|Home | About | Journals | Submit | Contact Us | Français|
Discussions of health care delivery and payment reforms have largely been silent about how behavioral health could be incorporated into reform initiatives. This paper draws attention to four patient populations defined by the severity of their behavioral health conditions and insurance status. It discusses the potentials and limitations of three prominent models promoted by the Affordable Care Act to serve populations with behavioral health conditions: the Patient Centered Medical Home, the Health Home initiative within Medicaid, and the Accountable Care Organization. To incorporate behavioral health into health reform, policymakers and practitioners may consider embedding in the reform efforts explicit tools – accountability measures and payment designs – to improve access to and quality of care for patients with behavioral health needs.
In the context of health care reform, various models for delivery and provider payment reform are being discussed, piloted, or implemented. These discussions have largely been silent about how behavioral health could be incorporated into reform initiatives. For example, the final rule for the Accountable Care Organization (ACO) program by the Center for Medicare and Medicaid Services (CMS)1 made little explicit mention of behavioral health care. Of the 33 quality measures to hold ACOs accountable for meeting minimum quality standards, only one (screening for depression) is directly related to behavioral health.
Behavioral health conditions, including mental health and substance use conditions, affected 26% (about 58 million) of U.S. adults in 2001–2003.2 The Affordable Care Act (ACA) will likely add at least 3.7 million individuals with serious mental illness and many more with less severe behavioral health needs to the health insurance system.3 The burden of behavioral health conditions is enormous in terms of pre-mature deaths, disability, lost productivity, and substantial increase in total health care costs.4
This paper draws attention to four patient populations defined by the severity of their behavioral health conditions and insurance status. It is along these two dimensions that important policy initiatives under the ACA may have the potential to serve people with behavioral health needs. These potentials are discussed in the context of three prominent reform models promoted by the ACA: the Patient Centered Medical Home (PCMH), the Health Home initiative within Medicaid, and the Accountable Care Organization (ACO). The paper further discusses tools policymakers and practitioners may consider embedding in the reform efforts to improve access to and quality of care for these patients.
The ACA has several other provisions with direct implications for behavioral health. Examples are revision of Section 1915(i) of the Deficit Reduction Act of 2005 that allows states to use home and community-based services to care for patients with behavioral health disorders and pilots to co-locate medical and behavioral health services.5 We do not address these provisions in our discussion below. The three reform models were chosen to be the focus of this paper because of their combined potential to reach a large population with varying behavioral health needs and because all three models embrace the whole-person approach, making them especially amenable to integration of behavioral health and medical care. For the rest of the article, we adopt the following definition of clinical integration by Shortell et al. (2000)6: “Clinical integration is the extent to which patient care services are coordinated across people, functions, activities, and sites over time so as to maximize the value of services delivered to patients”. (p. 129)
Below we define four populations by the severity of their behavioral health conditions and their insurance status. In a discussion paper prepared for the National Council for Community Behavioral Healthcare, Barbara Mauer proposed a four-quadrant conceptual model that categorizes patient populations by the levels of their behavioral and physical health complexities and needs.7 While our definition of the patient populations bears some relationship to the Mauer model, an important distinction is that we introduce health insurance status into the definition. Patients’ insurance is likely to be a key determinant of which of the three reform models we consider can potentially serve their behavioral health needs.
The first population with behavioral health needs includes patients with mild-to-moderate conditions such as anxiety and mild-to-moderate depression who are insured by a commercial health plan or Medicare, but not Medicaid. These patients most commonly receive care for both their medical and behavioral health problems within medical settings; they may or may not additionally receive care from behavioral health specialists. Frequently, care is not well-coordinated between the medical practice and behavioral health specialists. (Table 1)
The second population includes patients with mild-to-moderate behavioral health conditions who are insured by Medicaid. There are three reasons why the Medicaid-insured are explicitly differentiated from other populations. First, access to care in mainstream medical and behavioral health settings is often poor for Medicaid patients.8 Second, Medicaid patients with behavioral health problems often have a greater need for social and human services than patients with Medicare (but not Medicaid) or private insurance, leading to different policy implications, for example, in the area of service coordination between the health care and human services sectors. Third, several policy initiatives discussed below are relevant to the Medicaid population only. Like the first population, Medicaid patients with mild-to-moderate behavioral health conditions may receive care in both medical and behavioral health specialty settings, and care coordination between medical and specialty behavioral health providers is often lacking. Unlike the first population, these patients are often cared for by practices that specialize in caring for Medicaid and uninsured patients.8
The third population includes patients with serious and persistent behavioral health conditions (such as schizophrenia, bipolar disorders, severe substance dependence and major depression) who have commercial or Medicare insurance but not Medicaid. An example is patients in early stages of schizophrenia who receive insurance from their own or their spouses’ or parents’ workplace. Because of the debilitating and persistent nature of their conditions, these patients often transition to Medicaid over time.9 This paper does not explicitly discuss them below. This paper also does not consider children with behavioral health problems, the complex needs of whom may indicate different policy implications and deserve a separate discussion.
The fourth population includes patients with serious and persistent behavioral health conditions who are insured through Medicaid, including patients who are dually enrolled in Medicare and Medicaid. Typically, these patients’ primary contact with the health system is through a facility or provider that specializes in behavioral health care (e.g., a community behavioral health center). This population is at heightened risk of developing serous medical conditions such as hypertension, elevated cholesterol, and diabetes.10 However, their behavioral health providers often lack the capacity or incentives to effectively manage patients’ co-morbid medical conditions or to provide preventive and wellness care.11, 12
This section introduces and defines each of three reform models that have the potential to coordinate care for patients with behavioral health needs.
The concept of PCMH combines the tenets of primary care (first contact, comprehensive and coordinated care) with systematic improvement of the health of the practice’s patient population (via use of electronic information systems, disease management, continuous quality improvement, and so on).13, 14 As a prominent model to reform primary care, PCMH has been implemented in nearly fifty demonstration projects in 30 states.15 A large Medicare demonstration is currently underway.
The joint principles of the PCMH issued by major physician organizations16 and the public recognition standards specified by the National Committee for Quality Assurance (NCQA)17 define the attributes and functions of a PCMH. Because primary care is where most patients interact most frequently with the health care delivery system, primary care practices are in a natural position to become PCMHs. Meanwhile, the reform law does not preclude specialty practices from becoming PCMHs if the specialists address the majority of the needs of the patient and meet the same standards for primary care-based PCMHs.13
The ACA’s Health Home provision provides states with additional Federal funding (90% federal matching rate within the first 2 years of health home establishment) to pay for care management, coordination, and use of clinical information technologies by Medicaid providers. Minimum patient eligibility for receiving health home services includes having two or more chronic conditions, or one chronic condition and being at risk for another, or having a serious mental health condition. According to the legislation, providers of health home services may be a “designated provider” (Section 1945(h)(5)), a “team of healthcare professionals” (Section 1945(h)(6)), or a “health team” (Section 1945(h)(7)). In particular, examples of “designated providers” include but are not limited to physicians, clinical practices or clinical group practices, rural health clinics, community health centers, community mental health centers, and home health agencies. Many states are currently undergoing the planning stage for health homes.
According to CMS’ preliminary guidance to states,18 the Health Home model will build on states’ past experience with medical homes for the Medicaid population. However, the implementation of the Health Home will aim to “expand the traditional medical home models to build linkages to other community and social supports, and to enhance coordination of medical and behavioral health care, in keeping with the needs of persons with multiple chronic illnesses.” Therefore, the Health Home differs from the PCMH in that the former is a legislative provision specifically for Medicaid patients with special emphasis on care coordination activities involving community organizations that are critical in meeting the needs of Medicaid patients.
Accountable Care Organizations (ACOs) are provider-led organizations that manage the full continuum of care and are accountable for the overall costs and quality of care for a defined patient population.19, 20 A central design element of ACOs is the shared savings mechanism. Under this mechanism, ACOs could receive fee-for-service payment and share in the savings achieved relative to a risk-adjusted, pre-specified spending target for their patient population; alternatively, payment could be partially or fully capitated, with ACOs sharing in both risks and gains.1 The shared savings mechanism of ACOs arguably provides strong incentives for coordinating medical and behavioral health care. This is because behavioral health conditions are prevalent and strongly associated with treatment non-adherence,21 adverse health events and substantially increased total costs.4 Behavioral health thus represents an area that could be especially fruitful for ACOs to target.
The final rule for Medicare ACOs outlines seven types of provider organizations that are eligible to independently become an ACO.1 This expanded list, compared to what was proposed earlier,22 includes several important safety-net providers such as the Federally Qualified Health Centers (FQHCs). According to the final rule, ACO patient assignment will be based on utilization of primary care services provided by ACO professionals; minimum patient enrollment is 5,000.
It is worth noting that ACOs and PCMHs/Health Homes are not mutually exclusive. Realization of the ACO goals will require a strong primary care core. The ACO infrastructure beyond the primary care will enable the full realization of the PCMH/Health Home model.14 It is thus highly likely that ACOs will be built on a base of medical homes.
This section discusses the potentials and limitations of each of the three reform models in coordinating care for patients with behavioral health needs. It further discusses potential policy tools to better incorporate behavioral health into health care reform, including 1) accreditation standards, eligibility rules, and performance measurement and rewarding systems, collectively referred to as accountability measures, and 2) payment policies. (Table 2)
Our definition of the patient populations and our discussion below does not explicitly consider patients with behavioral health needs who are uninsured. This is because, first, for all three reform models we consider, payment plays a critical role in enabling and incentivizing the providers to deliver care the way intended by the model. This incentive does not exist for uninsured patients. Second, when the ACA is fully implemented, about 60% of the currently uninsured population with behavioral health needs will gain coverage through private insurance (including insurance exchanges) or Medicaid expansion.3 Our discussion pertaining to patients with private insurance or Medicaid apply to these patients. For patients who would remain uninsured after implementation of the ACA, the reform efforts may indirectly affect the way they receive care if providers of these patients (for example, the FQHCs) experience changes in care processes and quality that affect all patients regardless of insurance status. Meanwhile, it is important that federal and state governments preserve the direct-service funding to provide care to the uninsured.23
A PCMH may have the greatest potential to serve patients with mild-to-moderate behavioral health conditions (Populations 1 and 2). Patient-centeredness and the whole-person approach, two central qualities of PCMHs, are essential in meeting the needs of these patients since medical and behavioral health conditions frequently co-exist in this population and should be co-managed. Clinically, strong evidence supports a primary care-based, collaborative care approach to manage behavioral health conditions.24 A recent study demonstrated the potential of managing commonly-occurring “dyads” or “triplets” of behavioral health conditions and medical conditions.25
However, unless it is very large, a PCMH may lack the capacity to care for patients with serious behavioral health conditions (Populations 3 and 4).26 It may not have enough patients with serious behavioral health conditions to develop experience in referrals to a competent network of behavioral health providers or to coordinate care with them. PCMHs based in private practices may also not accept Medicaid patients (Populations 2 and 4). However, the Health Home provision in the ACA (discussed below) may offer options for creating “medical home” equivalents for Medicaid patients in either a primary care setting (for Population 2) or behavioral health specialty setting (for Population 4).
Accreditation is a central element of medical home design policies. The 2011 update of the National Committee for Quality Assurance (NCQA) accreditation standards has several provisions to require medical homes to meet their patients’ behavioral health needs.17 These include requirements for: 1) routine screening for behavioral health conditions, and 2) implementation of evidence-based guidelines for the management of one “health behavior” or one mental health/substance use condition in addition to two chronic medical conditions deemed important to the practice. Future refinement of PCMH standards may consider requiring practices to implement evidence-based guidelines for at least one behavioral health condition rather than keeping it optional. It may also require PCMHs to manage medical and co-existing behavioral health conditions simultaneously using evidence-based models.
Some PCMH accreditation systems (for example, the NCQA standards) adopt a tiering method to reflect different levels of capabilities to perform patient-centered activities. To enhance accountability of PCMHs for meeting patients’ behavioral health needs, the scoring system on which the tiers are based might be designed to reflect the varying extent of behavioral health integration. For example, such a system may assign a higher score to a practice that implements the collaborative care model for behavioral health conditions than if it merely tracks referral to behavioral health specialists.
Currently, the dominant payment arrangement for a PCMH is a per-patient-per-month care management and coordination payment in addition to traditional fee-for-service payment. Two policy tools may be available to embed incentives for integrating behavioral health into medical homes. One is to factor behavioral health conditions into the risk adjustment scheme of the per-patient-per month payment, as is seen in Minnesota’s state-wide medical home initiative.27 This tool recognizes the additional resource requirement for meeting patients’ behavioral health needs, thus tempering incentives to avoid patients with behavioral health conditions.23 Where PCMHs are tiered to reflect different levels of behavioral health integration, a second tool is to have higher-tier homes receive higher payment.
Medicaid patients with mild-to-moderate behavioral health conditions (Population 2) may be best served by primary care-based Health Homes. Medicaid patients with serious behavioral health conditions (Population 4) may most likely be served by Health Homes based in behavioral health specialty settings (for example, community behavioral health centers). Some primary care-based Health Homes who see a large number of patients with serious behavioral health conditions may have developed referral and care coordination programs with behavioral health and social service providers. A very small number of primary care-based Health Homes may also be community behavioral health centers; they have internal programs to provide and coordinate primary and behavioral health care through the care continuum, as in the case of Cherokee Health Systems in Tennessee. These primary care-based Health Homes have the potential to serve Population 4, but are likely to be small in number. The discussion below focuses on recommendations for behavioral health specialty Health Homes.
Behavioral health specialty Health Homes should focus on integrating primary care with specialty behavioral health care. Two organizational approaches are possible.26 First, these Health Homes may staff primary care clinicians (for example, advanced nurse practitioners or physician assistants) on-site, who receive supervision and consultation from an off-site primary care physician. Alternatively, they may partner with an off-site primary care provider (e.g., a federally qualified health center). Such partnership may be facilitated by an on-site care manager and take the forms of referral, clinical information-sharing, follow-up and case conferences. Both approaches are being implemented and evaluated in the primary care co-location grant program established by the Substance Abuse and Mental Health Services Administration and later expanded by the ACA.28
Even with these efforts to strengthen primary care, a behavioral health specialty Health Home may still have difficulty providing or coordinating medical care to its patients. Specifically, access to specialty medical care outside of emergency settings (e.g., to cardiologists, endocrinologists, and orthopedists) has been challenging for patients with serious behavioral health conditions because of the challenge of their conditions, because specialists are often reluctant to care for Medicaid patients, and because of the long-standing separation of behavioral health specialty from the rest of medicine.
The ACA provides great flexibility for states to design rules for designating providers as Health Homes. Policymakers are suggested to take the following into considerations when designating behavioral health specialty Health Homes. First, such Health Homes should meet minimum capacity requirement for primary care and preventive care, enabled by one of the two organizational approaches outlined above. Second, these Health Homes should implement mechanisms to ensure access to and coordination of care with medical specialty providers important for this population.
Section 1945(g) of the ACA requires designated providers of Health Home services to report to the State on all applicable quality measures as a condition for receiving payment. The CMS will develop a core set of quality measures.18 Given the special emphasis on coordinated behavioral health and medical care by the Health Home provision, CMS and states are strongly recommended to include measures of patient experience with coordinated behavioral health and medical care when developing this core set of measures.
The ACA and CMS guidance also provide states with flexibility regarding Health Home payment designs. Both a fee-for-service approach (a new line item in Medicaid claim) and a per-member-per-month approach are possible. Two strategies may strengthen incentives for primary care integration and care coordination. First, case-mix adjustment of the payment should reflect patients’ co-morbid medical conditions as well as behavioral health conditions. Second, if states choose to adopt a tiering system to reflect varying capabilities of providers to perform health home functions, the tiering method and accordingly, the payment, should differentiate between homes with distinct levels of primary care integration. The tiering system should also reflect the extent to which the Health Home provides its patients with access to other medical specialty care.
Nascent mainstream ACOs – those that focus on Medicare and/or commercially-insured populations - have the greatest potential to serve non-Medicaid patients with mild-to-moderate behavioral health conditions (Population 1). These patients are most likely to receive regular primary care from a physician who belongs to a mainstream ACO. Furthermore, mainstream ACOs are likely to have the scale (a relatively large number of patients with mild-to-moderate behavioral health conditions) and devoted resources to ensure access to and coordination with high-quality behavioral health specialists for this population.
Because ACOs are payer-specific, mainstream ACOs participating in shared savings programs with Medicare or commercial insurance plans will lack incentives to improve quality and reduce costs for Medicaid patients (Populations 2&4).
Pressed by spiraling Medicaid cost growth, some states are exploring the idea of regional ACOs for their Medicaid population, as seen in the state of New Jersey 29 and the state of Colorado.30 Because an adequate size of membership is critical to achieve ACO goals, Medicaid ACOs are most viable in states whose Medicaid populations are sizable and concentrate geographically, for example, in urban areas. High concentration of Medicaid patients among a small number of practices (for example, community health centers or hospital-based clinics)8 is likely to add to the viability as a result of concentrated, rather than dispersed, incentives faced by providers. These Medicaid ACO initiatives, if successful, may benefit Medicaid patients with various behavioral health needs. Specifically, they may deploy programs to improve care coordination between primary care and behavioral health specialty providers for the safety-net population, and to strengthen linkages to social and human services.
There are two areas in which ACO rules may play a role to incorporate behavioral health. First, such rules may require ACOs to secure members’ access to a sufficient array of behavioral health services and to have a mechanism in place for care coordination. Many behavioral health providers work in solo or small practice settings, independent from physician group practices and hospital-physician associations. They are also less likely to be part of health plan physician networks. Thus, for ACOs who do not have an adequate number of behavioral health providers within their own organizations, an alternative would be to partner with large behavioral health providers (for example, the Sheppard Pratt Health System in Maryland). Through such partnerships, ACOs may implement care coordination programs across medical and behavioral health settings. They may also push for adoption and use of information technology by behavioral health providers to facilitate integration.23 In return, behavioral health providers may obtain preferred referral status with the ACO, and receive assistance to build up the infrastructure for care coordination (for example, implementation of a data collection and monitoring system to track patient outcomes across settings).
Second, ACO performance measures should include well-vetted, standardized measures for high-quality behavioral health care.23 Measures in the final rule for Medicare ACOs are notably lacking in the domain of behavioral health care, with only one out of 33 measures directly related to behavioral health. NCQA’s ACO criteria specified several, largely process-based, measures regarding the treatment of behavioral health conditions. In contrast, of an initial core set of 51 quality measures recently issued by the Department of Health and Human Services for state Medicaid programs, 11 were behavioral health-specific measures.31 Five of the 11 concern the care of patients with serious and persistent behavioral health conditions (bipolar and schizophrenia). This set of measures will have important implications for the design of quality measures for Medicaid ACOs.
More importantly, measures should strike a balance between structures and processes associated with evidence-based care, and outcomes; they should guide ACOs in building the infrastructure and systematic processes to meet patients’ behavioral health needs.32 An example is a quality measure that requires the use of standard assessment tools for behavioral health conditions throughout the care continuum.33 Such a measure would spur implementation of standard assessments in the ACO’s clinical system and adoption by providers in diverse settings.
In theory, the shared savings mechanism provides ACOs with strong incentives for coordinating behavioral health and medical care. The challenge for both ACOs and policymakers is to make the incentives work as intended. The payers (Medicare, commercial insurance plans, or Medicaid) may consider sharing data and information with ACOs to help them identify patient subpopulations with unmet need for behavioral health care. ACOs may make use of these data to identify the settings and timings where interventions could be most fruitful, and to demonstrate the business case for investing in coordinated behavioral health and medical care. Also of importance is knowledge-sharing and group-learning among ACOs to accelerate adoption of best practices of care coordination, as the Center for Medicare and Medicaid Innovation is poised to support.
Key reform initiatives under the ACA hold promise for improving the quality and coordination of care for people with varying severity of behavioral health conditions.5, 23 To fulfill this promise, policymakers and practitioners should first identify the patient populations whom a specific reform model has the greatest potential to serve. We have shown in this paper that a definition of the patient population by the severity of their behavioral health conditions and their insurance status could be a useful starting point. Tools such as accreditation criteria and payment designs can then be used to explicitly incorporate behavioral health into reform efforts.
To inform this process, future research may seek to further shed light on how patient populations with varying behavioral health needs interact with different providers and how reform models may serve the needs of different populations. It would be valuable to identify critical settings and timings where poor coordination of behavioral and medical care contributes to poor overall patient and cost outcomes. It would be particularly useful to implementers of reform initiatives to know what proven clinical and delivery models are available for which specific settings and populations, what the business case is (in terms of cost of implementation relative to pay-offs), and how to implement and sustain the programs.
The general focus of our paper is on patients with behavioral health – mental health and substance use – conditions. Special challenges and opportunities exist for integrating substance use treatment with medical or even mental health care in the context of the reform models. These challenges exist in the domains of financing mechanisms, policy and regulations, workforce preparation and culture differences, clinical informational systems, physical facilities, and relative lack of research evidence supporting integration models involving substance use treatment.34 We reserve this as a topic for future investigation.
Dr. Bao received funding from the National Institute of Mental Health (K01MH090087, P30MH085943). Dr. Pincus received funding from the Irving Institute for Clinical and Translational Research at Columbia University (UL1 RR024156) from the National Center for Research Resources, a component of the National Institutes of Health; Mental Health Therapeutics Center for Education and Research in Therapeutics (CERT) at Rutgers. Additional funding for Dr. Pincus came from government and nongovernment organizations of the countries participating in the International Initiative for Mental Health Leadership project (Australia, Canada, England, Germany, Ireland, Japan, the Netherlands, New Zealand, Norway, Scotland, Taiwan, and the United States).
Conflicts of Interest
Drs. Bao and Casalino report no conflicts of interest. Dr. Pincus has submitted a detailed statement of potential conflicts of interest, which remains on file at the Journal’s editorial office; there appears to be no conflicts of interest.
Dr. Yuhua Bao, Department of Public Health at Weill Cornell Medical College.
Dr. Lawrence P. Casalino, Department of Public Health at Weill Cornell Medical College.
Dr. Harold Alan Pincus, Department of Psychiatry at Columbia University and the NewYork-Presbyterian Hospital.