An upcoming Medicare scheme that rewards hospitals that hit performance benchmarks might turn the government into a "reverse Robin Hood," according to a new study, as hospitals in poor regions with less educated inhabitants might lose out on funds they desperately need to their more advantageously situated counterparts.

The Centers for Medicare and Medicaid Services looks to roll out a plan in the next two to three years that rewards hospitals by how well they meet certain performance measures, in an effort to encourage better health care. But some investigators think the program falsely assumes all hospitals have the same financial and personnel resources to rapidly meet performance benchmarks. And the program could bring with it an unintended consequence: hurting hospitals in low-income areas already struggling to operate.

The idea for the research occurred to Dr. Jan Blustein, a professor of health policy and medicine at New York University's Wagner school of public service and lead author of the study, when she was consulting with hospitals in low-income, predominantly minority areas.

"It struck me that they didn't have what they needed in a lot of ways," Blustein told DOTmed News. "Asking them to perform at the level of a hospital I would be admitted to if I were sick is a bit silly."

In the study, published online Tuesday in open-access journal PLoS Medicine, Blustein and her colleagues analyzed 2,705 hospitals from 2004 to 2007 on Hospital Quality Alliances measures voluntarily reported to Medicare.

For the study, the researchers looked at HQA metrics for heart attacks and heart failure, such as whether patients having a heart attack got aspirin in the emergency room, or whether those with heart failure got the proper assessment of heart function. The researchers chose these measures because they depended on the hospital staff, and not on patient demographic factors, Blustein said.

The researchers then gave hospitals composite mean scores, out of 100, for each condition, to assess how well they did.

The researchers also modeled how pay-for-performance schemes would work in practice by assigning another set of scores based on a proposal described by CMS in a 2007 report to the U.S. Congress. In this, they graded for their absolute performance, or attainment, and also for how much they improved over the previous reporting year. Both scores were out of 10, with the hospital getting whichever of the two numbers was greater.

The researchers found hospitals in counties with longstanding poverty, the fewest college and high school graduates, and the most unemployment tended to fare the worst on these measures. For instance, hospitals located in areas blighted by decades of serious poverty -- with over a fifth of the population considered poor -- earned on average nearly three points less for mean composite heart attack performance measures in 2007, according to the study.

And for heart failure, hospitals in areas suffering from longstanding poverty had a mean HQA composite score of 73, compared with the hospitals not in poor regions, that got an 84, according to the paper.

Although low-performing hospitals, disproportionately from poor areas, improved the most in absolute terms over the four years covered by the study, they "still lagged behind their locationally advantaged counterparts," the researchers wrote.

Plus, the way scores are assigned in the Medicare model used by the study favored higher-performing hospitals, Blustein said.

"If you start low, it's like a ruler that's stretched out," she said, meaning the low-performing hospitals had to do more to gain points. "The attainment scores for better-off hospitals swamped improvement scores, so they ended up doing better."

REASON FOR PERFORMANCE SPLIT

It's up in the air what's causing the difference in performance outcomes, according to the researchers. Hospitals in richer, better educated areas not only reported higher scores, but were more likely to report, suggesting their ability to perform could have to do with being more wired with high-tech IT equipment, and having better trained personnel able to use it.

"To some extent that has to do with having good IT and good coders, and it may be entirely due to that," Blustein said. "We don't know."

More worryingly, the problems could be structural. In the paper, Blustein mentions that hospitals in poor areas can have a hard time attracting top talent, because of lack of good nearby schools or limited opportunities for the staff's spouses to work. Also, they could simply be poorer.

"They might not have the money to hire consultants to help them meet these requirements," Blustein said. "[Richer] hospitals can just buy [staff] essentially to helicopter in to show them how to do what they need to do."

Medicare will likely start rolling out the program in 2012. Similar plans have already been implemented in the UK, Australia and Taiwan, the researchers said.

Read the study: http://www.plos.org/press/plme-07-06-blustein.pdf

Check out this earlier DOTmed News story on pay-for-performance schemes hurting doctors who serve the poor and ethnic minorities.