is follow-up care.
The researchers in the Annals study used an extraordinary level of follow-up care for the patients they sent home, including a home visit by a nurse within 48 hours and outpatient clinic appointments seven and 30 days after the initial diagnosis was made. Few hospitalist programs can offer that level of follow-up care, however, which raises an important question: What kind of care do you need to offer low-risk pneumonia patients you send home from the ER or discharge early to adequately care for the patient “and protect yourself legally?

While Dr. Flanders emphasizes that he’s not a lawyer, he speculates that if a blood culture comes back negative and you’ve arranged for reliable patient follow-up in the next few days, you’ve likely met your obligation as a hospitalist.

“If patients have a follow-up appointment and they don’t have any positive results,” he says, “I think you’re meeting the standard of care. Ideally, you have a mechanism to check blood cultures, or you have someone else who is going to see the patient who knows what to check on when that patient shows up in the office, whether it’s results of the blood cultures, did they send sputum, or anything else that needs evaluation.”

Hospitalists can use other approaches to make sure these patients receive adequate follow-up care. The University of Michigan, for example, is working with several visiting nurses associations to care for patients who may have problems seeing their primary care physician or getting into a clinic. Ideally, Dr. Flanders says, a visiting nurse will see the patient at home within 72 hours of discharge.

Surprising results
Finally, even if you have no interest in using the PSI to determine which pneumonia patients can safely be treated at home, the tool has powerful uses in other scenarios. If you determine that your CAP patient has a low PSI score, for example, you might decide to give the patient oral, not IV, medications.

Some hospitalists say they use the PSI to gather more information on the patient and check their clinical judgment. Steven J. Atlas, MD, assistant professor of medicine at Harvard Medical School and a general internist at Massachusetts General Hospital, recalls a recent patient who presented with COPD, a pleural effusion and an infiltrate. While there was no question that the woman needed to be admitted, Dr. Atlas, who has published numerous articles on CAP, nonetheless found the PSI useful.

“I was interested in asking whether this patient was high risk,” says Dr. Atlas, “as a barometer of whether she was class 4 or class 5, whether she was likely to have a 5 percent to 10 percent mortality rate as a class 4, or higher because she was a class 5. The tool can provide some additional information on how sick the patient may be in the next 48 hours.”

Sometimes, Dr. Atlas adds, the results can be surprising. About a year ago, he admitted an obese patient who had been diagnosed with CAP and cognitive problems.

“When you looked at this guy,” Dr. Atlas says, “he didn’t look so bad, but his PSI score was quite high. When we went back and calculated his score, we realized that he could have been in an intermediate care bed. A hospitalist would have said that person didn’t have good clinical acumen, but the truth is that the physician didn’t get all the information available.”

Edward Doyle is Editor of Today’s Hospitalist.

Hospitalist-centered CME on pneumonia and more
COMMUNITY-ACQUIRED PNEUMONIA will be one of the six topics covered at the Fall 2005 Hospitalist CME Series. Scott Flanders, MD, associate professor of medicine at the University of Michigan and director of the University of Michigan’s hospitalist program, will present an update on CAP that focuses on recent studies, appropriate treatments and key processes of care. More information is available online.

Published in the April 2005 issue of Today’s Hospitalist