Launched with a campaign that sought to set the world on fire, the initial pilots using the Prometheus bundled-payment model found themselves struggling to get started as implementation proved to be as torturous as the project's acronym.

With a $6 million Robert Wood Johnson Foundation grant it received in 2007, the Health Care Incentives Improvement Institute developed Prometheus, which stands for “Provider payment Reform for Outcomes, Margins, Evidence, Transparency, Hassle-reduction, Excellence, Understandability and Sustainability.” The concept involves creating evidence-informed case rates, or ECRs, to calculate the cost of care episodes with probability risks taken on by payers and technical risks—those determined to be within a provider's control—being shouldered by providers.

Last November, health policy journal Health Affairs published a RAND Corp. report noting that, as of May 2011, no contracts had been signed, no bundled payments had been made, and implementation was full of unforeseen challenges—mostly because of the complexity of the payment methodology. Yet despite this, participants believed the program would eventually result in reduced costs.

And, while lessons have indeed been learned and obstacles to implementation removed, the report's authors concluded that delays associated with the use of claims data and the time needed to learn and use the complex MedAssets' Episode of Care Engine computer system “are likely to be experienced” by new Prometheus participants.

Francois de Brantes, the institute's executive director who has supervised the Prometheus pilot implementations, vehemently disagrees.

“It's absolutely untrue,” de Brantes says, explaining how the quick implementations with the second-round pilots in North Carolina and Wisconsin “completely contradict” the Health Affairs statement that even those who learn from history are still doomed to repeat it.

According to de Brantes, the CEOs at Blue Cross and Blue Shield of North Carolina and CaroMont Health implemented a knee-replacement program, worked out contracts and set bundled prices and risk-sharing agreements within four months. In Wisconsin, Anthem Blue Cross and Blue Shield and the Robert Wood Johnson Foundation's Aligning Forces for Quality effort worked out their own knee-replacement program with interested providers in nine months. (A diabetes program is still in the works.)

North Carolina Blues spokesman Lew Borman says some success is being seen in the early stages of the pilot in reducing avoidable complications and increasing collaboration among providers. "All indications are positive," Borman says.

De Brantes agrees. “Where there's CEO engagement, things are taking off,” he says, adding that without that engagement, “it's a heavier lift.”

The main author of the report, RAND policy researcher Peter Hussey, however, stands by the paper's findings and says the three initial pilots—in Illinois, Michigan and Pennsylvania—all had CEO engagement and characterizes them as “more prepared than most” healthcare organizations to take on a complex bundled-payment methodology.

“Even with CEO engagement, you need engagement of frontline staff to get this model to proceed,” he says. “That will take some time.”

Hussey adds what was particularly “eye opening” was the amount of information needed to track and evaluate the cost of an episode of care and the deficiencies in the claims data being used to do that.

The pilot in Rockford, Ill., involves the Employers' Coalition on Health, a healthcare-purchasing coalition of some 160 employers, the SwedishAmerican Health System and OSF St. Anthony Medical Center. A kickoff meeting was held in September 2009, but no bundled payments have been made yet regarding programs targeting coronary artery disease, diabetes and hypertension.

“We were expecting to move through the implementation much faster,” says Albert Ferrabone, management services organization quality manager at SwedishAmerican. “We came across the proverbial ‘devil in the details' from the very first starting point.”

Among the snags that were hit right out of the gate were problems with the Prometheus data-use confidentiality and business associate model agreements. “Our legal counsel felt they were not sufficient,” Ferrabone says.

Other problems included third-party payers not having unique member identifiers. Instead, numbers were assigned at the family level, and this led to problems calculating who was involved in the episode of care as it could not be determined whether it was individuals insured through their employer, their spouses or their dependent.

There also were problems with prescription data and slow feedback on quality. And “despite the best efforts” on the part of ECOH, Ferrabone says the employer coalition was “unable to secure commitment from employers on shared savings” payments.

All that said, however, Ferrabone credits the ECOH for having foresight for bringing Prometheus to the Rockford market.

“They certainly put their heart and soul into this,” he says.

Ferrabone recommends that any organization looking to adopt the Prometheus methodology partake in a self-assessment first to make sure everyone is on the same page and the information technology infrastructure can deliver what is required.

“One of the things Prometheus needs is quality data,” Ferrabone says. “Do you have an electronic medical record to retrieve that data? Does your IT system provide the granular level to deliver on that? … You don't want get into a project like this and, halfway through, find out your system doesn't provide the level of detail necessary.”

Ferrabone added that quality data was available from his organization's system, but utilization data from employer-supplied claims data was often insufficient to calculate the Prometheus evidence-informed case rate.

He says the organizations involved are still extracting quality data using Prometheus methodology and still pursuing the goal of bundled payments, but whether the Prometheus Rockford team stays together remains to be seen. “At this point, we're pausing and thinking about next steps and where do we go from here,” he says. “There's a lot of value, so my choice is that we move forward on this.”

The experience has been helpful as a bundled-payment laboratory, Ferrabone says, adding that improvement has been shown in lowering the cost of care for coronary artery disease. A Prometheus case study of the Rockford pilot also noted a key area of improvement: communicating with patients the importance of not eating before a visit in preparation for a cholesterol test.

Although it's unknown why costs decreased, Ferrebone says that, in an analysis of the same coronary artery disease patients, the average cost of an episode of care was $5,663 in 2009 with $1,000 identified as "potentially avoidable costs" (the costs providers are on the hook for). In 2010, the average cost of an episode of care for the same patients was $4,847, with $605 identified as potentially avoidable. Ferrabone says no specific intervention can be linked to the decrease, and he acknowledges that these were "only two data points in time," but adds that it shows the importance of keeping the project going.

Out West, the Colorado Health Foundation helped launch more pilots: First, with a $250,000 planning grant in 2010 and then a $3.2 million implementation grant to be paid over a three-year period that started in 2011. The grant will pay for three pilots in the Boulder/North Denver region, Colorado Springs and the San Luis Valley area.

“We went in with our eyes wide open,” says Kelly Dunkin, CHF vice president of philanthropy, saying the grants were a “calculated investment” that would advance the foundation goals of healthy living, increased health coverage and access to high-quality, coordinated care.

“We want to support innovation, and this is innovating how we pay for care,” Dunkin says, adding that a focus of the pilots is on care for chronic conditions.

Hussey notes that the Prometheus methodology is “cleaner” in covering episodes of care for orthopedic procedures than for chronic care as people being treated for diabetes often have other conditions that must be treated simultaneously as well. Also, with chronic illness, patients are more likely to also see other providers outside the network of Prometheus participants.

The point is also recognized by de Brantes, but he notes that while some savings can be achieved by reducing over prescribing of antibiotics or over-ordering of diagnostic tests, in the big picture, those are only pennies saved.

“The dollars are in patients with chronic illness,” he says.

Hussey notes that RAND is “still engaged” with the Prometheus model, but says it's also looking into how it fits into the landscape of accountable care organizations and patient-centered medical homes. “People think they are complementary, but how they fit together has not been analyzed yet,” he says, adding that he and his colleagues have given presentations about their Prometheus research to the CMS and commercial payers.

The CMS' Center for Medicare & Medicaid Innovation is launching its own “Bundled Payments for Care Improvement” pilot involving four models. It recently pushed back its application deadline for three of the models to April 30 from March 15 “because of the unprecedented public response to this initiative” and, according to its website, it needs additional time to “accommodate the extraordinary degree of provider interest.”

De Brantes says the CMMI Model 2 is identical to the Prometheus knee-replacement pilot project in North Carolina.

“The two parties have negotiated a bundled price for a knee replacement, (the Blues) continues to pay underlying (fee-for-service) claims submitted by the providers, and then does a back-end reconciliation,” he says. “If the providers come under budget, (the Blues) pays them the difference, and vice-versa.”

One reason why bundled-payment programs like Prometheus hold such promise is because the arrangement “allows physicians to be masters of their own destiny,” and to be rewarded for controlling what is possible for them to control.

“People have to have a clear line of sight on their results and why they matter,” he says. “You simply have an agreement that says, ‘If you meet this objective, this will be the result.' ”