Collaborative Care Models for Comorbid Medical and Behavioral Health Conditions

Clinical Context

A series of converging factors are leading to increased interest in integrated care. With the passage of the Affordable Care Act, attention is being focused on improving health outcomes and containing costs. The impact of mental and substance use disorders—hereafter referred to as behavioral health—on reaching these goals is being appreciated in new ways. Across the range of healthcare settings, opportunities to implement outcome-changing interventions are underway and being tested.

One in five U.S. citizens has a diagnosable mental disorder (1), with only 40% receiving any treatment for their condition. Of those who do receive care, only half of them see a behavioral health specialist (2), with the majority being treated in physical health settings by primary and specialty medical care clinicians or alternative medicine settings or social service agencies. In the primary and specialty medical care outpatient setting, patients with behavioral disorders are often not recognized or engaged in effectively delivered treatment, resulting in a mere 13% of patients receiving minimally effective treatment (3). In the inpatient medical setting, payment shortfalls and challenges in billing for psychiatric care make it difficult to access behavioral health services either in emergency rooms or on general medical and surgical units. The impact of untreated mental illness on total healthcare costs is significant (4) and can lead to decreased adherence to recommended medical/surgical treatments and lack of follow-up for care. These poor outcomes, coupled with the increase cost in overall health care associated with untreated behavioral illness, have led to the need for better treatment options in primary care and other medical settings. In addition, the lack of a sufficient number of psychiatrists to meet treatment demands necessitates developing models that leverage these limited resources more effectively.

Patients with serious mental illness have a life span shortened by as much as 25 years (5), with 60% of this attributable to untreated chronic medical conditions, leading to increased cardiovascular morbidity in particular (6). As seen in Figure 1, there is a high rate of physical illness in patients with behavioral illnesses (7). Further, the second-generation antipsychotics have a well-known contribution to increased cardiovascular risk (8). A number of other factors also contribute to this mortality, including failure to recognize medical and psychiatric disorders, lack of access to care, poorer quality of care, and inability to afford medical care. A look at baseline data prior to initiating the CATIE study showed a high prevalence of “current” untreated illness, including 30% of patients with diabetes, 66% with hypertension, and 88% with hyperlipidemia (9) with these disorders not receiving care.

In addition, high utilizers of health care resources, i.e., the 5%−10% of patients consuming 50%−70% of the health care dollars, as a group are being targeted for interventions (10). An estimated 70% of these have comorbid behavioral health problems (11). These high utilizers can be found in the population of patients with serious mental illness in public behavioral health settings as well as in general primary care, emergency rooms, and inpatient medical settings. They present an opportunity for collaborative medical and behavioral health interventions designed to improve outcomes and reduce costs.

Outpatient Treatment Strategies and Evidence

Terminology

The terms “collaborative” and “integrated” care are often used interchangeably, although in the former it is implied that independently operating medical and behavioral health groups support care coordination. A recently published lexicon defines integrated care as “the care that results from a practice team of primary care and behavioral health clinicians, working together with patients and families, using a systematic and cost-effective approach to provide patient-centered care for a defined population.” (12). Over the past decade, collaborative care models have often been considered to fall into two camps determined by the primary location of care. These describe common methods for service delivery when payment for medical and behavioral health services are separate. In the first camp, efforts are focused on “integrating” detection and treatment of behavioral health issues into the primary care setting. In the second, the goal is to improve health outcomes in patients seen in the behavioral health sector by addressing comorbid chronic medical problems through the introduction of medical professionals in the behavioral health setting, so called “reverse integration.” Each will be addressed with the understanding that some core principles apply to both and to integrated care in general. For the purposes of this article integrated care and collaborative care will be used interchangeably.

A third model, which is less common than the first two due to the misperception that medical and behavioral health patients fall into two distinct categories (i.e., those with primary medical problems need treatment only in the primary care sector and those with primary behavioral health problems need treatment only in the behavioral health sector), is one in which medical and behavioral health professionals participate in and are accountable for total health of the population they serve regardless of setting. In this model, behavioral health and medical care are considered core components of total health and are part of a single network of clinicians. In this model all practitioners, regardless of background, would be considered full team members. This comes close to being realized in level three patient-centered medical homes but remains a challenge to develop, since independent medical and behavioral health payers continue to drive separate delivery of medical and behavioral health services by maintaining segregated medical and behavioral health provider networks and payment systems that do not incentivize this care.

Integrating Behavioral Health into Primary Care

A robust evidence base for collaboration in primary care settings provides a framework on which to build an effective integrated delivery system. The preponderance of evidence comes from the Improving Mood Across Collaborative Treatment (IMPACT) trial and its many replications throughout the past decade. These studies have shown improved outcomes in depression treatment, lower overall healthcare costs, and patient satisfaction with the care received (13). Over 70 randomized trials in diverse populations conducted in primary care, mostly for the treatment of depression, have proven its effectiveness (14, 15). Another model that adapts a similar design, known as TEAMCare, includes targeted treatment of other chronic medical conditions in addition to depression. This trial demonstrated these interventions can lead to improvement in both medical and behavioral health indicators, such as HbA1c levels, blood pressure control, cholesterol levels, and Patient Health Questionnaire-9 (PHQ-9) scores in addition to saving on costs of outpatient healthcare (16). Both the IMPACT and TEAMCare models provide systematic approaches to care that meet the goals of the Triple Aim (17) and position primary care clinics to be accountable and reimbursed for the quality of care. In the most recent meta-analysis, Woltmann and colleagues (18) demonstrated that using collaborative chronic care models can improve outcomes in the treatment of various mental and physical health outcomes (see Figure 1 of the Woltmann et al. article, which is reprinted in this issue of Focus [p. 557]).

Central to the design of the collaborative care studies that form the evidence base is the Wagner Chronic Care Model (19). It emphasizes care delivery redesign through utilization of proactive care teams and “activated” patients. In May 2011, a summit of leaders met at the Advancing Integrated Mental Health Solutions (AIMS) Center at the University of Washington in Seattle. A set of Core Principles of Effective Collaborative Care (http://uwaims.org/overview-principles.html) were established and can serve as a useful tool for understanding what elements are considered crucial to effective interventions. They help distinguish collaborative care integration from other forms of care. The 5 Core Principles are as follows:

Utilizing these core principles, collaborative teams provide a system of care that begins with screening and patient identification through use of tools, such as the PHQ-9 for depression, moves on to patient engagement, and then provision of evidence-based brief interventions and pharmacotherapies. Registries allow tracking and adjustments in treatment as needed. Team members utilize aggregate data to improve the overall care. Treatment in integrated settings follows a system of “stepped care” (20) to allow effective treatment to be provided with the least amount of intervention necessary at that level. The person gets the care they need without over utilizing resources, leading to improved outcomes that are cost efficient.

The treatment team in the primary care setting adds two new members to accomplish these goals: a care manager, who may or may not be a behavioral health provider, and a consultant psychiatrist (Figure 2). The care managers (often trained in social work or nursing) provide evaluation and brief evidence-based interventions in addition to keeping track of patients in registries. They are usually located onsite with the primary care providers and are available for “warm handoffs” for immediate assessment and intervention. The consultant psychiatrist provides “curbside” (informal) consultations as needed to both the care manager and the primary care provider, conducts systematic registry review to discuss patients that are not improving and, less frequently, provides formal consultation with patients that are complicated and not improving. Utilizing this model, only a small number (5%−7% in the IMPACT trial) of patients are seen face-to-face by the psychiatrist. This allows extension of psychiatric expertise to a larger population of patients without suggesting that other face-to-face individual care should not be provided by psychiatrists (21). Psychiatrists best suited for working in the fast-paced cadence of primary care are those that are flexible, willing to be interrupted, agreeable to consultation across the life span if needed, and those not overly concerned about liability for curbside consultation. Their job is to continually support the onsite team in the clinic by phone or other communication methods as needed, allowing more immediate and accurate diagnosis and the implementation of evidence-based treatments.

Unfortunately, the most common form of integrated care in primary and specialty medical settings in the United States today, i.e., the addition of a nondoctoral psychologist or social worker into the medical clinic setting, does not provide these Core Principles. Not surprisingly, meta-analysis of clinical approaches in the United Kingdom, nearly identical to these now being introduced in the United States, have demonstrated that long-term outcomes and cost savings cannot be expected (22). Thus, it is important that the Core Principles are present so that outcome improving and cost savings, i.e., value-added care, can be expected as collaborative services are introduced and delivered.

Integrating Primary Care into Behavioral Health Settings

Determining best practices to promote evidence-based improvement in the health status of persons seen in public behavioral health settings, many of whom have serious mental illness, is a major ongoing area of research, lagging behind its better known counterpart in the primary care setting. Consequently, a robust evidence base for this practice has not been established. However, emerging examples of approaches exist that mirror the collaborative care models described above. These add care managers and specialist consultation to the healthcare team in integrated behavioral health settings. The Patient Care Access, Referral and Evaluation (PCARE) study (23) was one of the first studies published that showed if enhanced efforts by care managers to connect behavioral health patients to offsite primary care services was provided, measures of chronic medical illness could be improved. These efforts included serving as advocates and a communication link between the patients with serious mental illnesses and the primary care providers, providing education on health-related issues and support to dealing with the multiple system barriers that make this process even more difficult for this population.

Two new models are currently being tested nationally. The first is the SAMHSA/HRSA Primary and Behavioral Health Care Initiative (PBHCI) grant program, initiated in 2009 and financed by the 2010 Affordable Care Act to provide 4 years of funding to 93 sites across the country (24). Grants are being used to add primary care providers, care managers, and wellness coaches to create an integrated team to improve the health status of patients seen in behavioral health settings (Figure 3). The addition of care managers is a crucial component to this model where they function to track patients, monitor adherence to treatment, facilitate referrals to medical care, and alert the team to problems. The first study describing program characteristics and early implementation experiences has recently been published (25). While official clinical data from some of the early cohorts is not currently available, there are initial indications that measures such as blood pressure, HbA1c, and cholesterol are showing improvement (26). The addition of primary care staff into the community mental health center (CMHC) environment provides an opportunity for partnership and collaboration with the psychiatric providers, who have struggled with adequately screening and referring patients to other medical services for many years (27).

A second model being tested is the State Option to Provide Health Homes for Medicaid Beneficiaries with Chronic Illnesses (also funded through the 2010 ACA) initiated in 2011. This allows states to amend their Medicaid plan (often called State Plan Amendments or SPAs) and to receive additional funding for services provided to promote health. Patients with behavioral health conditions are qualified to receive services, and for the first time CMHCs can be the location of the health home where the care is provided (28).

Unlike the PBHCI sites, the health homes do not have to provide primary care services directly and instead must provide six additional health home services including comprehensive care management, care coordination, comprehensive transitional care, individual and family support, community linkages, and use of health information technology to link services (29). Eight states have implemented this model to date with Missouri having the most experience one year into implementation. A unique feature of the Missouri model is the addition of a primary care provider “consultant” to assist the CMHC care team with using aggregate data to identify problems and fine-tune medical and behavioral health interventions. This provides a general medical partner to the behavioral health staff for improving total treatment.

No published data exists on the outcomes of SPAs but early information from the Missouri model, which includes a medical care component (30), is showing reductions in emergency room and hospital visits and fewer 30-day readmissions. From provision of onsite primary clinics to the establishment of behavioral health homes, models are being tested with the promise of both improved outcomes and cost containment for these patients who are often high utilizers of healthcare resources.

Merged Clinics With Fully Integrated Medical and behavioral health Care

A final and less common model combines treatment for the total health of a given population into one organizational structure, designed with a diverse team of behavioral health and medical staff working in a unified manner to provide the type of care needed at any visit. These clinics operate under one board of directors with the mission to provide comprehensive collaborative care. Funding limitations, dual licensure requirements, and difficulty with electronic medical records are obstacles that must be overcome in these models. Several of these clinics currently exist nationally and provide a rich environment for the practicing psychiatrist to appreciate fully integrated treatment. Cherokee Health Systems in Tennessee (www.cherokeehealth.com) and Axis Health System in Colorado (www.axishealthsystem.org) are two examples of this model.

Addressing the Needs of Inpatient High Utilizers of Healthcare Resources

Interventions for medically complex patients who often have a behavioral health diagnosis and are high utilizers of healthcare are being identified. The term “hot spotting,” introduced in the 2011 New Yorker article by Atul Gawande, M.D. (31), has become a popular expression for this cohort of patients and entails finding high-cost patients and providing additional services to meet their needs and bring down the costs associated with their care. Jeff Brenner, M.D., from the Camden Coalition, the physician profiled by Gawande, describes a further subgroup of this set of medically complex and expensive patients as the “extreme bucket”, whose difficult to treat behavioral health needs pose an opportunity for intervention for consultant psychiatrists in inpatient settings.

Taking a proactive rather than reactive approach to case finding in hospital settings is an identified shift necessary in the way care is provided (32). One study utilized proactive consult liaison services to identify medically complex patients on general medical units with comorbid behavioral health issues and demonstrated earlier intervention could reduce length of stays (33). Another study demonstrated the utility of delirium prevention interventions in reducing length of stay in intensive care units (34). Utilizing medical complexity tools during the initial nursing assessment helps identify and prioritize which patients need immediate referral and help match the intervention to the level of need (35). Lastly, the idea of a “Complexity Intervention Unit” in inpatient settings has been proposed as an ideal location to take care of this group of patients (32).

Questions and Controversy

The issue of liability in integrated settings can be considered along two lines. The first is whether or not a doctor-patient relationship was established. Legally, this is usually determined by whether or not there was direct evaluation of a patient (in person or by televideo) and subsequent documentation of findings. Indirect or “curbside” consultations do not establish a doctor-patient relationship so there is presumed minimal liability (36). An additional area of consultation in collaborative settings in primary care is between the consulting psychiatrist and the behavioral health provider. Liability in this situation can depend on the role of the psychiatrist, which can range from consultative to collaborative to supervisory (37). There is increased liability if you are the supervisor of the behavioral health provider and are ultimately responsible for the care they provide. Finally, the primary care provider should remain in charge of the overall care provided, deciding if they wish to implement the consultant psychiatrist’s recommendations and prescribe all medications suggested. Figure 4 shows the nature of the liability and methods to reduce risk.

Recommendations