Home Q&A The case for faster treatment for community-acquired pneumonia

The case for faster treatment for community-acquired pneumonia

May 2004

Published in the May 2004 issue of Today’s Hospitalist

Peter M. Houck, MD, admits that when he first read that giving pneumonia patients an antibiotic within four hours of arriving at the hospital reduced mortality rates, he was skeptical. It was the early 1990s, and a relatively obscure journal was touting the benefits of treating pneumonia much more quickly than usual.

Over the years, a body of research has gradually made the case for speeding up treatment for community-acquired pneumonia (CAP). Organizations like the Infectious Diseases Society of America have endorsed the idea, urging physicians to treat the condition as soon as possible.

The notion of a four-hour window, however, has its share of critics. Some say it’s simply too difficult to administer antibiotics to CAP patients within four hours of their arrival at the hospital. Others worry about the overutilization of antibiotics.

That’s where Dr. Houck comes in. As chief medical officer for the Seattle region of the Centers for Medicare and Medicaid Services and clinical lead for the agency’s infectious disease quality improvement projects, he wanted to examine whether treating CAP inpatients within four hours really made a difference in areas like mortality and length of stay.

In a study published in the March 22 Archives of Internal Medicine, Dr. Houck and his colleagues examined data from nearly 14,000 Medicare patients. They found that among the 76 percent who had not received antibiotics in the outpatient setting, administering the drugs within four hours appeared to reduce factors like inpatient mortality (6.8 percent vs. 7.4 percent), mortality within 30 days of admission (11.6 percent vs. 12.7 percent), and length of stay exceeding the five-day median (42.1 percent vs. 45.1 percent).

In addition, mean length of stay was 0.4 days shorter for patients who received antibiotics within four hours of arriving at the hospital.

In an interview with Today’s Hospitalist, Dr. Houck talks about the research and its implications for hospitalists.

The timing of antibiotics in treating CAP has been a topic of discussion for several years. What does your study add to the debate?

Our study’s results push the administration time at which a statistically significant patient benefit has been demonstrated down from eight hours to four hours. For the Medicare population this has important implications, because pneumonia is the second-leading cause of Medicare hospitalization.

Some people have said that four hours is too short a time frame in which to administer antibiotics, that it’s impractical and unrealistic. But our study found that just under 61 percent of these patients received an antibiotic within four hours of arriving at the hospital, which we defined as the first time the patient’s presence was noted by hospital staff.

The point we’re trying to make is that many patients are getting the medicine within four hours, so it’s not a totally impractical goal. The vast majority of hospitals could administer antibiotics in at least half their patients within four hours, so we’re not talking about something that’s impossible to do.

What are some of the obstacles that hospitals face?
We recognize that many hospitals, particularly the larger urban teaching hospitals, will have more of a challenge. In the studies of inpatient pneumonia treatment we’ve done since 1993, we’ve typically found that smaller hospitals tend to do better on timing than larger facilities.

In smaller rural hospitals, the pharmacy is often within 30 or 40 feet of the emergency department, so it’s often very easy to get the drugs you need. In a big hospital, the pharmacy may be quite some distance away.

What are some typical issues that slow down the administration of antibiotics?

In some of our other research, we’ve been surprised at how many physicians do not order the first dose of drug to be given “stat,” who order ceftriaxone one gram “every 24 hours” or “daily.” If the patient happens to arrive on the floor and just misses the nursing unit’s “Q24” or “Q day” administration time, the patient could easily go 23 hours before receiving any antibiotics.

We’ve heard that a number of hospitals have changed their policies so every first dose of virtually any drug is considered a stat order. That may help improve the delivery of antibiotics to these patients.

Another issue comes up with the movement of patients in the hospital. The patient is in the emergency department when the drug is ordered but then moves to the floor before the drug is actually administered. The drug finally catches up to the patient on the floor, but it’s like a game of cat and mouse where the drug is chasing the patient around the hospital.

What are some solutions to get antibiotics to CAP patients faster?

Roughly 25 percent of CAP patients are admitted directly from an office or nursing home. The outpatient physician treating these people could give the first dose of the drug before the patient goes to the hospital. If the physician using this approach also wants blood to be cultured, the specimen should probably be collected in the outpatient setting as well, because the yield of cultures performed in the hospital would be reduced by the outpatient antibiotic.

Patients who go through the emergency department could receive their first dose of the drug in the emergency department before they are sent elsewhere in the hospital.

One issue we have heard about anecdotally is that emergency physicians sometimes don’t want to write orders for drugs until the attending from either the hospital service or the community practice writes the orders or says what he or she wants.

Hospitalists could work with emergency department physicians to develop a care pathway or standard CAP orders that are acceptable to the hospitalist and the emergency department physician. Hospitalists could ask that all CAP patients receive their first dose of antibiotic before being sent upstairs and use a certain protocol to select the drug.

Was timing as critical when patients had already received an antibiotic before going to the hospital?

Among the roughly 24 percent of CAP patients who had received some kind of prehospital antibiotic, our findings were not nearly as interesting. There was still an apparent significant reduction in length of stay when these patients received an antibiotic within four hours, but we could not detect any differences in mortality. With the exception of the difference in length of stay, we don’t have any data to say that it’s critically important to treat these people within four hours.

That said, we can think of these people as having failed their outpatient treatment. Maybe they received the wrong antibiotic, maybe they have a drug-resistant organism, but it’s still important for them to get an effective antibiotic. It makes intuitive sense that sooner is probably better than later.

Is there a benefit to giving antibiotics to CAP patients even faster than within four hours of their arrival at the hospital?

We found that that if you eliminate the people who received some kind of treatment before coming to the hospital, we could detect a significant benefit of giving antibiotics as quickly as three hours after arrival. For this study, however, we thought it was reasonable to choose four hours to actually give people time to evaluate a patient and start treatment. It’s also a time that was already being used by many hospitals, although some have two hours as a goal.

We probably won’t see recommendations for giving the medicine faster than four hours. You get into concerns about antibiotic utilization and whether you’re giving patients antibiotics without fully examining them.

What do you say to physicians who remain skeptical about the issue of timing?

We recognize the limitations of an observational study, so I would point out that we tried to make sure that the association between the timing of antibiotics and factors like mortality and length of stay was not the result of some kind of confounding factor or bias in the data.

For example, in some additional analyses we excluded all patients who had evidence of congestive heart failure, thinking that it could look like pneumonia and delay the diagnosis and increase mortality rates. When we eliminated those kinds of patients, we found the association was still there.

Despite my initial skepticism, I’m pretty comfortable that when you look at large populations of patients, there is an association between the timing of antibiotics and outcomes in CAP patients. When you look at an individual patient, by and large it probably won’t matter. But when you look at the 600,000 Medicare admissions for pneumonia a year, you’re probably going to save a lot of lives.