When it comes to blood clots in surgical patients, "vigilant hospitals get penalized because [when] they look more, they find more. Their rankings look atrocious," says the author of a report blasting a patient safety measure that he says is driving unintended consequences.
|
Quality leaders at 227 hospitals that according to Medicare data have "worse than U.S. national rates" for serious blood clots in their surgical patients" might have a plausible defense this week.
That's because a research paper in an influential medical journal has blasted PSI-12—a patient safety initiative that tracks preventable venous thromboembolism (VTE) such as deep vein thrombosis and pulmonary embolism—for being Exhibit A for the law of the unintended consequence.
Many of the hospitals with high rates look bad, not because they truly are; quite the opposite. They look bad because they're diligently looking for blood clots harder than other hospitals, and thus they are finding more often.
The "worse-than" label applied by the Centers for Medicare & Medicaid Services doesn't necessarily mean these hospitals are providing poorer quality care, or failing to safeguard their patients, the authors say.
"The problem is that not everyone looks for blood clots in the same way; not everyone is as vigilant," explains Karl Bilimoria, MD, director of Northwestern University's Surgical Outcomes and Quality Improvement Center at the Feinberg School of Medicine in Chicago and author of this week's report in theJournal of the American Medical Association.
"The vigilant hospitals get penalized because they look more, they find more, and their rankings look atrocious, and that's why this measure really is unfair," he says.
At least two unintended consequences are encouraged by this measure, he says.
First, some hospitals "may actually go in another direction, and start giving chemical (anti-clotting medication) prophylaxis to really low-risk patients, where the risk/benefit ratio doesn't favor that. That may increase the risk of bleeding and other complications.
"We don't want to do things that are outside the guidelines simply to address a faulty measure," Bilimoria says.
Second, "the worst thing is if a hospital's physicians think, 'Oh, I can drop my rate for VTE by looking less,' then they'll miss one, they won't get away with it, and that will be a deadly event."
In a nutshell, the problem is this: hospitals do imaging studies to look for blood clots in patients, even when they're asymptomatic. So hospitals that do more imaging studies find more blood clots than hospitals that don't do as many.
Some hospitals with trauma or burn units, or large orthopedic and cancer services, where surgical patients may be at higher risk for blood clots, and certainly academic medical centers, may look for clots in every patient after surgery, performing ultrasound or CT imaging tests to screen for any potential blood clot that could get into the lung and impair breathing.
Bilimoria and colleagues at Northwestern University's Feinberg School of Medicine who did this research believe the measure is so seriously flawed, that CMS should not be publicly reporting rates by hospital.
"The fact that it's being publicly reported should bother people," Bilimoria says.
Nor should it be using PSI-12 as a pay-for-performance measure, though it serves as such under authority of two sections of the Patient Protection and Affordable Care Act, he says.
Starting in less than one year, hospital reimbursement for Medicare patients will be adjusted based on how well they do on PSI-12 in two ways:
- Through the value-based purchasing incentive program (it is part of the AHRQ PSI-90 set of eight measures) and
- The new hospital-acquired condition penalty, where again it is embedded in PSI-90.
Though PSI-12's weight in each provision is virtually negligible, at most just a few cents for every $100 of Medicare payment, Bilimoria says that's not the point.
"Hospitals have wasted numerous years and millions of dollars trying to chase down their VTE problem and fix things around VTE, when actually they're being presented with flawed data," he says.
Indeed, the data for "serious blood clots after surgery," shows Northwestern Memorial, an academic medical center, with the 23rd highest rate of reportable blood clots among 3,094 hospitals whose VTE rates were reported by CMS. The Chicago facility had 11.85 VTE cases per 1,000 patient discharges.
"That is exactly the point," Bilimoria says. "Of course Northwestern Memorial would have a high VTE rate, because we are very vigilant in looking for VTE. The measure inevitably singles out organizations that are more vigilant in screening their patients for blood clots, but labels them as being worse for it."
Bilimoria says that hospitals should be scored instead on process measures to prevent clots. For example, does the hospital use appropriate screening tools to make sure the right patients get anti-clotting medications or mechanical prophylaxis, such as compression stockings?
This, however, is not yet in hospital pay-for-performance algorithms, and I'm not aware of any plans to put it there.
So if PSI-12 is so flawed, how did it get to be a performance measure? It was endorsed by the National Quality Forum in 2008, and was quickly adopted by the American College of Surgeons' National Surgical Quality Improvement Program (NSQIP) and by UHC, a performance improvement organization serving 118 academic medical centers, in their quality endeavors.
Apparently there's been little attempt to look back.
Greg Maynard, MD, director of the University of California San Diego Medical Center's Center for Innovation and Improvement Science and a VTE researcher, says he's been working with officials from the Centers for Disease Control and Prevention and with the Society of Hospital Medicine to get the measure amended or un-endorsed.
"We have sent letters to the National Quality Forum saying this measure for outcomes, and a measure for DVT prophylaxis, are not very good and we urged them to make changes," he says. "But they did not make those changes. They sent a response saying, 'we know we're not perfect, but we need to raise the bar.' "
Maynard adds, "to some extent, (the NQF and other PSI-12 defenders) generally make the argument that they raise awareness, and they get people to think about it. It's hard to argue with that. But to withhold payment based on that these (PSI-12) rates, as opposed to just looking at an increase or decrease within one institution, that's fraught with difficulty."
He adds that there are just too many factors that skew the results for any one hospital when it's compared with other hospitals across the country, such as the way hospitals code illness, other testing some hospitals do, the mix of patients with cancer or trauma, and many other variables.
Maynard says Bilimoria's paper "has made a very valid point, and raised awareness beyond what we've been able to do just by saying these measures aren't very good."
Asked for its response to the Bilimoria paper, specifically why it continues to back a quality measure about which providers are so strongly skeptical, Erin Reese, a spokeswoman for the National Quality Forum, replied in an e-mail:
"Any party may request an ad hoc review of any measure as long as there is adequate evidence to justify the review," for example the evidence has changed, implementation results in unintended consequences or material changes have been made to the measure.
"That said, the ad hoc review is available to parties who wish to request one, and the measure you reference seems to fit the criteria stated above."
I had mixed feelings about Bilimoria's paper. My mother suffered a pulmonary embolism while sitting in the lobby of a South Florida cancer treatment center, waiting for her second dose of chemotherapy. Though a code was called and she was revived, she did not survive that event.
No one looked for the possibility she might have a blood clot, though an exploratory surgical procedure she'd had a few weeks earlier put her at even greater risk.
But to have hospital physicians chasing their tales and gaming the system, overtreating some or ignoring risk to escape financial penalty and public shame would not solve this problem. And it certainly wouldn't save people like my mom.