Public-Academic Partnerships: The Continuous Improvement for Veterans in Care: Mood Disorders (CIVIC-MD) Study, a VA-Academic Partnership

The Department of Veterans Affairs (VA) health care system provides an opportunity for conducting mental health services research that can inform the translation of research findings to clinical practice ( 1 ). With its extensive electronic data on clinical status and utilization from a relatively stable population, the VA has been an important setting for mental health services research and psychiatric clinical research. Moreover the VA serves a more vulnerable patient population than the general U.S. population (older, with co-occurring illnesses and lower income) ( 2 ). Including these groups in research studies is crucial for the development and implementation of effective treatments for vulnerable populations elsewhere. The VA's quality improvement mission also encourages the rapid translation of research findings to clinical practice, which has greatly informed the emerging field of implementation science ( 3 ).

Many VA facilities have successful mental health research programs through partnerships with university psychiatry departments. However, establishing successful VA-academic research programs has remained elusive for some VA facilities. Barriers include the cultural and administrative separation of the VA and academic research settings. The priorities of intense, efficacy-oriented clinical trials characteristic of some academic psychiatry programs may conflict with busy VA settings where the priority is on caring for veterans.

We describe the development of a VA-academic collaboration at a VA mental health facility that had not had a strong research program, particularly in mood disorders research, with its affiliated university's department of psychiatry. This collaboration occurred within the context of a naturalistic cohort study, implemented from 2003 to 2006, which focused on bipolar disorder: the Continuous Improvement for Veterans in Care: Mood Disorders (CIVIC-MD). We discuss how the partnership was established, critical elements required for its success, and its unique challenges and opportunities that can inform the establishment of research partnerships elsewhere.

Development of the partnership

CIVIC-MD began as a collaboration of administrators and providers from the VA Pittsburgh Medical Center's Highland Drive Campus and academic investigators from the University of Pittsburgh's Department of Medicine and Department of Psychiatry. Designed as a naturalistic, longitudinal cohort study, the goal of CIVIC-MD was to identify the patient factors associated with processes and outcomes of care among VA patients receiving care for bipolar disorder. The study involved a baseline and follow-up survey of clinical status, behaviors, access to care, vulnerability (homelessness), adherence, and treatment perceptions. Medical charts were reviewed for comorbidity, patient preferences, provider decision making, and quality of care. Administrative data on patient utilization were also ascertained from local data warehouses, including pharmacotherapy, laboratories, and visits. VA administrators were initially supportive of CIVIC-MD because the principal investigator was part of a VA Center of Excellence that was affiliated with the university's medical school. The principal investigator was a full-time VA employee and had implemented similar health services research studies in the VA's general medical clinic.

Setting

Highland Drive is a tertiary mental health care facility that represents one of three hospitals that comprise the VA Pittsburgh Healthcare System. It has a strong tradition of providing inpatient, outpatient, and day treatment for psychiatric and substance use disorders for VA patients throughout the western Pennsylvania region.

Although Highland Drive is technically affiliated with the university's medical school, it is located seven miles away and has not had a strong research partnership with mood disorders researchers from the university's psychiatry department. Some VA providers attribute this partly to the perception that the VA was often used solely for patient recruitment for university-led psychiatric studies. Recruitment often occurred without fully consulting VA administrators or providers. Some of these clinical studies were perceived as too burdensome for VA patients (requiring multiple psychotherapy visits for lower-functioning patients or medication switching or washout). Prior attempts to establish a mood disorders research program were hindered because of concerns that VA providers would be disproportionately burdened with study operations, recruitment, and patient assessment.

Conceptual framework

Because of these concerns, we felt that the critical elements needed to implement CIVIC-MD and make this VA-academic collaboration successful were to obtain input from VA administrators and providers regarding the types of research questions they would like to see addressed and the logistics of implementing the study.

To accomplish both, we used a framework referred to as participatory management ( 4 ). The framework derives principles from community-based participatory research and involves the development of research agendas and operationalization of research programs via extensive input from stakeholders ( 5 ). Stakeholders (in this case, VA providers, administrators, and patients) directly influence all phases of the study and are encouraged to identify research priorities ( 6 ). Participatory management practices have been associated with increased acceptability and participation of vulnerable populations in mental health research and with improved patient outcomes in mental health settings ( 6 ).

Implementing CIVIC-MD

We focused on bipolar disorder because it is associated with substantial morbidity, mortality, and health care costs ( 7 ). The alternating episodes of mania and depression and the comorbidities associated with this illness make research with this patient population particularly difficult. Bipolar disorder was considered a priority condition by VA Highland Drive administrators and providers primarily because patient outcomes for this illness remained suboptimal despite the availability of efficacious treatments ( 7 ). Moreover, there were no active studies on bipolar disorder at Highland Drive.

Administrator meetings

An early challenge during this process was garnering support from VA Highland Drive administrators and providers. Before applying for funding in 2003, we sought their input on barriers to care and optimal outcomes for patients with bipolar disorder. They expressed initial concern that the study focused too much on measuring quality and not enough on identifying the modifiable patient factors associated with poor outcomes. They pointed to medical illness and mental health patient no-show rates, which exceeded 40% in one clinic, as key barriers to optimal continuity of care and outcomes.

Consequently we modified our study to focus more on determinants of poor access to care and outcomes in bipolar disorder. Administrators' input strengthened the study and led to increased support for the project. We added a more rigorous assessment of mutable factors associated with outcomes (insight, therapeutic alliance, and adherence).

Provider focus groups

The VA providers attending the administrator meetings also expressed concerns in regard to provider and participant burden. To address these issues, we conducted a 90-minute focus group of providers in the fall of 2003 from the mood disorders clinic where recruitment would take place.

Providers were concerned about balancing perceived burden of recruitment with their desire to be involved in the process. They wanted to be involved in identifying patients who were ineligible (that is, too medically ill or psychiatrically impaired to complete surveys). Provider support in recruitment was crucial because the VA institutional review board did not allow unsolicited direct contact with patients. Consequently, providers assisted in developing a mechanism that allowed the CIVIC-MD coordinator to preidentify potentially eligible patients, and providers then used a standardized checklist to confirm ineligible patients.

Partnership successes

Through this partnership, we enrolled a substantial number of individuals with bipolar disorder. The nature of bipolar disorder renders this population challenging to engage and to follow over time. A total of 468 patients were enrolled in CIVIC-MD between July 2004 and July 2006, or 76% of eligible patients at the VA. The primary reason for study refusal was lack of time to complete the survey (for example, because of a pending outpatient appointment). Of the 468 patients enrolled, 435 (93%) completed the survey, and of those, the mean± SD age was 49±10.6 years (range 21–78 years), with 62 (14%) women and 58 (13%) African Americans. CIVIC-MD patients reflected the underlying bipolar patient population at Highland Drive (12% women and 13% African American overall). The sample included a substantial number of vulnerable patients: 12% (N=53) were homeless, 28% (N=123) had substance use disorders, and 78% (N=340) were unemployed.

Challenges and lessons learned

Some challenges in implementing CIVIC-MD, although unique to this VA facility, may reflect challenges experienced in forging academic-community partnerships elsewhere. Foremost was balancing comprehensive data collection with participant and provider burden. Mental health clinics within VA medical centers are often busy because they serve large catchment areas and many patients are referred from smaller primary care clinics in more rural areas. Hence research may be secondary to the clinic's priority of caring for veterans. In some cases providers did not have enough time in their schedule to refer patients to CIVIC-MD. The recruitment period was therefore extended to two years. We also regularly held working lunches to update providers on the study and disseminate preliminary findings.

Some VA data from electronic medical records were not available or were recorded inconsistently. A few years before CIVIC-MD began, Highland Drive ceased transcription services; providers were responsible for entering their progress notes into the electronic medical record. As a result, busy clinic schedules precluded many providers from entering detailed information on patient clinical status that would be useful for research purposes (such as symptoms and treatment preferences). Therefore, the patient surveys were crucial because they were used to collect clinical, utilization, and health behavior information not available from these other sources.

The key to successful implementation of CIVIC-MD was the ability to both obtain and incorporate input from VA administrators and providers. Another factor was fostering the relationship between the principal investigator and the VA mental health program staff. This partnership facilitated patient recruitment. To address concerns about time required for recruitment, the recruiter's office was located in the mood disorders clinic adjacent to the providers' offices. This access facilitated communication between the provider and survey coordinator.

The strength of the partnership and the use of provider input led to a more innovative project that captured data on patient factors that had not been widely available in previous bipolar disorder research studies (including access to medical care, therapeutic alliance, use of complementary and alternative medicine, and vulnerability).

Furthermore, CIVIC-MD could not be successful without an understanding of the VA's mission and focus on providing comprehensive care for veterans. CIVIC-MD was not a replication of a university-based research protocol but rather a study designed to address crucial questions regarding the health care and outcomes of veterans. Moreover, the principal investigator's full-time affiliation with the VA, familiarity with the VA system, and commitment to a VA research career were other factors contributing to the project's success. Specifically, the principal investigator felt that fostering this VA-academic partnership would facilitate the development of a mood disorders research program that reflected VA priorities, both locally and nationally, and focused on improving the care of veterans through implementing research findings into clinical practice.

CIVIC-MD also provided an opportunity for VA providers to learn about the behaviors, clinical status, and outcomes of their overall patient population. VA providers received newsletters that summarized study findings. Study participants responded to satisfaction surveys to provide input on the process. Moreover, CIVIC-MD served as a patient registry to implement an intervention to improve medical care for patients with bipolar disorder ( 8 ).

The current research-to-practice gap in mental health care can be reduced by engaging providers in settings outside of academia to treat more diverse, vulnerable patient populations. Although many VAs have established academic partnerships in psychiatric research, some still face challenges in establishing successful academic partnerships that are common in community-based practices. For such partnerships to be successful, researchers need to actively solicit and incorporate input on study design and operationalization from administrators and providers. Incorporating the interests and priorities of VA administrators and providers up-front, adapting surveys and recruitment protocols to be appropriate for veterans and VA settings, and rapid dissemination and application of study findings were critical elements to the success of CIVIC-MD.

Acknowledgments and disclosures

This research was supported by the Department of Veterans Affairs, Veterans Health Administration, Health Services Research and Development Service. The views expressed in this article are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs.

The authors report no competing interests.