TFAD and JCAHO: What the Data Shows
Here's a familiar scenario: an elderly patient comes to the ER with symptoms that include a cough, abdominal pain, and confusion. It could be pneumonia, but the attending doctor's unsure; other symptoms suggest chronic obstructive pulmonary disease, or even heart failure. Meanwhile, the clock is ticking—the Centers for Medicare and Medicaid Services (CMS) measure how many pneumonia patients get antibiotics within four hours of hospital admission. And low CMS scores can influence the hospital’s reputation, not to mention looming pay-for-performance allocations under Medicare.
Faced with that situation, many doctors simply “shoot first and ask later,” meaning they give antibiotics for pneumonia before confirming the diagnosis. But while that might boost CMS performance measures, it's not necessarily good medicine. “It's inappropriate to give antibiotics if you don't know what you're giving them for,” says Mark Metersky, MD, a professor of medicine at the University of Connecticut School of Medicine. Not only does that accelerate antibiotic resistance, he explains, but it can also complicate treatment. Some patients have allergic reactions to antibiotics, while one in five develop antibiotic-induced diarrhea; a potentially serious condition.
CMS and the Joint Commission on the Accreditation of Healthcare Organizations (JCAHO) agree. Citing evidence that some cases can't be diagnosed in four hours, both organizations have revised data reporting requirements for “time to first antibiotic dose,” or TFAD, for patients with community-acquired pneumonia. Last July, they recommended that antibiotic treatment should be contingent on chest X-ray findings consistent with pneumonia. And beginning in June, CMS and JCAHO will collect six-hour TFAD data, in addition to the four-hour data they already collect now. A new CMS data element for “diagnostic uncertainty,” will also allow doctors to provide clinical reasons for delaying antibiotic treatment.
Dale Bratzler, DO, MPH, Clinical Coordinator for the Oklahoma Foundation for Medical Quality, a CMS contractor, says the six-hour extension could take some pressure off doctors forced to make fast decisions. “We're trying to avoid the unintended consequence of blindly giving antibiotics to anyone who comes in with a fever and a cough,” he says.
New Data Show Four Hours Not Enough
The revised measures respond in part to new data published by Metersky in the July issue of Chest. Metersky reviewed charts from 86 Medicare patients with a discharge diagnosis of pneumonia. He found that among them, 19 presented with symptoms that could delay diagnoses beyond the requisite four-hour time frame. Those symptoms included altered mental status and findings that were also suggestive of heart failure or COPD. Diagnostic uncertainty in some patients could also be due to the absence of rales (abnormal chest sounds) or oxygen desaturation; which are both common in pneumonia.
Metersky says his findings validate wide-spread complaints among doctors that the four-hour window is too short. “We’d been hearing a lot grumbling to that effect, but until now we didn't have the corresponding data,” he says. “I'm confident the results are applicable across all Medicare patients. The precise percentages may change, but the overall numbers of cases that require more than four hours to diagnose is significant.”
The Scientific Rationale
Pressure to shrink the TFAD stems back to the early 1990s. At that time, mounting evidence was showing that Medicare patients with pneumonia had better survival if antibiotics were given as soon as possible. Peter Houck, MD, a former director of the CMS' National Pneumonia Project (NPP), says the most convincing evidence came from large cohorts of patients aged 65 years or more. For instance, Thomas Mehan, et al. studied 14,069 hospitalized Medicare patients, and identified a 15% reduction in 30-day mortality with a TFAD of 8 hours. Those results were published in the December 17, 1997 issue of the Journal of the American Medical Association.
Other studies linked improved survival to even shorter TFAD. Khan et al., reported in JAMA on October 17, 1990, that survival among Medicare pneumonia patients improved if they were treated with antibiotics within four hours of admission.
These cumulative findings led Houck's group to recommend that hospitals treat the elderly for pneumonia with antibiotics faster. Taking that cue, CMS and JCAHO in 2001 both recommended a TFAD of eight hours. JCAHO incorporated compliance with the measure into its hospital accreditation processes, while CMS applied it to its own quality reporting procedures.
In 2003, both organizations reduced the measure from eight hours to four. According to Houck, that change resulted in part from his own reanalysis of an enormous dataset containing more than 18,000 Medicare patients. That investigation, published in a March 22, 2004 paper co-authored by Bratzler in Archives of Internal Medicine, showed a TFAD of four hours could significantly reduce mortality among elderly patients who hadn't yet received outpatient antibiotic treatment before hospital admission.
Not everyone agreed with those findings, however. In fact, two scientists have consistently challenged the notion that TFAD and pneumonia outcomes are correlated. Grant Waterer, MD, FCCP, from the University of Western Australia School of Medicine, and Richard Wunderink, MD, FCCP, from the Feinberg School of Medicine, in Illinois, argue that pneumonia patients with delayed treatment typically have altered mental status, or early signs of sepsis. Those patients would probably die regardless of TFAD, they say, which suggests the relationship with clinical outcomes doesn't exist. Their conclusions are published in the July, 2006 issue of Chest.
However, Houck (who wrote an editorial that appears in the same issue) and Bratzler counter that their own findings hold up even after adjusting for mental status. Moreover, Waterer et al's conclusions suffer from a lack of statistical power, they add. Only 451 patients were studied in their most recent analysis. And of these, just 158 were above the age of 65, which is the only age group for whom significant links between TFAD and mortality have been shown.
Houck stresses that because TFAD probably makes a small contribution to survival, it can only be detected in cohorts with thousands of patients. But he concedes the true contribution will likely never be known. “To determine that, you'd need a randomized, prospective trial of the effect of timing on outcome,” he says. “But no one's going to do this. For ethical reasons, you can't have a study that intentionally delays antibiotics among some patients. Furthermore, no company would ever pay for a study that delays administration of their drug. So, we're left with observational data, and our job is to try to make sense of it.”
The Stretch to Younger Adults
Thus, observational data drive TFAD reporting that, ironically, now applies to all adults of all ages, including those less than 65. Houck insists he's never been willing to argue for the inclusion of younger patients in the measure. So why must hospitals report them? “One could make the case that it's a more consistent and efficient approach,” Houck proposes. “And we at CMS were instructed to make the measure consistent with JCAHO down to the syllable, and JCAHO wanted it to apply to those over the age of 18. At the time, we didn't think it mattered too much since there are relatively few younger Medicare patients anyway.”
Bratzler adds that for many other infections, quick antibiotic treatment produces better results, regardless of age. Patients with septic shock, he says, do better when treated with antibiotics quickly, and so do those who develop pneumonia in the hospital. “It made sense that the same would apply to community-acquired pneumonia,” Bratzler explains.
What neither scientist anticipated was TFAD's emerging role in public reporting and reimbursement. But that role has since proven significant: In March, the Chicago Sun Times reported that local pneumonia patients were getting care that lagged behind the rest of the country. That highly-publicized conclusion was based on performance against the four-hour TFAD, which among some Chicago-area hospitals was close to zero. The current national average, meanwhile, is 77 percent, according to Hospital Compare, a DHHS-funded website that ranks hospital performance on CMS measures. Hospital Compare has become popular with health care consumers, who use it to choose care facilities according to their compliance with CMS performance measures.
Ironically, it's not year yet clear if performance on the six-hour measure will even show up on Hospital Compare. That's because the National Quality Forum, a non-profit organization that advises the website's administrators on content, has yet to endorse the measure. Sources from the National Quality Forum could not be reached for comment.
Meanwhile, Metersky suggests the revised measure doesn't go far enough. “I think six hours is fairly arbitrary,” he says. “It will remove some of the pressure about whether to give antibiotics or not, but it doesn't really address the issue.”
In Metersky's view, CMS needs to lower the performance target from 100 percent to 75-80 percent. Hospitals today compete for reimbursement under CMS reporting schemes by striving for full adherence with performance measures, he explains. And that makes antibiotics low-hanging fruit—doctors shooting for high performance numbers have a strong incentive to prescribe them even without a definitive diagnosis. Lowering the target would reduce that temptation, he suggests, while encouraging timely antibiotic use.
But Jarod M. Loeb, PhD, JCAHO's executive vice president for research, counters that performance targets under 100 percent can be problematic. “That gets into an area we call ”appropriateness,“ he explains. ”It's hard to second-guess when a given treatment is appropriate because medicine is both a science and an art. We don't have much experience setting policy on something like this—if you put ten people in a room, you'll get eleven opinions on how to do it.“
Neither Loeb nor Lisa Buczkowsky, JCAHO's clinical lead on pneumonia, would rule out the possibility of lowering performance targets, however. The trick, they say, is to set targets—or even target ranges—that allow for punitive measures when missed, and that point to clear opportunities for improvement. “This is under discussion,” Loeb says.
Looking forward, TFAD’s role on public reporting and reimbursement seems likely to increase. Geoffrey Page, from the quality improvement department at Mount Sinai Hospital in Chicago, IL, says chief medical officers are keenly aware of a growing transparency in health care. Mount Sinai recently participated in a CMS demonstration project that showed the hospital's adherence with TFAD reached barely 50 percent—an alarming figure, Page says. “Our CMO said 'we have to fix this,'” Page recalls. “There was some real feet to the fire from that. Pay-for-performance is coming, and if you fail on these measures you won't get your full Medicare increase. And that's a real concern.”