A collaborative primary and mental health care model with psychologist and psychiatrist working in GP practices: process evaluation of the implementation, challenges, and sustainability | BMC Health Services Research

Adaptation of the collaborative care model

Planning and recruitment

The project lead, a psychiatrist with extensive experience in working collaboratively with GPs [40], conceived the project after a visit to the Hamilton FHT in 2012. The HFHT model had been developed in a country with a similar health care system to Norway’s and offered a promising approach to collaboration between GPs, other community primary care providers and specialised mental health services, in line with national health policies in Norway [29]. It was also feasible for the assessment and treatment of a broad group of GP patients.

A multidisciplinary research group was established, including family physicians, psychiatrists, a sociologist, a psychologist, and a health economist, in addition to a user representative. Between them, the team had extensive experience in clinical work at primary and specialist levels, in quantitative and qualitative research methods, and in evaluating health services. This group planned and prepared the project, supported the implementation, and evaluated the process and the outcome of the CRCT.

Local leadership of GPs within each borough, the director of specialised mental health services at Akershus University Hospital, and the CMHC manager also agreed to join the project. Managers of other primary care services in each borough were also interested in participation, but uncertain as to how much their services could be involved in the project.

The research group designed the preliminary adapted collaborative care model and the CRCT to test it and obtained approvals from the regional and national research ethics committees and the hospital’s data protection officer. Recruitment of GP practices began in 2015, and when two GP practices had been recruited in each borough, they were randomised to participate in either collaborative care (intervention group) or to continue with care as usual (control group).

The CMHC manager assigned three clinical psychologist specialists to each work half-time in a GP practice. Additionally, a psychiatrist, who was also part of the research team, was assigned to work one to two hours per week in each GP practice. This decision to use experienced clinical psychologists instead of mental health nurses or social workers was based upon previous positive experiences with collaborating psychologists and was supported by the Norwegian GP association.

Preparation

In September 2015, 11 people from the participating services (GPs, primary care managers and health workers, CMHC managers, psychiatrist, psychologists, child-adolescent psychiatrist) and three from the research group (GPs, psychiatrist) participated in a study visit to the HFHT in Hamilton, Ontario. This three-day visit included workshops and small group visits to GP practices to gain insights into the organisation of the HFHT model, local staffing, practices, and experiences.

In October 2015, a joint meeting of the participating services reviewed the proposed adapted collaborative care model, as well as a list of issues that needed to be clarified. The goal and aim of the project and the adapted collaborative model were defined and agreed upon as “The goal is to strengthen the health services provided by GPs in the field of mental health” and “The aim is to achieve better treatment, increased treatment satisfaction among both patients and service providers, and to find a model that is useful and sustainable over time.” These are included in the description of the adapted collaborative care model in Table 1. GPs could initiate collaborative care for any of their patients and were not focused on or limited to any specific patient groups, diagnoses or treatments.

The participating services agreed to initiate the collaborative care project, set to be operational for 18 months starting in early 2016. They also wanted the CMHC to manage the project. It was anticipated that national health authorities would support reimbursement for the collaborative activities in the project, which could potentially demonstrate the advantages of the national health policies focused on enhancing collaboration among mental health services and treatments provided by GPs and primary care. At the conclusion of the 18-month trial period, the participating services would determine whether and how the collaborative care model should continue.

Several practical issues were clarified at this stage, including staffing, office space for the mental health specialists from the CMHC, the organisation of the collaboration, access to electronic patient records, and applying to the national health authorities regarding reimbursement for collaborative activities. It was agreed that a joint project group led by the CMHC and with members of the participating services in each of the three boroughs would further refine the model and determine how it could best be introduced in the practices. This work took three months, and the final adapted model is described below and in Table 1.

The adapted collaborative care model

The adapted model consisted of two major components. The core component (collaboration within the GP practice) was a collaborative team co-located at each GP practice and consisting of the GPs, a half-time psychologist and a visiting psychiatrist (for two hours a week) from the CMHC. Office space was provided on-site, and the GPs and mental health specialists could easily contact each other to discuss a specific patient, a clinical issue, medications or a referral to another service. The mental health specialists provided assessments and short-term therapies for patients that the GPs referred to them, and the GP occasionally sat in on a consultation. Both GPs and mental health specialists had access to the patients’ electronic patient records and wrote their notes and summaries of assessments and treatments in the same electronic record.

The extension component (collaboration with other service providers) included the involvement of and collaboration with additional professionals from the borough’s other primary care, including addiction services, as well as with other specialised mental health services, meeting regularly or on an ad hoc basis to discuss specific patients or situations. This extension of the collaborative care team allowed for the flexible utilisation of expertise from a wider range of services.

Theory of change for the model as a complex intervention

Recognising collaborative care as a complex intervention, the assumed mechanism or theory of change for the model would be that close interaction between co-located service providers with different expertise would facilitate establishing and utilising relationships, learning from each other while working together, and providing more comprehensive and earlier assessments and treatments for patients with emerging mental health problems. For GPs, this would include learning to better detect and manage mental health problems. See Fig. 2.

The model did not include formal training in new competencies but assumed that participating GPs and mental health professionals would learn from each other as the project progressed. The implementation of the adapted collaborative care model focused on the process of collaboration and integrating mental health professionals in primary care, rather than on specific treatments to be employed [21, 41].

Implementation and experiences of the collaborative care model

The collaborative care teams were established during the winter of 2016 and were operational for 18 months from March/April 2016 to October 2017.

Implementation of the collaborative care model

The collaboration activities of the team to implement the core component were as follows. The GPs initiated ad hoc discussions on patients or clinical issues, joint meetings with patients, and referred patients to the psychologist for assessment or therapy by reserving time in the psychologist’s schedule. In general, the GPs demonstrated a keen interest in evaluating and treating patients with mental health challenges, although there was some variation between different GPs.

The therapeutic work of the psychologists was similar across the GP practices, providing short-term therapy much like in the CMHC outpatient clinic. On average the psychologists saw five patients per day, with a range from three to eight. Some patients referred by the GPs to the psychologists for therapy were deemed too ill for short-term therapy and they recommended a referral to the CMHC instead. Situations like these increased the GPs’ knowledge and experience in determining who, when and where to refer and what information to include in referrals.

All three psychologists experienced fluctuations in the collaboration over time, with some initial challenges and frustrations while introducing the model, followed by a period of successful collaboration. Two positive turning points were identified. One was when the three psychologists after three-four months began reporting more positively about collaboration and workload, feeling integrated into the collaborative care team rather than being viewed as external professionals. Another was later when the GPs stopped expressing concerns about the uncertainty regarding reimbursements for collaboration activities (see below). Challenges could arise, however, when GP substitutes replaced the GPs in the practices during summer holidays. One psychologist also found on return from a one-month leave that the GPs were reluctant to attend collaborative care team meetings and instead were referring an excessive number of patients to be seen.

In the beginning the psychiatrist was assigned two hours a week at each of the three GP practices, but over time this was adjusted to respond to specific needs, spending less time in a practice with lower needs and additional time during a psychologist’s leave in another practice. The psychiatrist primarily assisted the GPs with reviews of the use and adjustments of psychotropic medications in specific patients, including long-term injectables, seeing the person if further clarification was needed. This increased the GPs’ competence and comfort in prescribing these medications. The psychiatrist observed that the GPs increasingly became more aware of the need to review long term medications and detect and manage mental health issues that were concurrent with physical symptoms.

The psychiatrist also responded to questions from the psychologists about medications, usually meeting the patient together. They also assisted the psychologists with more complex cases, sometimes assessing the patient in person (along with the psychologist) and sometimes just by providing advice regarding further treatment options or referral to the CMHC.

In addition, the psychologists established contacts with the municipal primary care and substance abuse services in their boroughs, and the local Employment and Welfare Administration. In one borough, the psychologist spent half a day a week with their primary mental health care service, meeting with and supervising personnel and seeing some patients. In the two other boroughs, the psychologist’s contact with the borough’s primary care service was as needed. Few of the borough service providers participated in the joint semi-annual collaborative meetings to discuss their experiences and future directions for the collaboration, indicating a possible lack of engagement.

Mental health specialists from CAMHS and the outpatient clinic for substance abuse in the specialised mental health services could also be included when needed, but it was soon concluded that this was not necessary, as the GPs had well-established collaboration with these services and could easily contact them when needed.

Wider stakeholder involvement in the project was more limited than intended. A service user representative from the local group of the national user organisation Mental Health was recruited to one of the collaborative care teams. She was involved in team meetings regarding the collaboration but not clinical meetings about specific patients. In the two other GP practices, the project was unsuccessful in recruiting service user representatives to the teams, or in involving family caregivers in the planning or implementation.

There were some differences in contexts and local adaptations of the collaborative care model across the three GP practices, designated as team A, B, and C. In Team A (with 4 GPs) initial reactions to the project were mixed. While the GPs became increasingly positive over time, some remained hesitant to participate in collaborative meetings and patient discussions. Each GP in Team A had separate funding, which might have contributed to difficulties in forming joint plans for the entire group. Collaboration with the borough services was most consistent in Team A, where the manager in the borough’s primary mental health care team engaged several local services in a weekly half-day session with the psychologist for supervision and case reviews.

Team B (with 3 GPs) was the most integrated team, holding monthly meetings and regularly engaging in productive dialogues and joint case discussions. The psychologist and psychiatrist found it easy to discuss patients with the GPs after consultations, and each GP often allocated time to discuss medication with the psychiatrist. The project was seen to enhance local collaboration across various services, service levels, and disciplines.

Team C (with 4 GPs) felt they had fewer needs for collaboration with the borough primary care services. This may have been due to fewer socioeconomic problems in their part of the borough or because they treated many patients from outside the borough. Patients referred to the psychologist tended to have fewer complex issues compared to those in the other teams, with fewer referrals to the psychiatrist for medication reviews. The borough provided low threshold “rapid mental health care” services with psychologists, allowing the team’s psychologist to refer patients when pressed for time. Team meetings were less frequent in Team C.

Challenges in the implementation of collaborative care

Three significant barriers to implementation were identified: the workload of the psychologists, the GPs’ lack of time to attend meetings, and uncertainty about reimbursement for collaborative activities at GP practices.

Initially, the psychologists’ workload was excessive as they were assigned too many patients per day. As their schedules began to fill weeks in advance, their ability to schedule new appointments for patients needing more therapy became limited. This issue was resolved by allowing psychologists to take more control over their schedules, restrict the number of patients seen per day, and use a scheduling block function to reserve slots for future therapy sessions. These adjustments were made after discussions with the GPs at their practices.

GPs’ lack of time for attending collaboration meetings was a recurring issue, as their increasing administrative workload limited the available time for meetings.

Financial challenges emerged early in the planning phase, as implementing the model without additional funding presented significant difficulties for the participating services. GP practices incurred various expenses, including office space, access to electronic patient records, and secretarial support, in addition to the time dedicated to collaboration. For the CMHC, assigning two full-time staff equivalents to collaborate with the three GP practices reduced the number of consultations that could be provided at the outpatient clinic, thus increasing the workload for other clinical staff and potentially impacting annual funding.

It was unclear from the outset whether GPs could be reimbursed for patient consultations conducted by the psychologist or for time spent on collaborative activities, and patients could not be charged fees for consultations that weren’t eligible for reimbursement. Similarly, the CMHC could not charge fees or obtain reimbursement for patients seen by a psychologist or psychiatrist at a GP’s practice.

It was anticipated that national health authorities would recognise the CRCT’s value in relation to national health policy and offer flexibility regarding reimbursement during the 18-month implementation phase. Despite multiple requests over time to increasingly higher levels of national health authorities, no special reimbursement arrangements were made. Consequently, the collaboration had to proceed without reimbursement for collaborative activities, causing significant frustration for participants and prompting discussions about the implications for long-term sustainability.

Experiences with the collaborative care model (from the post-project workshop)

The GPs expressed satisfaction with the model and wanted to see it continue. They noted several benefits, including shorter waiting times for mental health care and greater accessibility when care was provided within their practices. Additionally, they experienced personal gains, such as heightened awareness of mental health issues and increased confidence in assessing and managing mental health problems. They also gained a better understanding of the services offered by the CMHC and the borough, feeling that the personal relationships developed with personnel from other services facilitated easier collaboration.

The psychologists and psychiatrist also reported positive experiences with the model. They gained insights into how GP’s work, the comorbidity of mental and physical illnesses, the treatment of physical health issues, and the available borough services. Compared to their experiences in the outpatient clinic, they saw that clinical improvements could occur more rapidly, as they were able to engage with patients at an earlier stage when problems might be less severe.

Healthcare professionals in the borough’s services had varied experiences with the collaborative care model. Some had participated in the visit to Hamilton and were familiar with the model, while others needed more time to understand how borough services could be involved in collaborative care. Variations in population needs and the organisation of borough services also impacted their participation. Overall, the participating borough services appreciated having a common meeting point to familiarise themselves with each other, understand one another’s practices, discuss issues, and learn about other services. They also recognised that greater utilisation could have been made of borough services such as the “rapid back to work” mental health program, the local Employment and Welfare Administration, and substance abuse primary care if they had been more involved and found ways to contribute their specific expertise.

Conclusion regarding sustainability of the model

Ultimately, despite the successful implementation and positive experiences of the collaborative care model, the lack of reimbursement for collaborative activities led both the GP practices and the CMHC to conclude that the model could not be sustained beyond the 18-month implementation phase without additional funding. The financial losses made it impossible for the GP practices to continue, and the CMHC lacked the economic and human resources to maintain the collaboration model or expand it to additional GP practices. The borough services had not committed any specific resources and were not involved in the decision to discontinue the collaborative care model.