Symbol of Health System Transformation? Assessing the CMS Innovation Center

The Center for Medicare and Medicaid Innovation (CMMI), created by the Affordable Care Act (ACA), is catalyzing profound changes in U.S. health care. Congress gave CMMI $10 billion for 2011 through 2019 to test “innovative payment and service delivery models to reduce program expenditures . . . while preserving or enhancing the quality of care.”¹ It gave the Center power to expand any model without requiring new legislation if the model reduces costs without harming quality or increases quality without increasing costs. In just over four years since CMMI began operations, it has created a large and diverse portfolio of programs (see accompanying article). But the Center faces four key challenges.

First, has CMMI split its attention among too many programs? Nearly half the programs managed by CMMI were mandated by Congress as part of the ACA or previous legislation, so the number of programs is not entirely at CMMI’s discretion.² Nevertheless, lack of focus could result in poor program design, inadequate program management, and unconvincing program evaluations. On the other hand, important innovations are more likely to be discovered by exploring many different approaches, learning how they work in practice, and learning the extent to which health care organizations are willing to participate. Furthermore, CMMI has spent or obligated only about one-third of its $10 billion; since Republican opposition to the ACA makes its future uncertain, it may be better to spend the funds sooner rather than later.

CMMI’s second challenge is whether, as a government bureaucracy that must consider the views of Congress, the White House, and other CMS centers, it can make its programs responsive to the day-to-day responsibilities of health care providers? Operational details are critical to provider organizations. For example, how will CMS decide which beneficiaries are “attributed” to each organization? What cost and quality benchmarks will be used to measure performance? Will CMMI provide critical data to organizations in a timely manner? If CMMI cannot design programs in ways that make sense to providers, and quickly revise them when problems are discovered, organizations will leave, and the programs will fail.

CMMI uses multiple formal and informal mechanisms to gather and respond to feedback. CMMI’s recently announced “Next Generation ACO” program, ³ for example, includes major innovations in response to criticisms from participants in its flagship Pioneer Accountable Care Organization (ACO) program. CMMI continues to be slow, however, in transmitting relevant data to provider organizations in its programs. Such information sharing is a key function that some private programs — such as the ACO-like Alternative Quality Contract program of Blue Cross Blue Shield of Massachusetts — perform better.

The engagement of private payers may be essential to addressing a third key challenge — can incentives in CMMI’s programs be made strong enough to attract provider organizations to participate? It is difficult and expensive for hospitals and medical groups to create the infrastructure necessary to improve quality and reduce the cost of care. To succeed, they must hire nurse care managers, develop and deploy expertise in quality improvement, analyze relevant data (for example, to identify the patients who most need assistance), and report performance data to CMMI. Organizations may not choose to make large investments for a 3-year pilot program that covers only their Medicare or Medicaid patients.

CMMI has attempted to address these problems by recruiting multiple payers to participate in some initiatives (e.g., the Comprehensive Primary Care [CPC] initiative);⁴ providing up-front loans and grants to provider organizations that lack capital, such as rural hospitals and small, physician-led ACOs (e.g., in the ACO Investment and Advance Payment programs); providing technical assistance to provider organizations (e.g., through the Transforming Clinical Practice program); and providing large incentives in some programs (e.g., an average of $70,045 per clinician per year in care management fees, in addition to the usual fee-for-service payments, in the CPC initiative). Nevertheless, given that CMMI is not required to make its programs budget-neutral, at least in their early phases, it could do more to prime the pump by making it easier for providers to receive an early return on their investments in improving care.

Program evaluation is CMMI’s fourth major challenge: is it adequately evaluating programs to determine their impact and to learn what works and what doesn’t in programs’ day-to-day operations? CMMI uses a “rapid cycle” evaluation approach to address both these questions.⁵ CMMI involves the CMS Office of the Actuary in program and evaluation design, uses standard arms-length government-contracting procedures to award contracts for evaluations, and requires evaluators to use both quantitative and — to enhance learning about how programs and providers operate in practice — qualitative methods.⁵

It is particularly difficult to evaluate the impact of programs – for example, the Partnership for Patients – aimed at involving as many provider organizations as possible in collaborative efforts to improve quality. Critics argue that CMMI should conduct randomized trials when possible and that CMMI is unlikely to be able to provide decisive data on whether some of its largest collaborative quality improvement programs are effective.⁶ When large numbers of organizations are included, it can be difficult to standardize data collection and impossible to find a comparison group. For example, the Partnership for Patients includes 3700 hospitals — most of the hospitals in the country. Even when a much smaller number of provider organizations is included, it can be difficult or impossible to randomize by provider organization or by state. States and organizations cannot be required to join a program; conversely, those that wish to participate but are assigned to a control group may not cooperate with the program.

These problems make it difficult to definitively determine the impact of CMMI’s broad quality improvement collaborative programs, though much is being learned about how these programs work in practice. However, it is much more feasible to rigorously evaluate CMMI’s payment incentive programs — especially the ACO, bundled-payment, and primary care programs — and that is being done. Such evaluation is important because, whereas the quality improvement programs will not become ongoing parts of Medicare, the CMS actuary will decide, for each payment incentive program, whether there is sufficient evidence that the programs controls costs to warrant scaling it up to become a routine part of Medicare.

Republican control of Congress and possibly the presidency make the future of CMMI uncertain. Republicans generally favor using payment incentives to stimulate innovation, so in principle they should support CMMI. In fact, the recent bipartisan bill replacing the sustainable growth rate formula for calculating Medicare’s physician reimbursements specifically advocates increasing the use of such incentives. But CMMI was created by the ACA, which most Republicans vigorously oppose, and many Republicans are skeptical that a government agency can generate innovation.

Not all CMMI programs will be optimally designed or operate at maximum efficiency, and not all will succeed — but much can be learned from failures. If the ACO, bundled-payment, and CPC programs save money for CMS, improve the quality of care and patients’ experiences, and are made an ongoing part of the Medicare program, the impact on U.S. health care will be huge. Indeed, the impact will be substantial if even one of these programs succeeds.

But CMMI’s most important impact may come not from its individual programs, but from its role as a symbol and catalyst of health system transformation. Is the U.S. reaching a tipping point in health care leaders’ belief in the inevitability of movement toward new payment systems and delivery models? Are decisionmakers beginning to change their organizations accordingly? If so, how important is CMMI in changing the way they think?

Health plan ACO programs, the patient-centered-medical-home and patient-safety movements, quality-improvement collaboratives, and many other government, industry, and foundation-supported efforts have contributed to an atmosphere in which change seems both desirable and inevitable. Both the Secretary of the Department of Health and Human Services and the industry-based Health Care Transformation Task Force have recently announced plans to shift the U.S. rapidly toward value-based payment. It is not possible to determine the role of CMMI in these developments, but it is difficult to imagine the Secretary’s announcement occurring if CMMI did not exist to help create programs to make the shift possible. CMMI provides a visible symbol of health care transformation, offers incentives for private-sector innovation, provides cover for insurers who want to move toward value-based purchasing, and creates a constituency of thousands of health care providers and organizations working for change.

The annual $1 billion appropriated for CMMI represents an infinitesimal fraction – three parts in a million – of the $3 trillion that the United States spends each year on health care. At that price, CMMI looks like a good investment.