Large-Scale Dissemination of Collaborative Care and Implications for Psychiatry

Consider the overwhelming task of providing mental health care for 44 million adults in the United States. Sadly, there are shortages of mental health providers throughout the country, and access is particularly limited in rural areas. Access to child and adolescent providers is even more problematic (1). For individuals who receive treatment, there is a wide gap between treatments that are recommended and those that are received, especially for people with co-occurring mental and substance use disorders (2). A new delivery model that maximizes the impact of all mental health providers is needed. This column reviews the benefits of collaborative care and the challenges of widespread implementation of this model.

Effectiveness of Collaborative Care

Cost-effective, evidence-based treatments for depression, bipolar disorder, anxiety disorders, and several other mental illnesses are available. In a meta-analysis of results from 79 depression studies that compared collaborative care with usual care, Archer and colleagues (3) found an effect size favoring collaborative care of .34 for short-term care, .28 for medium-term care, and .35 for long-term care. Results for anxiety were similar, with effect sizes of .44, .33, and .20, respectively. These findings are similar to the effect size of .37 for U.S. Food and Drug Administration registration trials that compared antidepressants with placebo. Would we ever fail to use a drug that was as effective as collaborative care?

The Science of Dissemination

With such strong evidence of effectiveness, why isn’t collaborative care viewed as usual practice? The movie Field of Dreams was based on the idea that if you build it, they will come. Health care’s version is “Design it, and they will implement it.” Why should it take an average of 15 years for new research findings to move from publication to widespread dissemination? Collaborative care is a typical example: the first positive effectiveness trials were published in the early 1990s. Many innovations are examined in efficacy and effectiveness studies, but dissemination trials are rarely conducted, which deprives many people of effective treatments. Fitzgerald and colleagues (4) examined why some innovations are rapidly disseminated and others are not. What makes clinical professionals decide to adopt an innovation and use it in their clinical practice? They found that the innovation must be effective and applicable to a large population and cost-neutral (or must save money), must increase patient satisfaction, and must not be complicated to implement. Disseminating a new medication is relatively simple compared with disseminating collaborative care, which requires a fundamental change in the system of care. The current fee-for-service reimbursement model creates a substantial disincentive, because there is minimal funding for care coordinators and for collaboration between clinicians.

Fortunately, the rapidly growing science of quality improvement helps accelerate the dissemination process. The learning collaborative is a particularly effective method for designing and adapting a new model of care delivery to diverse settings. Introduced over a decade ago, and promoted in the Institute for Healthcare Improvement’s Breakthrough Series (5), learning collaboratives bring together content experts, change management experts, and motivated clinicians from multiple locations. This intensive process has been successful in overcoming inertia and barriers to dissemination. The key to this process is to start small with a group of early adopters willing to tolerate initial challenges. These early adopters use “plan-do-study-act” cycles to clarify adaptations to local system barriers that increase the chances that a new idea will work in a given environment (6) and to assist in navigating a common challenge described by Kilbourne and colleagues (7) as the “balance between adequate fidelity to the intervention and accommodating differences across organizations.” For complex system adaptations where there is some debate on the most workable approach in a given type of setting (for example, primary care) or when the approach is clear but may need to be adapted to a wide variety of environments (for example, a variety of primary care practice settings), a learning collaborative may prove invaluable. When systems are already in place but not ideal, other methodologies borrowed from industry may increase the quality and output of a given process (Six Sigma) or reduce waste (lean manufacturing). Although more research is needed on matching the appropriate methodology to the change envisioned (8), change leaders have a number of options for moving their practice toward effective and efficient collaborative models. Frequently, the main barrier is financial, and a study by Unützer and colleagues (9) suggested that linking key quality indicators to incentive payments can substantially improve fidelity and patient outcomes.

Successful Large-Scale Dissemination

Below we discuss six programs that involve adaptations of evidence-based models in real-world settings for ongoing patient care. Each includes the use of a registry, a stepped care approach, and a team-based approach that involves a change in mental health care delivery based on the collaboration of a team of mental health providers. The first program is COMPASS (Care of Mental, Physical, and Substance-use Syndromes), funded by the Center for Medicare and Medicaid Services to disseminate a modified version of the TEAMcare model (10) in 15 medical groups across the country. COMPASS includes patients who have depression with comorbid diabetes or cardiovascular disease and who are cared for in a primary care clinic. A key feature is the systematic case review (SCR), a process in which a care coordinator meets each week with a supervising psychiatrist and primary care provider to review the coordinator’s panel of patients and make recommendations to the patient’s primary care provider. The SCR is also a major component of two interventions described below—the MHIP (Mental Health Integration Program) and DIAMOND (Depression Improvement Across Minnesota Offering a New Direction).

The second program is MHIP, which has been implemented in more than 200 community mental health centers and community health centers in Washington State (11). MHIP serves a diverse safety-net population and is designed to integrate general medical and mental health services. MHIP uses a collaborative approach, including a primary care provider, care coordinator, and consulting psychiatrist assigned to each primary care–based team. The coordinator also acts as a liaison between the mental health center, community health center, and other groups, such as social services and vocational rehabilitation.

The third program is DIAMOND, which was coordinated by the Institute for Clinical Systems Improvement and is based on the IMPACT model (12). It has been implemented in more than 80 primary care clinics in Minnesota and in clinics in Michigan and Hawaii. Patients with major depression are referred by their primary care provider to a care coordinator, usually a nurse, who completes an intake evaluation, enters the information in a longitudinal registry, and reviews the patient in an SCR with a supervising psychiatrist. Recommendations are then made to the patient’s primary care provider.

The fourth program is RESPECT-Mil (ReEngineering Primary Care Treatment of PTSD and Depression in the Military) (13). This collaborative care program is adapted from the three-component model for reengineering systems, with the addition of routine primary care screening for posttraumatic stress disorder (PTSD) and depression, and has been implemented in 88 Army primary care clinics.

The fifth program is IAPT (Improving Access to Psychological Therapies) (14). This massive U.K. program is treating more than 590,000 patients a year and has involved training almost 4,000 therapists. It is designed for adults who have anxiety disorder or depression and uses a stepped care model with a low-intensity, guided self-help intervention and a high-intensity psychological therapy intervention.

The sixth is PC-MHI (Primary Care Mental Health Integration initiative) (15), which has been implemented in Veterans Health Administration (VHA) facilities. It includes three models: colocation of behavioral health providers into primary care; TIDES (Translating Initiatives in Depression into Effective Solutions); and Behavioral Health Lab, which is very similar to the IMPACT model, with the addition of a computerized assessment program and more resources for psychotherapy. Almost all VHA primary care facilities have a PC-MHI program, and all focus on helping patients who have depression, anxiety disorders, alcohol misuse and abuse, and PTSD.

Implications

The Affordable Care Act and patient-centered medical homes are rapidly changing the practice of psychiatry. In March 2013, James Scully, M.D. (16), then CEO and medical director of the American Psychiatric Association, said “The future of psychiatry in the new era of health care reform will involve more team-based integrative care than ever before. That means a shift away from the fee-for-service, volume-based model of care to which psychiatrists are accustomed. We are going to have to change the way we do business in order to survive.” The shift to a population perspective with total-cost-of-care reimbursement will provide incentives for groups to accelerate dissemination of collaborative care, but it is critical that these new systems retain the key elements of the collaborative model while allowing local adaptation of other elements. Expansion of mental health care benefits will substantially increase the demand for services. Implementation of collaborative care models will allow us to increase our reach to many more adults and children who have mental illness and to simultaneously address psychiatric and nonpsychiatric medical illnesses.

Next Steps

The optimal way to provide collaborative care is still evolving, with many unanswered questions. Should collaborative care focus on a single disease or on multiple diseases? Should the same care coordinator address mental illnesses and chronic general medical illnesses? Do we need “specialty care coordinators” with extra skills? Is in-person coordination necessary, or can most or all care coordination be done via telephone? What should a practice expect as a return on its investment and when? Is evidence of improved satisfaction and quality of life enough? Are improved clinical outcomes sufficient if costs do not rise? How should specialists be paid for this work? How will consumers learn about these programs and ask for them to be included in benefits? Social marketing may be needed and can be successful (17).

Conclusions

Basic scientific advances have not consistently led to improved health. Learning collaboratives and other scientifically supported methods of dissemination are often needed for rapid dissemination of effective but complex innovations. Large-scale programs to disseminate mental health collaborative care have been successful, and health care reform has the potential to accelerate the dissemination process. More mental health providers will be involved in team-based care that is closely woven into primary care practice. It is important that psychiatric residency programs and other mental health training programs prepare students to practice effectively in these new models of care. These rapid changes will be stressful, but ultimately they will create new opportunities for psychiatrists and all mental health providers and provide effective care for more people with mental illnesses.