SPECIFIC AIMS
Improving the value and performance of our health care system is one of the most important challenges facing our nation. However, despite considerable desire to both improve quality and “bend the cost curve,” recent delivery and insurance demonstrations have had limited impacts on the cost of care. The need for cost control is particularly acute for Medicaid programs, which cover a population with a high prevalence of multiple chronic illnesses, and face the prospect of continued budgetary constraints and increased enrollment from the Affordable Care Act (ACA).
In 2012, the State of Oregon will undertake an innovative and substantial restructuring of its Medicaid program in its initiation of Coordinated Care Organizations (CCOs). In many ways, CCOs reflect the national movement toward accountable care and primary care medical homes. However, Oregon’s CCOs differ from other pilots or demonstrations in two significant ways. First, savings in Oregon’s CCOs are enforced with a transition to a global budget that includes a large (11%) reduction in spending in the first year. Second, Oregon's program integrates payment for behavioral and physical health care, including funds that have been traditionally “carved out” for mental health and addiction treatment, giving CCOs flexibility in how these dollars are spent.
The CCO experiment provides a unique opportunity to assess two key economic questions of primary importance to public and private payers across all states. First, faced with a large, exogenous decrease in funding, how are savings accomplished? Is it primarily through reductions in payment rates, or through delivery systems that reduce unnecessary and expensive utilization? Second, what are the consequences of financial integration of behavioral and physical health care? Our proposal has four specific aims:
Specific Aim 1. Determine how a large exogenous decrease in spending in Oregon affects utilization, payment rates, and the composition of provider networks.
Specific Aim 2. Determine the impact of exogenous decreases in spending in Oregon on the quality of care.
Specific Aim 3. Determine the effect of bundling payments for behavioral and physical health care on the integration of care between behavioral and physical health care providers
Specific Aim 4. Complete a qualitative assessment to understand the key governance, organizational, and contextual variables affecting the transition to a restructured Medicaid program.
Our goal is to use the NIH Common Fund Mechanism to test the effects of these changes, which cross disease boundaries and have widespread generalizability to states, federal reform efforts, and commercially insured populations. This distinctive study frames Oregon as the intervention state and uses Colorado as a comparison state. Like Oregon, Colorado has restructured its Medicaid program with the creation of Regional Care Collaborative Organizations (RCCOs). The programs in Colorado and Oregon are similar, sharing elements of local control (coordinating care in a geographically contiguous area), an emphasis on primary care medical homes, and a reliance on performance benchmarks to monitor and encourage high quality care. The primary differences – and the foci of this proposal – are the enforced spending cuts in Oregon and the financial integration of behavioral specialty care and primary health care.
There is broad agreement that the United States must slow rising health care costs, yet little consensus exists on how best achieve this goal. Furthermore, appreciation for the connection between behavioral and physical health is growing, but uncertainty remains about the best way to implement and finance this integration. This project will apply econometric methods to generate new, important results on the cost savings and potential quality enhancements that can be realized from the natural experiment underway in Oregon. While the focus of this study is quantitative, we overlay a qualitative component to identify not only what works, but also why and where certain elements promote productive delivery systems. Our study is specifically designed to decouple the specific impacts of spending reductions and financial integration from the contemporaneous policy changes (such as reliance on the medical home) that are occurring in both Oregon and Colorado. The application will be strengthened through the assembly of a national advisory committee that will help assess the key issues around behavioral/physical health integration and identify the concerns of Medicaid officials. Together, accomplishing these aims will provide crucial knowledge for the development and design of innovations necessary to improve the long term value and affordability of health care.
RESEARCH STRATEGY
A. SIGNIFICANCE
We propose to fill knowledge gaps in two critical areas of health care delivery. We seek to understand (1) how reductions in spending are accomplished within a CCO and (2) how financial integration affects behavioral and physical health care.
A.1. Background on Oregon and Colorado. Colorado serves as an excellent comparison group because the state has already implemented their Medicaid policy changes in response to the ACA. Following national trends, their program (like Oregon’s) places heavy emphasis on patient-centered primary care homes and accountability for quality measures. Colorado began enrollment of individuals into its RCCOs in early 2011, with assignment to a primary care provider based on existing relationships. Oregon's CCO framework was established in 2011 and CCOs will launch on July 1, 2012. Table 1 summarizes key differences.
Table 1 Characteristics of Coverage Models
Coverage Model / Regional Focus / Primary Care Medical Homes / Accountability Measures / Cost Savings / Specialty and Ancillary Care / Financial Connection to Behavioral HealthCCO (Oregon) / Regional and community resources support the coordination of care across programs / All beneficiaries are assigned a primary care medical home, which serves as the primary agency for coordinating care / Core measures that align with national initiatives / Built-in through up front reduction; global budget / CCOs at-risk under bundled payment / Yes, integration of funding for physical and behavioral health
RCCO (Colorado) / Same as Oregon / Same as Oregon / Same principle as Oregon, although measure sets differ between states / Incremental reductions anticipated / RCCO is not at risk; paid FFS / No, behavioral health carved-out though RCCOs may also be related to carve-out
Oregon’s CCOs differ from Colorado in two important ways. First, initial cost savings in Oregon will be built into the global budget, with the equivalent of an 11% cut in spending planned in the first year. Second, in Oregon, specialty care and behavioral health care will be brought under the global budget. Oregon’s CCO funding captures the majority of available streams for Medicaid populations – including approximately 20% of funds that have been assigned to mental health “carve-outs” and 10% assigned to other services (e.g., non-emergency medical transport) – and gives CCOs flexibility in how these dollars are spent. Specialists in Colorado will be paid fee-for-service while behavioral health care remains under a regional carve-out.
A.2. Recent Experience with Programs Designed to Reduce Spending. Public and private purchasers desire delivery systems that can improve efficiency: care that attains a given level of quality at the least cost. Several high profile demonstrations have attempted new reimbursement mechanisms with the potential for shared savings, but the results have been mixed. In particular, while these innovations have generally succeeded in improving quality, their impacts on costs have been relatively minor. For example, the Medicare Physician Group Practice (PGP) – the prototype for Accountable Care Organizations – included 10 large multispecialty groups that were eligible to receive shared savings if they met quality goals and exceeded a savings threshold of 2%. However, while the 10 groups did very well on achieving their quality metrics, only 2 of the participants were able to exceed a 2% savings threshold in the first year, and only half surpassed that threshold after 3 years.1,2 The Blue Cross Blue Shield of Massachusetts’ Alternative Quality Contract (AQC), launched in 2009, signed 12 medical groups and used a global budget tied to quality metrics, with groups with higher quality scores eligible for greater shared savings. The AQC was associated with an improvement in quality and a reduction in the growth of expenditures, although the reduction was modest (7% growth in spending in the AQC compared to 8.8% growth in a control group), and spending by the health plan on bonuses for quality was likely to have exceeded total savings.3,4 The Group Health Medical Home (GPMH) showed improvements in patient experience, quality, and clinician burnout, with savings estimated at $120 per member per year (approximately 2% of total costs).5 The Community Care of North Carolina (CCNC) is a statewide restructured Medicaid program with enhanced medical homes. An actuarial analysis estimated that CCNC saved the state between 2% to 5%.6,7 A recent review by the Congressional Budget Office evaluated the effects of 10 major disease management and care coordination programs and concluded that most have not reduced Medicare spending.8 Taken as a whole, these studies suggest that these approaches – medical homes emphasizing coordinated care, with shared savings based on meeting quality targets – are successful at improving quality but have demonstrated only modest improvements in the cost of care.
Our study aims to provide valuable information on achieving cost savings. Our contribution stems from the study of Oregon, which is enforcing an 11% spending cut in its transition to CCOs, and from our econometric specification. Our study is designed to elicit causal inferences and to provide detailed and policy-relevant information about how savings are ultimately achieved, and how these changes affect the quality of care. In particular, our study is designed to test whether savings are achieved primarily through reductions in payment rates and selective contracting (which appears to have been the primary mode by which managed care organizations in the 1990s achieved savings9,10), or through more appropriate or “efficient” care, which is the hope of many policy-makers. Knowing what savings can be squeezed from integration and care coordination and how much must come from reductions in price is a critical test of new health care models. Furthermore, it is crucial to determine whether large decreases in spending can be managed without reducing the quality of care.
A.3. Financing the Coordination of Behavioral and Physical Health Care. A large number of studies – including more than 30 randomized trials – have demonstrated that integrated primary care and behavioral health care has been associatedwith improved quality and patient outcomes.11-16 Furthermore, many programs that focus on integration and care management have been associated with lower health care spending, but results are mixed as to whether these savings offset the initial investments.14,17 Despite the widespread interest in this area, and the ongoing piloting of integrated and collaborative models, relatively little is known about how the methods of financing affect the degree of coordination and its effectiveness. The literature is clear that there are gains from coordination, but there is insufficient evidence about how to optimally structure reimbursement to sustain coordination. Typically, Medicaid programs pay separately for addiction care, mental health case management, psychiatric day treatment, and related services. The State of Oregon has identified over 26 separate programs (most of which relate to behavioral health care) that would be moved under the global budget. This integration may lead to innovations and improvements in the ways in which the whole patient is treated. Our study will identify the benefits and potential adverse effects of financial integration (Oregon’s CCOs), compared to a model where the funding streams remain separate (Colorado’s RCCOs).
B. INNOVATION
The primary innovation of this project builds on the unusual changes in Oregon's Medicaid transformation: a large reduction in spending coupled with financial integration of physical and behavioral health care. Although many models promoted by the ACA are similar in design to those adopted by Oregon and Colorado, Oregon has pushed forward more aggressively with cost saving and integration.
The regional nature of the Oregon and Colorado programs eliminate incentives to cream-skim healthy patients or skimp/stint to avoid patients with high expected costs. The lack of patient selection incentives provides us with a relatively clean natural experiment to study the effect of major changes in provider incentives and overall reimbursement. Furthermore, the use of Colorado as a comparison group allows us to control for the emphasis on primary care homes and accountability that are part of a larger national trend. The unique nature of Oregon's CCO's enables us to focus on the effect of mandated declines in per capita spending and the effectiveness of financial integration of primary care and behavioral health care. This study will capitalize on the unique opportunity to examine the effects of dramatically large changes in incentives in a Medicaid program, using a well-matched contemporaneous group to control for distinctive program features.
C. APPROACH
The study has four specific aims. Aims 1-3 use econometric methods to assess the effects of spending reductions and integration. These methods use the common difference-in-difference framework, but are extended to allow for decomposition of the effects of interest. In Aim 4, we will use qualitative methods to more fully understand the organizational structure and services provided by Oregon’s CCOs and Colorado’s RCCOs. This inquiry will also place in context the importance of factors that may lead some CCOs (or RCCOs) to be particularly successful (or unsuccessful) in controlling costs, improving quality, or achieving integration.
C.1. Overview of the Research Team
The proposed project is founded on an existing collaboration between investigators with backgrounds in economics, behavioral health services, and primary care medical home models. The research team has a long history of research on state health policy, insurance design, and mental health services, and extensive experience in working with State officials to acquire data and share results. The Oregon team is led by K. John McConnell, PhD (Oregon Health & Science University) and Neal Wallace, PhD (Portland State University). The Colorado team is led by Richard Lindrooth, PhD (University of Colorado Denver). Drs. McConnell, Wallace, and Lindrooth will be assisted on the qualitative analysis by Debbie Cohen, PhD (Oregon Health & Science University) and Jeanette Waxmonsky, PhD (University of Colorado Denver). The project team also includes: Bentson McFarland, MD PhD, (Oregon Health & Science University) a psychiatrist and health services researcher; Dennis McCarty, PhD, (Oregon Health & Science University) a health services researcher with a focus on addiction treatment; David Dorr, MD (Oregon Health & Science University) a primary care physician who has worked closely to develop medical home models.