(This was originally posted on the Health Affairs Blog on 6-26-12.)
At its best, the free market rewards companies that provide a superior service at the best price. A few years back, Medicare tried to bring similar incentives to its market-oriented Medicare Advantage program, which was created to give seniors more control over their own health dollars. Retirees can choose among competing, privately administered health plans that must provide the same benefits as Medicare's traditional, fee-for-service program. MA has increased in popularity over time; today, one-quarter of all Medicare beneficiaries are enrolled in the program.
In 2008, in order to help seniors choose among the various MA plans, the Centers for Medicare and Medicaid Services (CMS) instituted the star-rating system, in which CMS would assign one to five stars to plans based on their performance on features such as managing chronic conditions, routine care, and customer service.
President Obama's Affordable Care Act built upon the star-rating system by assigning bonus payments to plans that performed well, and charging penalties or rebates to plans that performed poorly. However, a recent report from Douglas Holtz-Eakin, Robert Book, and Michael Ramlet at the American Action Forum shows that the ACA bonus system doesn't live up to its billing. Without significant reforms, it promises to prevent, rather than encourage, better plan performance.
Insurers Can't Travel Back In Time
The first problem with the ACA's pay-for-performance system, the report's authors note, is that CMS's performance criteria are applied retroactively. "For example," they write, "CMS published criteria in October 2011 to be applied to plan performance between January 2010 and June 2011." Intuitively, CMS needs to set out its performance guidelines beforehand, so that plans can adjust accordingly for some future period. CMS did not.
In addition, the authors point out that not all of the performance criteria are measured over the same time frame: "For example, the October 2011 criteria included cancer and cholesterol screenings for calendar year 2010, but flu vaccination for February through June 2011." Indeed, the criteria measured 36 different performance endpoints over six distinct time frames, some of which did not overlap.
Even worse, MA plans couldn't make adjustments for next year based on the October 2011 guidelines, because they were required to submit their 2012 plan designs in June 2011.
Retroactive Criteria Incentivizes Favoritism
The retroactive nature of the new pay-for-performance system makes CMS vulnerable to charges of favoritism. Because CMS designs the criteria to reward past performance, CMS could end up picking winners and losers, either intentionally or unintentionally, on a non-meritocratic basis.
In addition, say the report authors, "the star system could be used as a back-door budgeting tool" by setting goals that plans are realistically unable to reach. They cite the example of all-cause readmission within 30 days. MA plans have a readmission rate of 15 percent, compared to 20 percent for traditional fee-for-service patients. However, the October 2011 guidelines required a rate of 5 percent to gain a 5-star rating, which is virtually impossible.
The Lake Wobegon Of MA Plans
We also know from a recent report from the Government Accountability Office that the Department of Health and Human Services has made some curious, and apparently unscientific, alterations to the pay-for-performance system, so as to avoid reducing MA subsidies prior to 2013.
Under the department's "Medicare Advantage Quality Bonus Payment Demonstration," according to the GAO, HHS modified the pay-for-performance system, such that nearly every plan--covering 90 percent of MA enrollees--will receive a bonus in 2012 and 2013. Medicare Advantage has seemingly moved to Lake Wobegon, where nearly every plan is above average.
In reality, however, not all localities are like Lake Wobegon. Some regions have higher poverty rates than others, with correspondingly poorer health outcomes. It would make sense for CMS to adjust its ratings scale based on these socioeconomic factors. So far, it has not done so.
What Would A Real Pay-For-Performance System Look Like?
A proper pay-for-performance system would look very different than the one that has been applied to Medicare Advantage.
First off, as the GAO recommends, the Quality Bonus Payment Demonstration program should be scrapped. That program undermines the idea behind pay-for-performance by rewarding nearly all plans in similar ways, regardless of performance.
Second, CMS should prepare its guidelines prospectively, such that they apply to future plan performance and are issued at a time of year when plans can incorporate those guidelines into their future designs.
Third, CMS should adjust its star ratings to take into account regional differences in the quality of care.
Paying for performance is a good idea. Those who favor a strong role for public sponsorship of health insurance have a special incentive--and a special responsibility--to get it right.