Published in the December 2008 issue of Today’s Hospitalist
DOCTORS LIKE TO THINK they practice medicine based on evidence. But a recent study has found otherwise, at least when it comes to ordering elective percutaneous coronary intervention (PCI) for patients with stable coronary artery disease (CAD).
“There is a bias toward intervention,” which holds true for hospitalists and internists as well as for cardiologists, says Grace A. Lin, MD, a general internist at the University of California, San Francisco. Dr. Lin was the lead author on a study published in the Oct. 15 issue of Journal of the American Medical Association (JAMA). That study found that less than half of patients nationwide underwent stress testing at least 90 days before getting an elective PCI. National guidelines, however, call for documenting ischemia first through stress testing.
Physicians exhibit, says Dr. Lin, a bias toward what she and her colleagues call “oculostenotic reflex,” meaning that if physicians see a lesion, they will fix it, even if the “fix” won’t help the patient. Doctors also, she adds, are predisposed to trusting their own judgment more than a clinical guideline.
However, doctors exhibit those tendencies differently depending on where they live. The study found wide variation in the rate of stress testing prior to elective PCI according to geographic region. Those rates ranged from a low of 22.1% to a high of 70.6%.
“This really drives home the point that local practice and where patients go for care matters a lot,” Dr. Lin says. “This is somewhat troubling, because you would think that where you live shouldn’t affect you that much, particularly when there is a guideline.”
Even regions that border each other vary widely. In San Francisco, for example, 46% of patients going for elective PCI had stress testing before angioplasty, while right across the bay in Oakland, the rate was 70%. In nearby Fresno, Dr. Lin says, the rate was 20%.
“We know the rate shouldn’t be 100%,” she explains. “There are clearly situations where someone’s symptoms are so clear and they have been treated previously with maximum medical therapy, or they can’t get a stress test for some reason. But less than half is probably too low.”
Those wide variations, which had no corresponding differences in the population being treated, the availability of the procedure or of testing, or in basic hospital characteristics, raise a red flag about potential under- and overuse of PCI.
Gender, racial variations
According to Dr. Lin, the study found that women were less likely to have stress tests ordered before PCI than men. Researchers also found that blacks were more likely than whites to have stress testing before PCI. (Because the study was based on 2004 Medicare data, it didn’t look at insurance status.)
The researchers found that specialty training made no difference in doctors’ decision-making. But physicians’ age did: Patients of younger physicians (less than 40 years old) were more likely to get a pre-PCI stress test than those being treated by physicians between ages 50 and 69. Patients of the most senior doctors, those older than 70, were the most likely to have a stress test ordered.
“You have to wonder if physicians in the middle-age category are so comfortable with their own clinical judgment that they don’t necessarily feel the need to have an objective test,” Dr. Lin says. “This all makes us think that there are other things affecting doctors’ judgment than just the patient and the guideline.”
The affect of nonclinical factors
While most of the tests and procedures in the study were ordered by cardiologists, Dr. Lin says that internists, some of whom were hospitalists, were just as unlikely to order pre-PCI stress testing.
Because the study used AMA data to discern physician characteristics, it did not identify hospitalists per se. But Dr. Lin says that hospitalists are getting more involved in managing these patients both as admitting physicians and in comanagement arrangements with cardiologists. The study did find that physicians with the highest PCI volumes are less likely to use stress testing.
Dr. Lin has been trying to parse out why doctors seem to trust their opinion more than the evidence as it pertains to PCI. In other studies (one published in the Aug. 13/27, 2007, issue of Archives of Internal Medicine and another in the Sept. 23, 2008, Journal of General Internal Medicine), Dr. Lin and her colleagues explored physicians’ preference for performing elective PCI in patients with stable coronary artery disease, despite evidence that the procedure benefits such patients only minimally.
What they found was that “nonclinical factors appear to have substantial influence on physician decision-making” and often trump evidence-based guideIines. “Psychological and emotional factors” often prevailed “over evidence from clinical trials,” the authors concluded in the Archives study.
Specifically, that study found that doctors believe in the benefits of PCI, particularly in the value of drug-eluting stents. They also think patients want a procedure, and they fear being sued for malpractice if they don’t order PCI. They also express what Dr. Lin calls “anticipated guilt,” meaning that they would regret it if something bad happened to a patient for whom they hadn’t ordered PCI.
“A lot of physicians had a firm belief that stenting works and improves outcomes,” Dr. Lin explains. That belief persists despite evidence that angioplasty doesn’t prevent heart attacks or deaths in patients with stable coronary artery disease any more than medical therapy.
“They feel that because it opens up the blockage, it must be better.” Physicians also find it difficult, she adds, to interrupt a natural “cascade” of medical interventions that starts with the first test ordered for CAD.
The question of overuse
While the main reason to add PCI to optimal medical therapy is to reduce ischemia, Dr. Lin points out that cardiac catheterization and angioplasty are not risk-free. Downsides include the need for repeat revascularization, thrombosis and “a decreased quality of life if performed in patients with minimal symptoms,” according to the authors in JAMA. Stress testing prior to PCI, on the other hand, is linked to lower overall costs, shorter hospital stays and lower rates of revascularization.
Nonetheless, Dr. Lin and her colleagues wrote, PCI for stable patients now accounts for “the majority of PCIs performed” in the U.S. PCI has also become a “common treatment strategy for patients with stable CAD.”
With Medicare spending between $10,000 and $15,000 per PCI and with PCI accounting for at least 10% of the increase in Medicare spending since the mid-1990s, the study’s authors concluded that this area is ripe for investigation into how to cut spending for potentially inappropriate procedures. The ultimate carrot to sway physician decision-making, wrote Dr. Lin, may be restructuring Medicare payments “to reward hospitals and physicians who adhere to guidelines.”
Deborah Gesensway is a freelance health care writer based in Sierra Madre, Calif.