Look up any hospital in the US: Hospital database
Patients line up at Maimonides Medical Center in Brooklyn, but that doesn't mean they like it. One in five complain their rooms are "sometimes" or "never" clean, that help may be late in coming and that the nightly commotion makes it hard to sleep. Fewer than half give the 700-bed Maimonides high scores for patient satisfaction. Just over half say they would recommend it to others, lower than the 67 percent average for hospitals nationwide.
Yet, by measures that arguably matter more -- what percentage of patients survive killer conditions, such as heart attacks, heart failure and pneumonia -- Maimonides ranks among the best hospitals in the USA.
See also: Mortality Rates: Where do Jacksonville's Hospitals Rank?
Medicare data released today shows that Maimonides is one of 13 of more than 4,700 hospitals nationwide with below-average death rates for all three conditions: 11.2 percent for heart attacks, compared with a national average of 15.9 percent; 7.3 percent for heart failure, compared with the national average of 11.3 percent and 6.8 percent for pneumonia, compared with 11.9 percent.
Hospital officials say they can't help but be upset by the patient ratings. "We wouldn't be human if we weren't disheartened by them," says Sheila Namm, Maimonides' senior vice president of professional affairs.
There's a flip side to the perception gap. A USA TODAY analysis of the Medicare data found that more than 120 hospitals well-liked by patients have death rates for heart attacks, heart failure or pneumonia that are significantly worse than the national average.
Twenty-two highly recommended hospitals had above-average death rates in two of these categories. In each case, the hospitals were recommended by two-thirds of their patients or received patient satisfaction scores of 9 or 10 on a 10-point scale.
Yale cardiologist Harlan Krumholz, who helped fine-tune Medicare's approach so that the hospitals are rated fairly, says personal experience can only tell you so much about the quality of your medical care.
"You can judge how it feels to be in the hospital," Krumholz says. "But you can't judge whether they're doing everything well on the medical side."
Medicare's analysis of more than 4,600 hospitals found that 323, or one of every 14, had above-average death rates for heart attack, heart failure or pneumonia. Two hospitals -- Piedmont Medical Center in Rock Hill, S.C., and Southwest Mississippi Regional Medical Center in Macomb -- had high death rates in all three categories. Thirteen hospitals had low death rates across the board.
Veterans Administration hospitals performed well, according to data released for the first time this year. Ten hospitals had lower death rates than average for heart failure; two were lower for heart attacks and five for pneumonia. All VA hospitals were as good as or better than the national rate for heart attack and heart failure. Only nine had higher death rates than average, all for pneumonia.
Over the past five years, rising health costs and concerns about patient safety have propelled the push to track and publicly report patient satisfaction and more concrete measures of patient care. Health spending in the U.S. passed $2.3 trillion in 2008. Almost a third of this money goes to hospitals, and experts say that as much as 30 percent of all medical care may be unnecessary, the Kaiser Family Foundation reports.
Experts say the goal of this approach to quality improvement is nothing less than a sweeping transformation of medicine, one in which doctors and health facilities will be judged on their performance and rewarded with when they do a good job. This effort gained new momentum with the passage in March of the Affordable Care Act, says Carolyn Clancy, director of the U.S. Agency of Healthcare Research and Quality.
Earlier this year, Medicare phased in a long planned pay-for-performance program for kidney dialysis centers. Similar programs are planned for hospitals, inpatient rehabilitation centers, hospices and cancer hospitals, says Tom Valuck, of the National Quality Forum, a consortium of medical groups, hospital organizations, health plans and others that has endorsed Medicare's approach to measuring quality.
The newspaper's analysis relies on the Medicare patient-satisfaction survey called HCAHPS, for the Hospital Consumer Assessment of Healthcare Providers, begun in 2006. A year later, Medicare launched its analysis of heart attack and heart failure death rates, which first appeared in USA TODAY and on Hospital Compare, a Medicare website.
Medicare's database, which covers patients 65 and over hospitalized between July 2007 and June 2010, has since been expanded to include pneumonia and hospital re-admissions, a measure of how many patients land back in the hospital within 30 days of their discharge. This year, for the first time, Medicare also included Veterans Administration hospitals. All death and readmission rates have been statistically adjusted so the hospitals can be compared fairly, despite differences in size, patient population and other factors.
It's tough to get it right, says John Wennberg, founding editor of the Dartmouth Atlas of Health Care, a critic who believes Medicare's method of analysis may underestimate death rates at some hospitals.
Researchers say the differences between hospitals that perform well and those that perform poorly were apparent in a study of 11 of the best and worst at caring for heart attack patients, says Krumholz, of Yale. The best hospitals, he says, have a culture devoted to quality. "You go into a good hospital and it's immediately apparent," he says.
Consumers don't get that opportunity, Wennberg says, and grapple with a different question: "How do we make health care decisions based on information we trust?"
Judith Hibbard, of the University of Oregon, and an expert in how patients perceive health care, says the answer is to factor in all kinds of information, when making important health care choices, just as a person would when buying a house or a car.
"Patients are good at judging which places are clean, whether people responded to their needs and whether they're getting adequate pain relief," Hibbard says. "But there's lots of information that's not captured in patient data that people should also pay attention to."
By Steve Sternberg and Christopher Schnaars. Contributing: Anthony Debarros and Luke Kerr-Dineen