Hospital rankings: More than meets the eye

Medicare's pay-for-performance program ranks and rewards hospitals according to how well they meet certain guidelines for clinical care. But researchers at Duke Clinical Research Institute say the program penalizes hospitals that care for the greatest numbers of the poor and needy by not taking into account their greater clinical burden.

Studies show that age, race and severity of disease can influence which patients get certain treatments and how they fare, and these factors vary significantly from hospital to hospital. "That means that hospitals serving large groups of the elderly, women, poor, uninsured or African American patients might have problems competing with institutions whose patients are younger, wealthy, insured, and white," says Dr. Eric Peterson, a cardiologist at Duke and the senior author of the study. "Hospitals are simply not starting out on the same playing field."

Under the current pay-for-performance system, hospitals in the top 20 percent of the rankings receive financial reward; those in the middle 60 percent receive nothing, and those at the bottom stand to loose Medicare reimbursement money. Currently, Medicare does not consider demographic variables and patients' existing health problems in figuring hospital rankings.

Peterson and colleagues in the American Heart Association's Get With the Guidelines program wanted to find out if those rankings would change if patient mix was included in the calculations. They examined Medicare beneficiary records of 148,472 heart attack patients in 449 hospitals across the country from 2000 to 2006. They analyzed the hospitals' process performance on eight measures of clinical care taken from guidelines established by the Centers for Medicare & Medicaid Services. The measures included appropriate use of certain drugs, like aspirin, beta-blockers, ACE inhibitors and anti-clotting medications; and procedures, like angioplasty and counseling to support smoking cessation.

Investigators ranked the hospitals according to crude composite process performance scores and then grouped them according to Medicare's current system. Next, they ranked the same hospitals again, but this time taking into account the patients' demographic variables, their clinical characteristics and eligibility for certain procedures.

They found that the hospitals with the lowest crude composite scores tended to be smaller, non-academic institutions that treated a higher percentage of older, sicker and minority patients than those in the top group.

While there was general agreement on performance between the two ranking systems, researchers found that when taking into account patient characteristics and treatment opportunity, 16.5 percent, or 74 of the hospitals would fall into a different financial status category.

So why doesn't Medicare consider patient mix in tallying rankings? "On the surface, it may well seem to be the right thing to do, but some feel such a move would 'legitimize' less-than-optimal care," says Peterson. "At the same time, not taking these factors into consideration is like comparing apples to oranges."

Peterson says one solution might be to reward hospitals for improvement in adherence to
evidence-based treatment, rather than rewarding a single score or ranking. Another option might involve separately reporting adherence data for older patients, women, or minorities. "That would surely draw more attention to any gaps in care, and might prompt better compliance."

Source: Duke University Medical Center