The government has identified hundreds of hospitals whose Medicare patients are incurring especially high bills, a first step toward using bonuses and penalties to encourage more efficient care.
The new Medicare figures show wide variance among hospitals around the country, even ones just a few miles apart. In Los Angeles, for example, the average patient admitted to Los Angeles Community Hospital cost Medicare nearly $24,644 during the stay and in the month afterward, 37 percent above the national median. Across town, according to the data, an essentially similar patient admitted to Ronald Reagan UCLA Medical Center cost Medicare $17,628, or 2 percent below the median.
Medicare believes that identifying these discrepancies will help clamp down on excess medical care, which some health care researchers believe could comprise as much as a third of the nation's $2.6 trillion health care spending and be responsible for much of Medicare's fiscal woes. As the nation's biggest insurer, Medicare's proposal to link reimbursement to efficient care is expected to reverberate throughout the health care system when it starts in October 2014. The plan has alarmed hospitals, particularly since Medicare intends to hold hospitals accountable not just for the cost of services provided while at the hospital and in follow-up care, but for every test, treatment or stay that occurs 30 days after discharge, no matter the reason.
"I wouldn't take any of these cost measures to the bank," said Dr. Tom Rosenthal, the chief medical officer of the UCLA Health System. He credited Medicare for making a good first, although imperfect, stab at measuring efficiency, but said, "I think we need to know a lot more before we start publishing lists of hospitals and doctors and declaring who is most efficient."
Kaiser Health News analyzed the new Medicare data and found that even after the severity of illness and overall health condition of patients were taken into account, Medicare paid 5 percent or more above the national median to care for patients from 657 hospitals, about one in every five. The cost for caring for Medicare patients at 1,150 other hospitals — or one out of every three — was at least 5 percent below the national median, according to the KHN analysis.
Medicare plans to judge hospitals in two ways, with hospitals getting more or less money based on how their patients' spending levels compare with other hospitals and how the hospitals' spending compares with previous years. So even hospitals whose patients are expensive to Medicare may not lose money if they are reducing their patients’ spending rates by a fast clip.
The new data appear to reaffirm decades of studies showing that patients in some regions of the country are bigger users of health care services than other places. Patients treated at most or all hospitals in Las Vegas, Fort Lauderdale, Newark, Miami, Los Angeles and Orange County, Calif., tended to cost more than the national median, which is $17,988. Patients treated at most or all hospitals in Anchorage, Des Moines, Honolulu, Minneapolis and Portland, Ore., tended to cost Medicare less.
The data provide new insights into the differences in costs to Medicare of patients at specific hospitals, allowing for comparisons between hospitals from different parts of the country. Medicare said it removed some factors that had confounded previous analyses, such as the fact that hospitals in some places get more Medicare reimbursements to offset higher wages, the large numbers of low-income patients they treat and the number of medical resident they train.
Medicare also said it took into account the differing mixes of patients hospitals treat, allowing for academic medical centers that handle lots of complex cases to be compared to small rural community hospitals.
The data, known as the "Medicare Spending per Beneficiary measure" and published on Medicare's Hospital Compare website, also provide new avenues of inquiry into one of the most perplexing questions about America's high health care costs: why do similar patients end up using more — or more expensive — health services than others?
In some regions, such as Los Angeles, the most expensive patients were ones who had been admitted to community hospitals. (Officials at Los Angeles Community Hospital’s corporate headquarters did not respond to requests for comment.)
In other places, such as Washington, D.C., the most expensive patients came from a teaching hospital: The average patient of MedStar Georgetown University Hospital cost Medicare $18,708, or 4 percent above the national median. Yet the hospital with the least expensive patients was also a teaching hospital, albeit one that generally serves lower income residents. Patients at Howard University Hospital cost Medicare $16,369, or 9 percent below the median.
Georgetown officials didn't respond to requests for comment, while Howard officials said they were interested in why the hospital's patients cost Medicare less. "We'd have to dig into that calculation," said Aldwin Lindsay, Howard's chief financial officer.
Trying To Explain The Gaps
Experts offered various possible explanations for the gaps. Some hospitals have more specialists working on a case, which drives up costs. If a hospital did a subpar job, a patient could end up getting more tests and procedures afterward or be readmitted, also running up Medicare’s costs. Others said some hospitals might be more likely to discharge patients to expensive places to recuperate, such as long-term care hospitals.
"I do think post-acute care is a major driver" of the cost differences, said Jennifer Faerberg, director of health care affairs at the Association of American Medical Colleges.
Nancy Foster, a vice president at the American Hospital Association, said the data do not answer key questions: Did the patients that got more services fare better than others? Could the patients that cost Medicare less actually have benefitted from more care? "What we don't know is if those additional investments yield differences in outcomes," Foster said.
A number of experts expressed caution and suggested it is very difficult to accurately adjust for the fact that patients at some hospitals are sicker than those elsewhere.
"The efficiency measures are complex and contentious, and I do think they will be debated," said Dan Mendelson, CEO of Avalere Health, a consulting company based in Washington. "While hospitals understand that the system is moving there, they have significant questions, particularly about risk adjustment."
The Centers for Medicare & Medicaid Services' analysis offered a very different assessment than did the Dartmouth Atlas, which pioneered the study of differences in the amount of medical services patients received depending on where they live and which hospitals treated them. Dartmouth has identified New York City hospitals as places that are expensive to Medicare, but CMS found that patients from the city's hospitals were in line with their peers nationally.
The differences between the two analyses may have to do with the fact that CMS discounted the extra government money that went to New York because it has so many teaching hospitals. Hospitals in McAllen, Texas, another region of the country that Dartmouth judged as so extremely expensive that its name became nearly synonymous with excess health care use, hewed close to the national median in the Medicare analysis.
Elliott Fisher, one of the main researchers from the Dartmouth Atlas, questioned the practical usefulness of the new information. "As a hospital administrator I would go, how does this help me?" he said. "We just don’t know whether a lot of specialists are running through the hospital doing everything they can to every patient who is horizontal, or whether they're discharging every patient to a rehab facility. Those are two very different causes of high costs."
In an ideal world, patients with the same health and diagnosis should cost the same to treat, since Medicare pays set fees and doesn't haggle with providers like private insurers have to. But the Medicare data, while ranking hospitals and areas differently than did Dartmouth, appeared to back the fundamental point of the Dartmouth research: that some places are delivering health care more cost effectively than others. In the Nevada and Arizona areas surrounding Las Vegas, patients from 14 hospitals -- all but two in the region -- cost Medicare 9 percent or more above the national median.
The Medicare data indicated big spending differences in areas of the country that have not generally been thought of as high users of Medicare services. In Kansas City, Mo., the average patient admitted to St. Joseph Medical Center cost Medicare $19,247 during a stay and in the month afterward, 7 percent above the national median. Fifteen miles away, according to the data, an essentially similar patient admitted to Truman Medical Center-Lakewood cost Medicare $15,290, or 15 percent below the national median.
The owner of St. Joseph, Corondelet Health, which is part of the nation's largest Catholic nonprofit system, Ascension Health, declined to comment. Truman said in a written statement: "It is important that there be a comprehensive analysis of this data and its variables before final reimbursement conclusions are reached."
Paul Matsui, executive director at the Advisory Board Co., another Washington consulting firm, said his firm has already started analyzing the data to see what factors were linked to more efficient hospitals. He said whether a hospital tended to treat wealthy or impoverished patients did not seem to make a difference in Medicare’s ratings. His initial analysis indicated that rural hospitals and smaller hospitals all tended to have lower cost patients than urban hospitals and large hospitals, he said.
"It's all the factors you can’t do anything about," Matsui said.