| NEW YORK
NEW YORK (Reuters Health) - Publicly reported statistics about how patients fare after a common heart procedure often are inaccurate, says a new study that suggests publishing the poor quality numbers may do more harm than good.
The proportion of patients who die within 30 days of a medical procedure is sometimes used to gauge the effectiveness of the treatment, or of the hospital or doctor administering it.
After analyzing thousands of patient records at one U.S. hospital, researchers found that less than half of the deaths following percutaneous coronary intervention (PCI) - also known as coronary angioplasty - were related to the procedure.
That's far fewer than the number of deaths usually attributed to the surgery, which involves clearing a narrowed artery then inserting a mesh tube called a stent to prop the vessel open.
"I think what this really points out is that (with) the clinical complexity of medicine (it) can be tough to get a policy that's sort of one size fits all," said Dr. Karen Joynt, a cardiologist at Brigham and Women's Hospital in Boston.
Joynt, who was not involved with the new study but has done similar research, said there are a number of factors that determine whether or not a person dies after PCI - many of them unrelated to the actual procedure.
About 500,000 angioplasties are performed in the U.S. every year. It's often used in emergency situations, for instance to stop heart attacks caused by blocked blood vessels.
New York State first reported the number of people who died after PCI in its hospitals in 1991. Since then, Massachusetts and Pennsylvania started to report that figure as well.
According to the new study's authors, the U.S. Centers for Medicare and Medicaid Services have also proposed using the number as a quality measure, in the hope of improving patient care.
For the new study, researchers led by the Cleveland Clinic's Dr. Mehdi Shishehbor used patients' medical records to see whether deaths following PCI could be blamed on the procedure.
Out of 4,078 PCIs performed between January 2009 and April 2011 at one U.S. hospital, they found 81 deaths within 30 days of the surgery, representing 2 percent of patients who had the procedure.
Forty-seven, or about 60 percent, of those 81 patients died of a heart-related problem, but only 34, or 42 percent, of the deaths resulted from a complication of the procedure itself.
Previous research has attributed up to 60 percent of deaths following PCI to the procedure, but those figures were often based on causes of death listed on death certificates, which are frequently inaccurate, according to Shishehbor and his colleagues.
In the current study, based on close reading of every patient's chart, many of the deaths were among patients who were very sick when they got to the hospital and might have died with or without PCI, the researchers report in the Journal of the American College of Cardiology.
Fifty eight percent of the patients who died arrived in cardiac arrest, cardiogenic shock or with the most lethal type of heart attack, the authors point out.
But Joynt said doing PCI on these patients may have given them a better chance of survival, even though they ultimately died.
"That's the kind of overuse (of PCI) that we're probably willing to accept in many cases," she said.
However, Joynt's own research has suggested that hospitals avoid doing PCI in the sickest patients when their states start reporting the number of 30-day deaths.
In 2012, her study found that hospitals in Massachusetts, New York and Pennsylvania performed fewer PCIs - especially among the sickest patients who could possibly have benefited the most.
"Public reporting is a good thing and needs to be done, but when done incorrectly there are a lot of downstream consequences, such as denying patients treatment," said Dr. Duane Pinto, head of interventional cardiology at Boston's Beth Israel Deaconess Medical Center, who wrote an editorial accompanying the new study.
"We have to recognize and adapt to the unintended consequences of public reporting and fix them when we identify them," Pinto said.
SOURCE: bit.ly/10L9DPz Journal of the American College of Cardiology, online May 7, 2013.