Researchers Find Lessons in Failed Bundled Payments Pilot

A major demonstration project designed to gauge the effectiveness of bundled payments exposed the complications of implementing such a system. Officials say the 3-year study fizzled after participation waned and the number of applicable cases proved too few to be statistically relevant.

A major demonstration project designed to gauge the effectiveness of bundled payments exposed the complications of implementing such a system, according to an article in the August edition of Health Affairs.

The California-based Integrated Healthcare Association (IHA) was awarded $2.9 million in 2010 to study the feasibility of bundled payments in that state. However, despite enthusiasm from providers and health plans, officials say the 3-year study fizzled after participation waned and the number of applicable cases proved too few to be statistically relevant.

“None of the barriers encountered in IHA’s demonstration signal a ‘death sentence’ for bundled payment,” M. Susan Ridgely and Peter S. Hussey, 2 of the paper’s authors, wrote in a post on the Rand Corp.’s blog. “However, the demonstration clearly shows that bundled payment is difficult to implement.”

The project was designed to institute bundled payments for orthopedic procedures performed on patients under the age of 65. Six of the state’s largest health plans, 8 hospitals, and one independent practice association initially agreed to participate, with a goal of completing 500 cases in the first 2 years of the study.

Instead, researchers ran into a series of road blocks. Participants had trouble defining the services included in bundles, a process hampered by competing interests and distrust among the participants. There was also regulatory uncertainty, leading to months-long discussions over contract language and other concerns. Participants found their software wasn’t equipped for bundled payments. When such software was developed, providers said it wasn’t cost-effective to purchase the software given the low number of patients for whom bundled payments were used.

In the end, only 3 health plans and 2 hospitals ended up joining the demonstration project, generating only 35 cases.

“It did not fail because its participants lacked enthusiasm or did not work hard, but because design decisions made early on in the process set a course from which the pilot could not recover,” the authors wrote.

Still, the study’s organizers see value in the experience, even though it failed to meet its objectives.

Writing in the Health Affairs blog, IHA’s Tom Williams and Jill Yegian said their study should be interpreted “as an early, hard-fought battle offering important lessons.”

Among those lessons, Williams and Yegian suggest the next demonstration project begin with a “retrospective” payment model, where fee-for-service claims continue to be collected in an effort to determine savings created by using bundled payments. They said such accounting would also help smooth out the regulatory confusion surrounding bundled payments.

The article also urges future researchers to iron out technical issues and simplify definitions. They say the designers of future studies need to find ways to ensure a sufficient volume of cases and work with health plans to design benefits so as to steer patients toward bundled payment services.

Finally, the authors recommend that language in bundled payment plans include criteria to lay out which procedures are appropriate and which are not, and also find better ways to manage risk.