Establishing the Research Agenda for Collaborative Care
This collection of three research papers represents the fruits of the AHRQ-funded Collaborative Care Research Network Research Development Conference in Denver in October 2009. At the meeting, participants took steps toward establishing a research agenda for collaborative care among primary care and mental health clinicians.
Collaborative care emphasizes the recognition and care of mental health problems in primary care settings and the effective collaboration of primary care and mental health clinicians. Collaborative care is regarded as an important function of the patient-centered medical home (PCMH), and one avenue toward achieving PCMH goals for health, patient experience, and affordability.
The three papers are:
- Establishing a Research Agenda for Collaborative Care (Miller, Kessler, Peek, and Kallenberg).
- A Framework for Collaborative Care Metrics (Kessler and Miller).
- A Collaborative Care Lexicon for Asking Practice and Research Development Questions (Peek).
Benjamin F. Miller, Psy.D., University of Colorado School of Medicine Department of Family Medicinea
Rodger Kessler, Ph.D., ABPP, University of Vermont College of Medicine
C.J. Peek, Ph.D., University of Minnesota School of Medicine
Gene A. Kallenberg, M.D., Department of Family and Preventive Medicine, University of California, San Diego
The ongoing crisis in health care continues to be driven by the twin concerns of cost and quality. Recent healthcare policy changes promote significant system reorganization (e.g., patient centered medical homes and accountable care organizations) aimed at increasing coordination and comprehensiveness of care as a way to both contain cost and increase quality. Improvements in the coordination between mental health and primary care offer a prominent example of an area of healthcare reorganization that can contribute to both better quality and lower costs. The phenomena and practice of mental health and primary care have been linked inextricably, and a body of research highlights the benefits of integrating mental health into primary care and addresses mental health and physical health simultaneously. However, despite significant positive outcome data on integration, most research on mental health in primary care has been disease specific, using targeted interventions not always indicative of standard clinical practice. Systematic reviews on integrating mental health and primary care have concluded that despite the benefits of integration or the benefits of increased attention to mental health problems in primary care, more research is needed to understand the effects of specific strategies, levels of integration, care processes, or financial models on outcomes. Taking into account these gaps in evidence along with what is already known, this paper reports a research agenda for mental health in primary care created at the Collaborative Care Research Network Research Development Conference in Denver. This manuscript will propose two sets of research questions for the field of integrated mental health and primary care.
It is well known that the majority of mental health-related complaints presented by patients are to their primary care providers1-4: consider that psychosocial issues are present in approximately 70% of primary care visits,5 and conditions such as depression are some of the most common seen in primary care.6 New models for integrating mental health care into primary care have been proposed, yet research on this integration has yet to catch up with clinical innovation, often leaving many practices wondering which model or strategy is best for them.7 This discussion deals with the component of patients' mental health issues that are routinely brought to the primary care setting—problems that are already considered part of the mission of primary care.8 While some patients with severe and persistent mental illnesses are seen in primary care, the majority of this population is followed in mental health clinics.9 These patients are very much in need of primary care services integrated into their mental health settings. Many of these patients die prematurely of their medical illnesses due to lack of medical intervention largely because their mental health issues make it difficult for them to seek and receive primary care in primary care clinics.10,11 When one examines the literature on mental health and primary care, the majority of the findings support specific interventions for targeted disease conditions. 7,12
The Agency for Healthcare Research and Quality (AHRQ) commissioned a report systematically reviewing the literature on integrating mental health and primary care in late 2008. This Minnesota Evidence-Based-Practice Center report on Integration of Mental Health/Substance Abuse and Primary Care synthesized the evidence and concluded that integration of mental health into primary care works, but it remains impossible to tease apart the model or level of integration from the added attention paid to a specific disease (e.g., depression).7 The report went on to challenge the healthcare field to begin to study which elements of integrating mental health into primary care (including financial models) lead to desired outcomes. While the report offered suggestions for areas of future research on integration, it did not set a research agenda or identify how a research agenda could fill the gap between what we know and what we need to know.
For example, the majority of studies used in the AHRQ systematic review focused on depression, using a specific model. Therefore, empirical questions around what other models, diseases, and strategies could benefit from integration could not be created and addressed. The twin challenge of this inquiry then is to create a research agenda that combines studying the effect of integration or collaboration as a system attribute in action across specific treatments or conditions and studying the effect of specific treatments on specific patient groups or specific conditions. In other words, a research agenda for mental health in primary care needs to ask questions about the generalized effects on clinical and financial outcomes of different kinds of collaboration and collaborative care teams and the effects of specific collaborative care treatments on outcomes for different patient groups. This makes a research agenda conceptually challenging, but this is the reality and the gaps that require study.
For the purposes of this paper, we are using the global definition of collaborative primary and mental health care proposed by Miller, Mendenhall, and Malik13, which identifies collaborative care as the larger construct addressing the integration of mental health providers and services into primary care. The reader will note that clarity in language will become a prominent theme in this paper both in terms of content and research agenda.
The Collaborative Care Research Network Research Development Conference in Denver, Colorado in October 2009 brought together a group of national leaders in primary care, mental health, practice-based research networks (PBRN), research, and healthcare systems. The meeting was audiotaped and transcribed to ensure accurate reporting of information.
The White Papers
One important aspect of this conference was the series of white papers, solicited and written prior to the conference, covering areas germane to collaborative care and research. To prepare participants for setting a research agenda, several areas needed to be reviewed and addressed. The paper titles and authors are listed in Table 1. While each paper represented an original and significant contribution to the conference, the collaborative care lexicon paper (the basis for the Peek paper in this series) will be highlighted here because, without clarification and definition of a common language, creating a research agenda would be difficult. The financing paper is not included in this current collection because it pertains more to billing and payment issues than a research agenda. The paper has been used by Ms. Reynolds in her work with the National Council for Community Behavioral Healthcare.
Table 1: White paper topics developed for this conference
|A Framework for Collaborative Care Metrics: Kessler and Miller (included in this series)|
|Financing of Collaborative Healthcare Services: The Who, What, When and Where of Billing Issues: Reynolds and Miller|
|Toward a Conceptual System for the Field of Collaborative Care: A Starter Lexicon for the Collaborative Care Research Network: Peek (precursor to Peek article in this series)|
Toward a Collaborative Care Lexicon
The Collaborative Care Research Network (CCRN), a sub-network of the American Academy of Family Physicians National Research Network and a PBRN, was created to pose and answer collaborative care research questions in a way that can be consistently understood in practices across the Nation. By enrolling primary care practices that had integrated mental health, the CCRN would be able to answer research questions germane to mental health and primary care integration. The conference leadership also led the CCRN and believed that the network could be the platform to carry out the research agenda. But the program committee's initial experiences framing research questions revealed confusion about the terms in common use, especially regarding the components of collaborative care that are the subject of research. The program committee also realized that funding agencies and policymakers need consistently articulated concepts for this field. Such conceptual clarity, or pre-empirical work, has preceded the empirical success we associate with mature scientific fields. The lexicon paper established a common vocabulary to ensure that all discussion participants were using the same language to talk about mental health in primary care. This lexicon or conceptual system for this new scientific field was created as a product for AHRQ and was used to formulate research questions.
A method for creating a lexicon that would have standing in the field would have to be consensual and analytic (a disciplined transparent process—not a political campaign); involve actual implementers and users who know the field well; focus on what functionalities look like in practice (not just on principles or values); and be amenable to gathering around it an expanding circle of "owners" and contributors (not just an elite group coming up with a declaration). The lexicon white paper employed methods for defining complex subject matters that meet these requirements called "paradigm case formulation" and "parametric analysis."14 A paradigm case formulation is a vehicle for creating a definition that maps both similarities and differences—a consistent core and acceptable variations. A parametric analysis is a vehicle for being explicit about how one instance of collaborative care might differ from another instance.
It was clear early on in the planning of the conference that the lexicon paper would play a significant role in framing the discussion and creating a research agenda. Similarly, the group recognized the need to use a consistent lexicon when describing collaborative care. Collaborative care is used in this document to describe an ongoing relationship among clinicians (i.e., mental health/behavioral health/substance abuse and primary care) over time.15 This is not a fixed model, but rather a larger construct consisting of various components which, when combined, create models of collaborative care practice.16,17
Collaborative Care Metrics and Evaluation Framework
To research and evaluate mental health in primary care, we needed a unifying construct to serve as a framework for measuring the success of collaborative care. The Kessler and Miller metrics white paper included in this series defines the conceptual relationships among structure, process, and outcome, building on the work of Miller, Mendenhall, and Malik.13 Peek18 has discussed a "Three Worlds" model of collaborative care, which says that all actions in collaborative care have clinical, operational, and financial dimensions and that these must be made to work together. For example, a new clinical model will no doubt require operational changes and often business model changes. This approach is consistent with other contemporary concepts and measurements within healthcare, such as the Triple Aim described by Berwick and colleagues.19
The Method for Producing the Research Agenda—the Conference
The work flowed over the 1.5-day conference from:
- Orientation to the perspectives and desired conference outcomes brought by participants.
- Introduction of the lexicon as a vocabulary for use during the conference.
- Orientation to the research picture in the field so far—from clinical and financial perspectives.
- Exploring and synthesizing collaborative care research agendas dealing with clinical, operational, and financial aspects of collaborative care
- Examination of challenges for comparative effectiveness research in collaborative care—and how that relates to policy issues and diffusion of research.
The time was divided between large group conversations and small group tasks that were brought back to the large group. A facilitator was employed to keep the group on task and on time. Scheduled time for informal conversations that helped generate ideas and synthesis also occurred at breaks and meals.
The following results propose a new research agenda for collaborative care building on what is known from the white papers and reports such as the AHRQ systematic review. Current knowledge is limited regarding the extent and characteristics of collaborative care practice. Further, the lack of a consistent evaluation model with specified outcomes has impeded large scale translational opportunities. The following research recommendations concern primarily 1) the knowledge that ultimately needs to be gained (what kinds of collaborative care strategies work best for whom—with triple aim outcomes in mind); and as a means to that end, 2) the need to acquire a level of robust organized knowledge of what is presently occurring in practices in the field that claim to be practicing collaborative care. This latter point is especially salient in creating a practice-based research network agenda for the CCRN.
Therefore, two sets of research questions need empirical investigation in collaborative care. Part A questions are descriptive. What is going on in the field by way of collaborative care? What models and functions are being employed in the field? Part B questions are evaluative. What collaborative care functions or organizational supports being used lead to improved outcomes and for whom? This research agenda has been created to find out what makes a positive difference (or not) among the many variants of collaborative care seen in practice. The agenda will be an ongoing mixture of mapping existing practice across the country and discovering which forms make a positive difference to whom.
While having Part A questions fully answered before Part B questions could help describe collaborative care settings and outcomes of any studies of care delivered in those settings, it is not required. To help tease out the effects of collaborative care as a strategy from the effects of greater effort and attention being paid to the assessment and treatment of mental health problems, a better understanding is needed of what constitutes a collaborative care practice. Also, the overall collaborative care characteristics may play critical outcome roles in all phases of mental health problem management from identification/screening to treatment to improvements in patients' mental health status to effects on other aspects of patients' health care. It will be necessary to place a practice or practices participating in any outcomes studies on the continuum of collaborative care as part of the methods section of any such study.
- Descriptive Questions: What is currently going on in collaborative care—the extent to which variations in practice models, target populations, and other dimensions are using the lexicon as a nosology or definitional framework. This amounts to creating a systematically articulated picture of collaborative care practice as actually done—specifying the similarities and differences in language people can consistently understand.
- Evaluative Questions: With "A" Questions in view, what collaborative care arrangements work best for whom—with triple-aim outcomes in mind.
The research questions below are divided into parts A and B. It is recommended that part A questions be undertaken in the first phase of work and then move as quickly as possible to part B questions that can actually be studied effectively in existing practices.
The Part A research questions look at how practices vary along the dimensions of the collaborative care parameters created by Peek in his lexicon white paper in this series; answers to these questions will demonstrate how various collaborative care practices may differ. The following research agenda will require referencing Peek's parameters (Go to Figure 5, Parameters of Collaborative Care Practice, in "A Collaborative Care Lexicon for Asking Practice and Research Development Questions in this collection.") In brief, a parameter is a key characteristic of collaborative care that contributes to the conceptual whole. Parameters might describe the types of clinicians or clinical functions included on the team, the types and frequency of interactions between primary care and mental health clinicians, the type of clinical setting involved, characteristics of the patient population, and the like.
Populations Seen in Collaborative Care (Parameter 3)
- Who is being seen in collaborative care practice (genuine cases of collaborative care as delimited by the paradigm case). Proportion, number, and other descriptors of patients with mental health presentations in each range category of mental health need (Go to Figure 5 in "A Collaborative Care Lexicon for Asking Practice and Research Development Questions" in this collection.)
- Medical presentations that need behavioral /mental health treatment as part of the plan
- Medical or mental health presentations that require improvement in health behaviors
- Comorbid or multimorbid medical and psychological/mental health presentations
- Social conditions affecting physical or mental health--social determinants of health
- Mental health and substance abuse conditions within the scope of primary care management
- Severe mental health and substance abuse conditions beyond the scope of primary care management
- In what care settings, venues, or sectors are these patients receiving collaborative care?
- Primary medical care
- Specialty medical care
- Specialty mental health care (where primary care providers are part of mental health teams)
- In what stage of life are the patients who are receiving collaborative care?
- Adults/young adults—nongeriatric
- Advanced illness/end of life
Characteristics of the Practices Seeing These Patients
- How are these patients identified for collaborative care (Parameter 4)?
- Targeted as members of specific populations, such as a disease or risk group vs. non-targeted, such as any patient deemed by anyone to need collaborative care—"all comers."
- Identified by patients or clinicians vs. system indicators or systematic screening
- What is the distribution of practices along a continuum of integration/collaboration?(Parameter 2)
- Coordinated—basic collaboration at a distance
- Co-located—basic collaboration on-site
- Integrated—in a partially or fully integrated system
- Do practices move along this continuum of integration or level of collaboration? In what directions and on what timeline?
- What is the distribution of clinical functional capabilities that are represented on the team or that can be quickly called in for care of patients? (Parameter 1) A more simplistic approach is to list the distribution of professional degrees or disciplines among clinicians working as teams or collaborating in integrated models.
- What is the range of patient perspective on the purpose, course, and usefulness of collaborative care treatment (and the degree to which they recognize it as such)?
- What is the distribution of practices along a continuum of organization development or maturity?
- From being starter pilots to being larger and more visible projects to being mainstream implementations (Parameter 5)
- Level of office practice consistency/reliability for collaborative care—from informal/nonstandard to partially routinized to standardized processes (Parameter 7)
- Business model type—e.g., from fee-for-service to addition of bundled payments to models of total cost/revenue (Parameter 8)
- Ability to collect and use practice-based data—from little or none to mature data collection and use (Parameter 9)
- How representative are Collaborative Care Research Network (CCRN) practices of collaborative care practices outside the network?
The Part B research questions look at which collaborative care arrangements work best for whom, considering the triple-aim outcomes. Part B questions are asked within the specified outcomes and evaluation framework. Therefore, research questions about "outcomes" can and should include:
- Metrics for access, identification, and treatment
- Metrics for clinical experience, patient experience, financial performance, and provider experience
These results are intended to help build a national database assuring representation from a broad range of practice conditions so that underrepresented practice and patient types are included in analyses.
Populations Seen in Collaborative Care
- Is improved patient access to mental health in primary care associated with:
- Frequency of patient self-identification of mental health problems?
- Overall patient volume (% of panel)?
- The frequency of appropriate patients identified by the practice?
- The frequency of patients referred to outpatient mental health outside the primary care team?
- Treatment initiation, completion, and attendance rates—including successful referrals?
- How do collaborative care interventions affect the outcomes of selected comorbid conditions or disease clusters, e.g., diabetes, depression, and coronary artery disease?
- Are patient experiences better in collaborative care practices?
- Are clinical outcomes better for care of conditions or situations in collaborative practices than in usual care?
Performance Characteristics of Practices Seeing Patients in Collaborative Care
- What functional components of collaborative care have the greatest effect on outcomes? That is, what makes the difference? A vocabulary for pointing to those components and functions appears in Peek's collaborative care parameters.
- Does progression along a continuum from less to more collaboration or integration produce better outcomes? For which patients or patient groups? How does the level of communication between mental health providers and primary care providers impact clinical care? (Parameter 2)
- Does choice of target population or how patients are identified for collaborative care affect outcomes—do some populations benefit more than others? (Parameters 3 and 4)
- Are there differential effects of care management models, team composition, involvement of various disciplines, and/or roles and responsibilities on clinical, operational, and financial outcomes? (Parameters 1,2,4)
- Do practices with higher levels of organizational development, reliability, consistency, and ability to collect and use practice-based data get better outcomes? Do they yield clearer or more powerful data for research? (Parameters 5,7,8,9)
- What role does the delivery setting (Parameter 3-A—primary medical setting, specialty medical setting, or specialty mental health setting) play on outcomes for patients with severe mental illness in a collaborative care model?
- Can we profile practices according to Peek's parameters at a more detailed level taking into account model specific information and operational components so that results from these profiles will differentiate kinds of practices and provide the initial basis for a scale of collaboration, likely effectiveness, or practice development level? Or simply: Can we show that placement on scales of collaboration and/or practice Development level is positively associated with clinical outcomes?
Financial Research Questions
A method for collecting cost and cost outcome data needs to be developed and made part of regular data collection for practices engaged in collaborative care. General recommendations:
- Target the research on financial and other outcomes that matter for the stakeholders—patients, providers, States, schools, payers, employers, Medicaid population, Medicare population.
- Employ broad categories of collaborative care outcomes meaningful to that range of stakeholders, such as access outcomes, identification outcomes, and treatment outcomes.
- What does it cost to establish and maintain a collaborative care practice, including the relative costs of key elements (i.e., mutual recordkeeping, time for behavioral health and physician consultation) at both individual patient and practice level?
- What are the startup costs of a practice moving from usual care to a collaborative care model? This includes the incremental costs of changing primary care team composition and function from usual care (primary care physician and nurse team) to collaborative care by adding care manager, onsite behavioral health provider, or other mental health providers and the operational, infrastructure, or IT support that these entail.
- What are the ongoing costs to a practice for maintaining these collaborative care functions once established and start-up costs accounted for?
- Which collaborative care practice models or characteristics achieve the desired financial outcomes for practices, plans, purchasers, employers, and patients?
- Are different financial models of support for collaborative care associated with differences in clinical and operational outcomes? Are different financial models associated with achieving those results sustainably over time? Are some business models better than others?
- Does level of integration/collaboration have a financial impact—initial cost, long-term cost, total cost of care?
- What financial metrics and financial outcomes for collaborative care are most convincing to stakeholders: e.g., reductions (or improved appropriateness) in utilization, effect on per patient per month, lost employee productivity, patient out-of-pocket expense, and other financial variables?
If the field of collaborative care is to advance, it needs a more substantial evidence base on which specific components or functions in collaborative care practices lead to desired outcomes for patients, providers, and other stakeholders. The field needs a research agenda, but first needs a defined lexicon. The evidence is especially important as part of the imperative to improve quality while bending the cost curve in primary care—such as through the patient centered medical home. Identifying such gaps in the evidence, the Agency for Healthcare Research and Quality supported a small conference on creating a research agenda for collaborative care as the first step in addressing this need.
As described in this paper, a two-part approach is needed for developing this collaborative care evidence base. At the very beginning, robust descriptive data will need to be collected to take a snapshot of the field and who is doing what and where—a descriptive study of the naturally occurring experiments and developments in collaborative care in all its various forms. Thus far, there have been no attempts to do this nationally. While this descriptive work is underway, a second set of studies can begin to evaluate the cost and effectiveness of various models of collaborative care delivery taking into account the clinical, operational, and financial perspectives of health care system design and the care, experience, and affordability perspectives of patients. To articulate and answer these collaborative care research questions, a basic conceptual system for this important subfield must be developed and in place in this project—something that enables researchers, clinicians, and policymakers to talk to each other using a common vocabulary and an organized way of specifying the required components of collaborative care. As the field expands the evidentiary support for collaborative care, clinicians, administrators, insurers, and policymakers will have more information to inform decisions on healthcare.
1 Green LA, Fryer GE, Jr, Yawn BP, Lanier D, Dovey SM. The ecology of medical care revisited. N Engl J Med Jun 28 2001;344(26):2021-2025.
2 Hogan MF. The President's New Freedom Commission: recommendations to transform mental health care in America. Psychiatr Serv Nov 2003;54(11):1467-1474.
3 Institute of Medicine. Committee on Crossing the Quality Chasm: Adaptation to Mental Health and Addictive Disorders. National Academy of Sciences. 2006.
4 Kessler RC, McConagle KA, Zhao S, et al. Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States: Results from the National Comorbidity Study. Arch Gen Psychiatry 1994;51:8-19.
5 Fries J, Koop C, Beadle C. Reducing health care costs by reducing the need and demand for medical services. N Engl J Med 1993;329:321-325.
6 Sartorius N, Ustun TB, Costa e Silva JA, et al. An international study of psychological problems in primary care: Preliminary report from the WHO Collaborative Project on Psychological Problems in General Healthcare. Arch Gen Psychiatry 1993;50:818-824.
7 Butler M, Kane RL, McAlpin D, et al. Integration of Mental Health/Substance Abuse and Primary Care, No. 173 (Prepared by the Minnesota Evidence-based Practice Center under Contract No. 290-02-0009.) AHRQ Publication No. 09-E003. Rockville, MD: Agency for Healthcare Research and Quality; October 2008.
8 Starfield B, Shi L, Mackino J. Contributions of primary care to health systems and health. The Milbank Quarterly. 2005;83:457-502.
9 Daumit GL, Pratt LA, Crum RM, Powe NR, Ford DE. Characteristics of primary care visits for individuals with severe mental illness in a national sample. General Hospital Psychiatry. 2002;24(6):391-395.
10 Piatt EE, Munetz MR, Ritter C. An examination of premature mortality among decedents with serious mental illness and those in the general population. Psychiatr Serv Jul 2010;61(7):663-668.
11 Parks J, Svendsen D, Singer P, Foti ME, Mauer B. Morbidity and mortality in people with serious mental illness. Alexandria, VA: National Association of State Mental Health Program Directors;2006.
12 Blount A. Integrated primary care: organizing the evidence. Families, Systems, & Health Vol 21(2) Sum 2003, 121-133. 2003.
13 Miller BF, Mendenhall TJ, Malik AD. Integrated primary care: An inclusive three-world view through process metrics and empirical discrimination. Journal of Clinical Psychology in Medical Settings. 2009;16:21-30.
14 Ossorio PG. The Behavior of Persons, Vol V. Ann Arbor, MI: Descriptive Psychology Press; 2006.
15 Doherty WJ, McDaniel SH, Baird MA. Five levels of primary care/behavioral healthcare collaboration. Behavioral Healthcare Tomorrow 1996:25-28.
16 Craven M, Bland R. Better practices in collaborative mental health care: An analysis of the evidence base. Canadian Journal of Psychiatry 2006;51 (suppl 1, May).
17 Peek CJ. Collaborative care: aids to navigation. Presented at the AHRQ-sponsored Collaborative Care Research Network Research Development Conference, Denver, CO; October 2009.
18 Peek CJ. Planning care in the clinical, operational, and financial worlds. In: Kessler R, Stafford D, eds. Collaborative Medicine Case Studies: Evidence in Practice. New York: Springer; 2008.
19 Berwick DM, Nolan TW, Whittington J. The triple aim: Care, health, and cost. Health Aff May 1, 2008;27(3):759-769.
a Corresponding author: Benjamin Miller, PsyD. Benjamin.Miller@ucdenver.edu