Best and Worst Hospitals for Cardiac Cases Revealed

For the first time, a government study publicly unveils death rates by hospital.

BySteve Sternberg and Anthony DeBarros

Aug. 20, 2008— -- Motorists heading through the Lehigh Valley from Allentown, Pa., earlier this year passed two giant billboards proclaiming: "Fast Heart Attack Care Saved My Husband's Life."

What the billboards didn't say was just how fast. It took 24 minutes for Richard Silverman's doctors at Lehigh Valley Hospital to clear a 100% blockage from his heart's most vital artery. That's a third of the 90-minute goal that hospitals strive for.

FULL MAP: Compare hospitals near you on heart attack, heart failure and pneumonia

STANDOUTS: Hospitals with highest and lowest death rates

ON THE WEB: Compare hospitals in your ZIP code

"Maybe five minutes more and I'd be gone," Silverman, 63, co-owner of Pro-dent, a dental laboratory in Allentown, says his doctor told him.

Doctors at Lehigh Valley are proud of their speed. It's one reason the hospital boasts the lowest heart attack death rate in the country, 11.6%, in a new government analysis obtained by USA TODAY. Among those at the other end of the spectrum is Virginia's Danville Regional Medical Center with death rates for heart attack of 19.6% and for heart failure of 15.5%.

Until today, hospital death rates were closely guarded secrets, discussed in board rooms but beyond the reach of patients whose lives are on the line. That changed this morning when USA TODAY posted on its website the government's best estimates of heart attack, heart failure and pneumonia death rates for every U.S. hospital for two years.

Now anyone with access to a computer can directly compare a local hospital with the one across town to see how it stacks up against the biggest medical institutions nationwide.

Death rates from heart attack, heart failure and pneumonia are widely viewed as yardsticks of a hospital's overall performance.

"We're in an era of change at last," says Donald Berwick, president and CEO of the Institute for Healthcare Improvement, a non-profit in Cambridge, Mass., that works with hospitals to improve care and eliminate errors.

Last year, the U.S. Centers for Medicare and Medicaid Services (CMS) released a broad comparison of death rates for heart attacks and heart failure, noting how hospitals compared with the national average — better, worse or no different — without releasing the death rates themselves.

This year the agency decided to disclose them to consumers.

The agency shared the information in advance with USA TODAY to reach the widest possible audience. The agency also posted its new mortality estimates on a government website (, along with more than two dozen other measures of how well hospitals meet patients' needs.

Among them are statistics on what percentage of a hospital's patients get appropriate care for a variety of ailments, including childhood asthma, and 10 measures of patient satisfaction with the hospital experience.

All three types of measurements give hospitals ways to assess — and improve — their quality of care, but many health officials regard the number of patients who die in the hospital or soon after discharge as the ultimate measure of performance.

"That's why we think this is so meaningful," says Barry Straube, the chief medical officer of CMS.

Knowing a hospital's death rates also gives consumers more power to influence the quality of their medical care, says Lisa Iezzoni, associate director of the Massachusetts General Hospital Institute for Health Policy.

"What the mortality rate does is give you an entree to talk to your doctor and say, 'Look, is this hospital stay going to kill me?' "

That's not an easy question to answer with any certainty.

By trying, officials knew they were courting trouble with the hospital industry, Iezzoni says. An earlier effort by Medicare to report on hospital death rates faltered in the early '90s.

The agency wilted under relentless criticism that its so-called death list didn't give adequate weight to a hospital's mix of patients, including how sick, poor and old they were.

This time, the architects of the new analysis took a different approach. They tallied death rates for common life-threatening conditions, not the hospitals' overall mortality rates. And they chose a strict statistical formula that allows them to say with 95% confidence that a hospital's death rates fall within a certain range.

For Lehigh Valley, the range was 9.3% to 14.4%, with 11.6% representing the best estimate of the heart attack death rate.

But there's a rub, experts say. Using this method of analysis, only a handful of hospitals stand out as better or worse than the national average.

"There's great conservatism in calling people better or worse," says Yale cardiologist Harlan Krumholz, who helped develop the approach, adding that Medicare "got burned in the past."

Leah Binder, CEO of Leapfrog, a consortium of Intel, Boeing, Marriott and other corporations aiming to lower health costs to their employees by improving the quality of care, says employers want much more information than what the government is willing to provide.

"The problem with the CMS data is that most hospitals look average, which isn't what employers want. What they want is to compare hospitals." She says Leapfrog hopes to have its own rating system by this fall.

Berwick, of the institute, has studied the "death list" controversy. He says all the politics around measurement add up to an effort to protect hospitals against false alarms.

"It's protect the hospital or protect the patient," he says. "You can't have it both ways."

The new formula, developed by teams led by Krumholz and Harvard's Sharon-Lise Normand, captures all deaths among 35 million Medicare beneficiaries that occurred within 30 days of the patients' hospital admission.

They also factored in the hospital's patient mix and how many deaths might be expected in a hospital with that population. By tracking deaths from all causes, Krumholz says, the agency was less likely to miss a death related to hospital quality.

Including deaths that occurred within 30 days after admission made it tougher for a hospital to game the system by shipping risky cases somewhere else.

John Rumsfeld of the Denver VA Medical Center helped evaluate the approach for the non-profit National Quality Forum. He says the method makes it hard for hospitals to complain they've been misjudged.

"When your numbers aren't what you like them to be, it's not because you didn't have the same patients as the hospital across the street," Rumsfeld says. "It's because you didn't do as well as you could with the hand you were dealt."

With all its imperfections, Berwick says it is a useful tool.

"This is turning the lights on for providers of care, as well," he says. "Doctors and hospitals can't decide to do better unless they know how they're doing."

For the doctors at Lehigh Valley, the new analysis confirms what virtually everyone knew. "We have a culture of excellence," says Michael Rossi, chief of cardiology.

The hospital decided to treat heart attack patients by clearing their arteries with a balloon long before it became the gold standard. The hospital's chief quality officer, Anthony Ardire, says he routinely briefs board members on infection rates, pressure ulcers and scores of other benchmarks of patient care.

"If we fall back a bit, or if we're not making progress," Ardire says, "they want to know what we're going to do about it."

The key to the hospital's success is teamwork, a willingness to accept criticism and attention to every detail, he says. "There's no one thing that determines whether you save a life."

Consider the Silverman case. Just a day earlier, the hospital had given paramedics the power to diagnose heart attacks in the ambulance using an electrocardiogram. Doctors in the emergency room confirmed the diagnosis and rushed Silverman upstairs for angioplasty. What they hadn't counted on was the wait at the elevator doors.

"We learned that if we lock down our elevators when we know a patient's coming, we can save three or four minutes," says Nainesh Patel, the doctor who cleared Silverman's artery.

At Cedars-Sinai Medical Center in Los Angeles, patients with pneumonia are similarly fast-tracked to treatment. A triage nurse sends those with symptoms for a chest X-ray without waiting for a doctor. Those who are positive are quickly placed on antibiotics, says James Loftus, co-chair of emergency medical care. The result is a death rate of 7.5%, one of the lowest in the USA.

Sometimes even hospitals known for their excellence discover problems, including Baylor Health Care system, winner of this year's National Quality Health Care Award.

When last summer's CMS report came out, one of the 11 hospitals in the Dallas-based system, Baylor All-Saints Medical Center in Fort Worth, was found to have a heart failure death rate of 14.6%, higher than the 11.1% average.

What leapt out of a review of the patients' records was that just 10 of 31 deaths occurred in the hospital, suggesting that some deaths were due to follow-up care by local doctors and nursing homes, says Paul Convery, Baylor's chief medical officer. "This was a signal that we have to be responsible for patients after they've left our halls."

Finally, there are 115 hospitals, such as Danville (Va.) Regional, that have been singled out for having higher death rates than the national average. Danville last year ranked higher than the national average in both heart attack and heart failure. In the new report, Danville has improved. Although its heart attack death rate is still high, it now falls within the average range.

Michael Moore, the hospital's chief of medical education, says the numbers don't account for the poverty and lack of education pervasive in southern Virginia. Patients with heart attacks don't seek care quickly enough, he says, while those with heart failure don't follow doctors' orders.

"We take these reports very seriously," he says. "We continue to work to try to improve."

But Ardire of Lehigh Valley says the challenge is to provide top-notch care no matter who your patients are. "We have those patients here, too," he says.

Contributing: Julie Appleby

ABC News Live

ABC News Live

24/7 coverage of breaking news and live events