Many in healthcare, including leaders at the CMS Innovation Center, are betting on experiments with more coordinated services and value-based payments to show it's possible to spend less for better outcomes. For two decades, the Medicare program has tried some of these ideas—and now a new federal report has some bad news about the results.

Last week, the nonpartisan Congressional Budget Office reported that a study of 10 major demonstration projects in care coordination and value-based payment showed that most programs did not reduce Medicare spending. Despite the discouraging conclusion, providers and researchers alike say there's too much promise to give up on the demonstration concept just yet.

The CBO's findings examined the outcomes of six disease-management and care-coordination demonstrations and four value-based payment projects. For the former, the programs aimed to improve the quality of care for beneficiaries with chronic illnesses whose healthcare is particularly expensive. In the latter category, providers were given financial incentives to improve the quality and efficiency of care—as opposed to payments based on volume and intensity of services delivered.

According to the results, in nearly every program that involved disease management and care coordination, spending was either unchanged or increased relative to the spending that would have taken place if the program didn't exist—when the fees paid to the program were taken into account. Similarly, the results for the value-based payment projects produced little or no savings for the Medicare program, the study showed.

Randy Brown, vice president and director of health research at Mathematica Policy Research, said the study's conclusion that the disease-management and care-coordination projects didn't lower Medicare spending is not surprising. First, it's hard to save money on a very sick patient population with chronic problems such as heart disease or diabetes. For example, Brown said a patient in this group could have both arthritis and heart disease—and could be told by an arthritis specialist to stay off his feet, while his cardiologist instructs him to get on a treadmill.

“Couple that with changing physician behavior,” said Brown, who provided comments to the CBO on an early draft of the report. “Physicians don't get paid under Medicare fee-for-service for talking to each other—and they don't,” he said. “Either the physician refers them (the patients), or they go on to their own specialist, and information doesn't get back to the primary-care doctor,” he continued. “It's really complicated to do, and there's not any structure to do it or any incentive—so no wonder it doesn't happen.” That's why, Brown said, there needs to be the right “carrot and sticks” to get significant results.

Brown's work with other researchers—which was published in the Journal of the American Medical Association in 2009—analyzed the Medicare coordinated-care demonstration that ran from 2002 to 2006 and was extended until 2008 for 11 of the 15 programs. The programs targeted mostly patients with congestive heart failure, coronary artery disease, chronic lung disease or diabetes. The Medicare coordinated-care demonstration is one of the six care-coordination demonstrations evaluated in the CBO report. Among the high-risk populations in those 11 programs, only four were successful, and they each reduced hospitalizations by 11%. But the results of the successful programs—and the methods used to achieve success—indicate that demonstration projects have value.

“We’re in the beginning phase of understanding.” —Dr. Ken Coburn, Health Quality Partners

According to Brown, researchers found certain program features associated with three or four of the successful programs—but in few or none of the unsuccessful programs. These included frequent face-to-face contact with patients (averaging about once a month for the program's first year); opportunities for ongoing face-to-face contact with patients' physicians to improve communication; care coordinators serving as a “communications hub” in which they made sure all providers have essential information about patients from other providers; a strong, evidence-based patient education intervention, including how to adhere to their medications; and a timely, comprehensive response to care-setting transitions.

Only one of the programs in the Medicare coordinated-care demonstration—Health Quality Partners in Doylestown, Pa.—has continued. Dr. Ken Coburn, CEO and medical director there, was one of the founders of HQP and serves as the not-for-profit organization's only physician. The organization's program was deemed to be budget-neutral because it saved as much as the program costs and produced benefits to patients, he said. Because of this, HQP asked to extend the demonstration program for the group of patients who have at least one chronic disease (heart failure, coronary heart failure, diabetes or chronic obstructive pulmonary disease) and who have had at least one hospitalization in the past year.

“That is the group we've been extended to serve because our model had a big impact on reducing hospital admissions and saving CMS money,” said Coburn, who added that the program has been extended until June 2013.

Coburn is scheduled to give a presentation this week at the CMS Innovation Center's Care Innovation Summit in which he will show that HQP's program results achieved CMS' “triple aim” of better healthcare, improving health and lowering costs. The program was a randomized, controlled trial that was implemented in more than 90 practices in a four-county area. From April 2002 through today, the program has involved more than 2,600 Medicare beneficiaries—and served higher-risk patients since late 2010.

Coburn acknowledged that there needs to be more research to show for certain why the program was successful, but he offered some preliminary reasons why. To begin with, the program provided a wide range of preventive interventions that nurses used with patients.

“It's a broad portfolio within the scope of nursing and they do it with a long-term, continuous relationship with frequent contact: half or more are in person, or one-on-one in the home, or a follow-up by phone,” he said. “It's a very contact-intensive work.”

In addition to a broad range of preventive services, the program cast a wide net for the patient population it served to include low-risk, medium-risk and high-risk patients (which were determined through a health-risk assessment tool). According to statistics about the program, net costs increased by about 9% for the “all-risk” group, but reduced significantly—about 28%—for the higher-risk group.

“Basically, the lower-risk groups were not consuming much healthcare to begin with—i.e., they were already at lower cost by dint of being healthier,” Coburn said later in an

e-mail. “So it was harder to save enough money in those groups to offset the cost of the program. That said, there is good evidence that the program still improved the health outcomes of some of these lower-cost folks. For example, participants in our program with coronary artery disease who had not been using much healthcare (i.e., were low cost) still had lower death rates than their control counterparts, but didn't show much net savings.”

Coburn said he would recommend a “robust design” for other groups involved in demonstration projects, meaning they should assemble as many evidence-based interventions as possible that have been proven to improve health outcomes. He also cautioned against giving up on a demonstration before realizing what works. “It takes focus and understanding,” he said. “We're in the beginning phase of understanding.”

The same need for understanding and evaluation applies in the value-based payment demonstrations, said Stuart Guterman, vice president of payment and system reform at the Commonwealth Fund. Guterman served as the director of the CMS' Office of Research, Development and Information from 2002 through 2005, and worked with the Physician Group Practice demonstration and the Premier Hospital Quality Incentive, two of the four value-based payment programs the CBO evaluated.

Results varied for the four value-based payment demonstrations, according to the CBO report. In one of the programs, Medicare made bundled payments that covered all hospital and physician services for heart bypass surgeries and spending for those services was reduced by about 10%. The other demonstrations for value-based payment produced little or no savings for Medicare, the study found.

Guterman said it is important to monitor projects along the way, rather than set a clock and determine the program's worth after the buzzer goes off. “You set something in motion; you partner with the folks doing it to make it work; you objectively evaluate how it's going on; and you stop it if it's obvious it won't work,” he said. “But you try to make it work.”