A great medical secret — hospital death rates — will be made public in a bold federal initiative
By Steve Sternberg and Anthony DeBarros
Contributing: Marie Skelton
USA TODAY
Copyright 2007 Gannett Company, Inc.
All Rights Reserved
NEW HAVEN, Conn. — At 8:35 a.m. on Good Friday, paramedics raced Eugene Cummiskey into Yale-New Haven Hospital in a real-world test of the hospital’s response to a heart attack.
Doctors found that the 76-year-old had a blockage in an artery so vital to the heart that they call it the “widow maker.”
Within minutes, orderlies whisked him to the angioplasty suite, where doctors cleared his artery from within by inflating a tiny balloon. “It was like a crack military Special Ops unit,” Cummiskey says. “A lot of precision.”
Now recovering at home, the retired prep school administrator from Guilford, Conn., owes his life as much to geography as to his doctors.
He lives within minutes of Yale-New Haven, which has one of the nation’s lowest heart attack death rates, according to confidential data obtained by USA TODAY from a Medicare analysis of death rates at more than 4,000 hospitals.
Cummiskey couldn’t have known he’d be in such good hands. Hospital death rates are among the best-kept secrets in American medicine. The Internet may be crowded with consumer information, from school report cards to airline safety records, but death rates for most hospitals are still as closely guarded as the formulas for Kentucky Fried Chicken and Coke.
That will begin to change in June, when the Centers for Medicare and Medicaid Services (CMS) plans to post the first broad comparison of the death rates for heart attack and heart failure on its website, Hospital Compare (hospitalcompare.hhs.gov).
The federal initiative marks a bold departure for an agency that has long been the repository of private information on Medicare patients. More than a dozen top hospitals provided USA TODAY with an exclusive look at the government’s initiative by sharing their confidential Medicare death-rate report cards. The reports are drawn from death rates of heart attack and heart failure patients who died between July 2005 and June 2006, of any cause, within a month of entering the hospital.
The analysis reveals just 17 of 4,477 hospitals had heart attack death rates that were better than the national rate. Thirty-eight of 4,804 hospitals had heart failure death rates that were better than the national rate.
There was cause for alarm, too. The analysis reveals 42 hospitals where patients are more likely to die from heart attacks and heart failure than patients who go elsewhere, including at least one whose 24% heart attack death rate topped the national rate by nearly 10 percentage points.
Hospitals on notice
The agency will name the high-risk hospitals along with all the others in June, but it does not plan to take corrective action. Instead, officials say, they hope to shame them into doing better.
“If I’m running a hospital, and I see that I fall in a category that’s worse than 98% of hospitals, that’s going to grab my attention,” said Michael Rapp, CMS director of quality measurement.
While it may crack open the door to its trove of information, CMS won’t throw open the vault.
Despite multiple requests, officials declined to release the names of hospitals with high death rates before they’re posted, though doctors there continue treating fragile patients with heart disease, the nation’s leading killer. CMS officials also declined to reveal specific death rates of hospitals in the analysis.
“Detailed information won’t specifically be made available to the public,” Rapp says.
Concerned about the potential backlash from hospitals fearful that a mediocre report card will drive patients away, CMS has chosen to highlight a small percentage of hospitals with the best and worst performance compared with the national death rate.
The approach has provoked controversy among doctors and hospital administrators who fear that the analysis doesn’t give enough weight to how sick, poor, rural or urban their patients are.
“I feel strongly that the public has a right to see how hospitals and physicians perform,” says cardiologist Steven Nissen at the Cleveland Clinic. “It’s got to be done carefully. If not, it can backfire, and the whole system can fall down.”
Consumer advocates agree the move is a valuable first step, but they say people are being shortchanged by the agency’s cautious approach, which withholds specific death rates and leaves 98% of hospitals in the USA statistically indistinguishable from one another.
“As a consumer, I would want to know if my hospital has higher death rates than the hospital across town,” says Minna Jung, an expert on the quality of medical care at the Robert Wood Johnson Foundation.
Not everyone agrees medicine is best practiced by the numbers.
“A friend of mine says statistics are like clothes. What they show is much less than what they hide,” says Jerome Groopman of Brigham and Women’s Hospital and author of How Doctors Think, which explores medical decision-making. “Hospitals will market around these numbers, physicians will be paid for performance around these numbers, and fulfilling these measures will become the primary imperative for every caregiver.
“It’s almost impossible to measure the heart and soul of medicine. It’s nearly impossible to measure what people come to a doctor for. This risks taking an art and a healing relationship and turning it into a quality-control assembly line.”
Accessible information
If measurement is one pillar of the quality control movement, sharing the results with employers, insurers, hospitals, doctors and consumers is another. And by providing even a narrow glimpse of its vast trove of medical information, analysts say, CMS is helping launch a new era of relative “transparency” in which consumers can rely on information about medical quality to shop for care the way they would shop for a car.
If the movement takes hold, information on hospital and doctors’ performance will soon be a mouse-click away. Patients will be able to decide where to seek medical care based on a simple report card.
The effort also marks the beginning of a broader transformation of medicine, one in which hospitals and doctors will be routinely judged on their performance -- and rewarded with fatter paychecks when they do a good job.
That change, too, has begun. Last year, Congress authorized CMS to develop a pay-for-performance plan by 2009.
The agency launched a pilot study with the Premier Inc. network of non-profit hospitals, which involves about 260 hospitals in 37 states. The hospitals were given bonuses for taking better care of patients who were hospitalized for heart attacks, heart failure, pneumonia, bypass surgery and hip and knee joint replacement procedures.
“The surf’s up on this,” says Donald Berwick, president of the Institute for Healthcare Improvement, a group advising thousands of community hospitals on how to improve their care. “Hospitals know things are changing. They know they’re not going to be in control of transparency, and it’s not just Medicare.”
Two years ago, the agency made reporting mandatory by docking hospitals a portion of their earnings, now up to 2% a year, if they failed to report on what percentage of their patients get standard treatments, including those for heart attack and heart failure. President Bush gave his support to the movement in August with an executive order authorizing federal agencies to take steps to prompt “more transparent and high-quality care.”
What Americans don’t know can hurt them. In 1999, the Institute of Medicine reported nearly 100,000 patients a year die from medical errors. In 2003, researchers from Rand Corp. reported that only half of patients got recommended medical care.
Yet the 800,000 people who are wheeled into emergency rooms with heart attacks each year and the 400,000 Medicare patients who are hospitalized with chronic heart failure have little or no way of knowing how good or bad their care will be.
Until now, the standard yardstick for hospital performance has been whether patients get every treatment shown to be safe and effective in clinical trials. But some studies have raised questions about whether following treatment guidelines alone can improve the outcome of care. Doctors say they must make sure no other factors, such as long waiting times in the ER or lack of communication between doctors, have been overlooked.
“Any of us who’ve worked in hospitals know that if you look deeply in cases of patients with bad outcomes, there are often a lot of clues to what you could have done differently,” says Yale’s Harlan Krumholz, who led the team that developed the statistical methods used by CMS.
To help hospitals track down those clues, CMS has included in each hospital’s death-rate report coded patient information that will allow doctors to pull medical records and look for patterns that might explain why too many patients are dying. Medicare officials say the purpose is to give hospitals a new way to measure how well they’re caring for some of their sickest patients -- and to spur problem hospitals to improve.
Comparatively little information will be provided to the public. People who visit Hospital Compare will find a grid showing how each hospital’s 30-day death rate compares with the national heart attack and heart failure death rate. Using the grid, consumers will be able to scroll to a hospital and see a check mark in a box showing whether the hospital has scored higher, lower or on par with national death rates. Because the agency says it won’t reveal specifics, most of checks will land in the box marked “no different” than the national rate.
“Isn’t it amazing -- 4,477 hospitals that treat heart attacks are all the same?” asks Richard Lange, chief of clinical cardiology at Johns Hopkins University.
Rapp says Medicare officials made this choice because they want to be able to state with 95% confidence that they haven’t mislabeled any of the hospitals.
“I think this would have been more useful if they set the cutoffs so consumers could tell which hospitals were in the top 5% and which hospitals were in the bottom 5%" says Mayo Clinic’s Raymond Gibbons, president of the American Heart Association.
Wider range of death rates
Death rates vary more than the CMS presentation will reveal, USA TODAY’s review indicates. For instance, death rates for heart attack patients who died between July 2005 and June 2006 ranged from 12.3% at New York-Presbyterian to 18% at University Hospital in Cincinnati. Heart failure death rates ranged from 8.4% at New York-Presbyterian to 12.4% at AtlantiCare Medical Center in Atlantic City. Similar differences emerged in December, when CMS calculated hospitals’ 2003 death rates and supplied them to hospitals as a dry run of next month’s analysis.
Even small differences in hospitals’ death rates may signal unnecessary deaths, Krumholz says.
Just a 4% difference between two hospitals that treat 250 heart attacks a year adds up to 10 extra deaths annually.
In a report published in the journal Health Affairs in January, Krumholz and his co-workers calculated that if every hospital in the USA were able to achieve the same level of performance as the top 25%, nearly 10,000 lives could be saved each year.
He says the study adds to evidence that “where you’re taken can have as powerful an impact on whether you survive as whether you get important medications.”