A Framework for Collaborative Care Metrics

A National Agenda for Research in Collaborative Care

Rodger Kessler, Ph.D. ABPP, University of Vermont College of Medicinea
Benjamin F. Miller, Psy.D., University of Colorado School of Medicine Department of Family Medicine


The difficulties in identifying and accessing services for mental health concerns and symptoms are well documented. This is particularly challenging in primary medical care where many persons present mental health or stress-related symptoms that interact with their overall health picture, but are not ready to accept referrals to specialty mental health and behavioral health care. Systematic reviews on integrating mental health with primary care (collaborative care) have concluded that despite the benefits of integration (or of increased attention to mental health problems) in primary care, more research is needed to understand the effects on outcomes of different specific approaches to clinical integration, care processes, or supporting financial models.

These discussions about collaborative care have been limited by the lack of a shared language and conceptual framework that define its core elements or allow us to identify which elements combined in which ways lead to positive outcomes. A consensus-based lexicon of collaborative care was developed for (and by) participants in the AHRQ-funded Collaborative Care Research Network Research Development Conference in Denver, Colorado in October 2009. (go to "A Collaborative Care Lexicon for Asking Practice and Research Development Questions " in this collection.) From that work emerged five defining clauses (a paradigm case) necessary for inclusion in a collaborative care practice and thirteen elements to be measured (parameters) that specify the acceptable differences between instances of collaborative care practice. This now allows us to calculate standardized metrics to benchmark and evaluate the process and outcomes of collaborative care delivery. In addition, such standardized metrics provide the ability to generate profiles and patterns of practice that can be used to evaluate the effectiveness of different aspects of collaborative care. This paper presents the rationale, framework, and examples of initial metrics derived from the paradigm case and parameters of collaborative care. Where needed, project-specific metrics can be developed using this framework.

Return to Contents 

Access to Behavioral Care for Medical Patients: Insufficient

Quality mental health services are identified as the most difficult subspecialty for primary care physicians to access.1 A survey of 6,600 primary care physicians reports that two thirds cannot gain access to outpatient mental health services for their patients. Difficulty in accessing mental health care is at least twice as high as for any other medical subspecialty.1 A lengthy literature suggests that the detection of mental health issuesand referral to appropriate treatment resources, if available to primary care, have only marginally improved over the last fifteen years.2,3 Even if access and identification are improved, specific treatment(appropriate evidence-supported interventions responding to different patient problems and needs) continues to be generally unavailable to primary care patients and their physicians.4 Taken together, lack of access, lack of identification of care need and unavailability of evidence supported treatments represent troubling structural and process limitations to achieving overall quality of care. Taking into account these gaps in evidence, a research agenda for mental health in primary care was created at the 2009 research development conference that is reported in a companion paper in this collection. (go to Establishing the Research Agenda for Collaborative Care).

In response, there is a growing trend and a variety of efforts to collocate and integrate mental health, substance abuse, and health behavior services into primary care practice. However, as of yet, support for such efforts has not been widespread throughout the larger healthcare policy, planning, or delivery systems. One potential reason for this is the lack of a shared lexicon describing the essential functional dimensions of collaborative care and the differences in how these are carried out between one practice and another. Without such a lexicon, policy and practice development efforts and discussions are much more difficult because it is not clear to all participants in the discussion exactly what is being talked about. Without such a lexicon, there cannot be a standardized evaluation model and set of metrics to measure the effectiveness of the plethora of collaborative mental health initiatives that have emerged and will continue to do so. 

Return to Contents 

Uncertainty About Characteristics of Ideal Care

A 2008 AHRQ evidence report (Butler et al) suggests that while collaborative care appears effective, we cannot presently identify the elements that contribute to that effectiveness. Also, there is no way to discriminate between the relative impact of interventions that have demonstrated effectiveness and the impact of  mere increased organized attention to patients' mental health problems.5 Further, the amount of collaboration in a practice necessary to enhance effectiveness has not yet been demonstrated. The present range of mental health interventions in primary care includes enhanced referral to specialty mental health care; collocated specialty care practice located in primary care; care managers, such as nurses or medical assistants under psychiatric supervision with a focus on a single psychiatric diagnosis (usually depression); and specially trained mental health clinicians who provide services in collaboration with primary care providers. These interventions are provided by a variety of mental health and non-mental health professionals with varied training and background conducting a broad range of clinical activities with varied degrees of organizational engagement and working with a broad range of financial models. Patients served are sometimes homogeneous and sometimes heterogeneous in diagnosis, gender, race, and ethnicity. Organizational characteristics and locations of settings vary, as does financing of the initiatives. Across this breadth, there has been no consensus in the field about how to define or evaluate the effectiveness of each of these models or components.

Return to Contents 

Discussing Collaborative Care With a Common Consensually Derived Language

 In  "A Collaborative Care Lexicon for Asking Practice and Research Development Questions," included in this volume, Peek defines the parameters of the paradigm case of collaborative care "lexicon" and the methodology used to generate it. This lexicon articulates in detail five defining functions necessary for a practice to qualify as a collaborative care practice (a paradigm case), and nine dimensions that provide a vocabulary for how one genuine collaborative care practice might differ from another on key dimensions (parameters). This lexicon amounts to a vocabulary with which to ask research questions and specify metrics. Some have referred to this as an "operational definition" that maps both similarities (essential functions) and differences (how one practice might legitimately differ from another one). In this paper we identify metrics corresponding to the functions identified as core parameters of the lexicon. We suggest that these metrics can serve as an evaluative framework for collaborative care, allowing us to benchmark collaborative care practice and conduct comparative effectiveness research.

Return to Contents 

The Need for Definition and Measurement

The definition offered by Peek's collaborative care lexicon in and of itself advances the discussion surrounding the inclusion of mental health as part of health care reform's emergent models of primary care.  Crucial to the argument for inclusion is the ability to 1) consistently describe collaborative care functions in their different forms, 2) measure and then evaluate the degree of presence or absence of these functions, and 3) look at practices to discover which combinations of these functions are associated with desired outcomes (evaluation of effectiveness) of collaborative care. Evaluation of effectiveness requires both dimensions to evaluate and specific measures/metrics within each collaborative care function. Such parameters and metrics must parallel those being used to evaluate the rest of contemporary healthcare. This implies that collaborative care (i.e., mental health care integrated into primary care) should be held to accountability, monitoring and structure, process, and outcome standards that Donabedian has suggested as the key dimensions to enable the evaluation of overall quality of healthcare.6

Return to Contents 

Parameters of Measurement

Peek's paradigm case provides a framework for defining collaborative care and allows us to observe and evaluate relationships between structure, process, and outcome. The paradigm case contains not only clinical but also operational and financial functions because all "three worlds" working simultaneously in harmony are required for consistent and sustainable success. This position is consistent with the contemporary conceptual and measurement dimensions of the Triple Aim (care, health, and cost) identified by Berwick and colleagues.7

The Triple Aim states that improvement in health care requires simultaneous attention to three aims: 1) Improve the health of the population; 2) enhance improving the patient experience of care, and 3)  reduce, or at least control, the per capita cost of care.7 There are considerable similarities between Peek's and Berwick's ideas. Attention to the clinical world implies improvements in health as well as patient experience; attention to the operational world reminds us of the operational and organizational aspects inherent in care; and attention to the financial world reminds us that practices require sustainable business models and that affordability is an aim for healthcare which is the direct concern of both schemas. Thus, both the "Triple Aim" and the "Three Worlds" paradigms support care driven by a team with a shared mission, using improved clinical systems to deliver improved care to a population of patients supported by operational and financial systems. Such care is continuously evaluated through improvement processes and effectiveness measurement.

Return to Contents 

Collaborative Care Measures and Metrics

Eden and Simone8 suggest a structure for presentation of metrics to evaluate health care. They suggest that in addition to identifying the content and methodology for constructing the metric, one can identify sources balanced among research, practice, and the utility of the metric for patients, primary care practices, and larger systems. So the target is a measure, clearly stated, with strong sources of support and with utility to multiple stakeholders.

Peek's paradigm case and parameters of collaborative care provide a framework for examples of metrics derived from those parameters and values. We do not suggest that they are exclusive of other metrics particularly suited to particular quality improvement or research projects and questions. We hope that this initial set will generate thoughts and ideas that will augment this effort. Again, the purpose of collecting a set of metrics for collaborative care practices is not to establish a preordained hierarchy. Rather, we will generate practice patterns to enable comparison of these patterns of metric performance with other performance and outcome variables so that we may understand any potential associations among them. Table 1 presents the elements to be measured, the metric to be calculated, and the source for the metrics.

It is likely that, initially, there will be few fully realized examples of the paradigm case of collaborative care. So it should be in a developing field. The intent of this paper is to advance the effort, advance the work in practices, and provide the opportunity for contrasts. These contrasts will help researchers evaluate the effectiveness of the field and test models and elements. They are not seen as the right metrics nor are they all encompassing. Rather, they allow for a translation of Peek's work in a fashion that supports consistent measurement and thus consistent description with common language.

Return to Contents 


A limitation to the progress of planning, implementation, policy, and financial changes necessary to eliminate the divide between mental health and medicine is the lack of a clear lexicon describing the parameters and values of collaborative care. Such a lexicon would have widespread use advancing the field of collaborative care and support both a research agenda for the field and a set of metrics consistent with the lexicon that can be used to operationalize the research agenda. This paper provides the framework and specifications of a set of metrics that can allow quality improvement within practices as well as provide a tool to assist in research to assess the comparative effectiveness of collaborative care. It allows us to respond to the call of the Butler et al evidence report to identify specific elements of collaborative care models that contribute to such effectiveness.5

The researchers suggest that the approach outlined in this paper is consistent with the major themes of contemporary quality and focuses on the dimensions of structure, process, and outcome that are core to all quality care evaluation. This paper may offer an opportunity for a national system of evaluating collaborative care of mental health and health behavior in medical settings


  1. Cunningham PJ. Beyond parity: primary care physicians' perspectives on access to mental health care. Health Aff 2009;28(3):w490-w501.
  2. Coyne JC, Thompson R, Palmer SC, Kagee A, Maunsell E. Should we screen for depression? Caveats and potential pitfalls. Applied & Preventive Psychology Spr 2000;9(2):101-121.
  3. Gilbody S, Sheldon T, House A. Screening and case-finding instruments for depression: a meta-analysis.[see comment]. CMAJ Apr 8 2008;178(8):997-1003.
  4. Williams JW Jr, Barrett J, Oxman T, et al. Treatment of dysthymia and minor depression in primary care: A randomized controlled trial in older adults. JAMA Sep 27 2000;284(12):1519-1526.
  5. Butler M, Kane RL, McAlpine 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.
  6. Donabedian A. An Introduction to Quality Assurance in Health Care. New York: Oxford University Press; 2003.
  7. Berwick DM, Nolan TW, Whittington J. The triple aim: care, health, and cost. Health Aff May 1, 2008;27(3):759-769.
  8. Eden J, Simone J. Assessing the Quality of Cancer Care: An Approach to Measurement in Georgia. Washington, DC: National Academy Press; 2005.
  9. Doherty WJ, McDaniel SH, Baird MA. Five levels of primary care/behavioral healthcare collaboration. Behavioral Healthcare Tomorrow 1996:25-28.

aCorresponding author: Rodger Kessler, Ph.D. Rodger.Kessler@uvm.edu

Page last reviewed June 2011
Internet Citation: A Framework for Collaborative Care Metrics: A National Agenda for Research in Collaborative Care. June 2011. Agency for Healthcare Research and Quality, Rockville, MD. http://archive.ahrq.gov/research/findings/final-reports/collaborativecare/collab2.html