Background

People with serious/severe mental illness (PWSMI) have a 13-to-30-year shorter life expectancy than the general population [1,2,3]. A number of factors contribute to this including higher rates of modifiable cardiometabolic risk factors such as smoking, hyperlipidaemia, hypertension, obesity, diet, and sedentary lifestyle [4, 5]. Use of psychotropic medications are also associated with increased risk of obesity, dyslipidaemia, Type 2 diabetes and cardiovascular risk [6]. Subsequently, this cohort is more likely to die of cardiovascular disease and respiratory disease than the general public [7]. Around 80% of PWSMI are diagnosed with at least one chronic physical health condition, and around 55% endure at least two [4]. The true number may in fact be higher due to low rates of routine screening and detection in these individuals [5, 8].

Poor uptake of preventive and proactive use of primary care is related to a number of health system issues (e.g., stigma, poor communication), consumer issues (e.g., cognitive difficulties, social circumstances and low health literacy) and social determinants of health [9, 10]. Nonpsychiatric inpatient admissions, length of stay, hospital readmission rates, and emergency department visits have been shown to be higher in medical patients with severe/serious mental illness (SMI) compared to patients without SMI [6]. Within Australia, complex funding arrangements mean primary care and mental health services are administered and funded separately, and this separation adds to the challenges of providing quality physical health care to PWSMI [11]. Primary care is funded through the Australian government and a universal public insurance scheme (Medicare) [12]. Responsibility for funding and regulating mental health services in Australia is shared among the Australian, state and territory governments, however the state and territory governments fund and manage community mental health services [13].

Over recent years, various models of care have been developed that aim to improve the way primary care and mental health services work together [14]. These models have largely been used to improve the management of mental health conditions (particularly depression) within primary care. Many different terms are used to describe these collaborative activities [15]. Integrated care is a broad term that refers to the provision of health services to meet the needs of individuals in a coordinated way [16]. Such care may be organised in a combination of ways ranging from simple linkage, through to coordination and the full integration of services [17]. Shared care is one system for achieving integration [18]. It refers to the planned delivery of care with enhanced information exchange [19] where both parties maintain ongoing involvement in patient care, share information and clinical responsibilities and proactively agree on common processes [20].

Systematic reviews of shared and collaborative arrangements between primary care and specialist services (including mental health) have shown mixed results for patients with chronic disease such as hypertension, kidney disease, stroke, depression, and anxiety [18, 21]. Shared or collaborative arrangements for those with SMI is less researched and tends to focus predominantly on how these models affect mental health management and outcomes [18]. A notable gap in the evidence includes how mental health services and primary care can better collaborate to create systems of care in which PWSMI are regularly seen by a General Practitioner (GP) while receiving specialist mental health care, and how mental health and primary care services can improve communication and integration of services [22].

This scoping review aimed to inform the work being conducted in Sydney Local Health District through the Shared Health Arrangements Research & Development (SHAReD) trial (https://www.unsw.edu.au/research/cphce/research/projects/shared--shared-health-arrangements-research---development) by identifying the barriers and facilitators to the participation and engagement of primary care (specifically GPs) in shared-care arrangements with community mental health services for preventive care of PWSMI.

Methods

The protocol for this review was published in Open Science Framework [23].

Research question and definitions

The research question addressed by this scoping review is: What are the barriers and facilitators to the participation and engagement of primary care in shared-care arrangements with community mental health services for preventive care of PWSMI?

This review adopted the following definitions:

  • Shared care: “A structured system for achieving integration of care across multiple autonomous providers and services with both primary and secondary care practitioners contributing to elements of a patient's overall package of care. Shared care involves some agreement about the shared activities and levels of responsibility for each provider and appropriate communication processes to support this integration. A shared-care arrangement may involve any combination of government, non-government or private sector providers” [18].

  • Serious mental illness/Severe mental illness: The terms serious or severe mental illness (SMI) are often used interchangeably [24] and refers to mental illness “which is severe in degree and persistent in duration, that causes a substantially diminished level of functioning in the primary aspects of daily living” [25]. SMI includes (but is not limited to) psychotic illnesses (primarily schizophrenia or bipolar affective disorder), severe depression, and severe anxiety disorders. It affects about 3% of Australians [26, 27].

  • Preventive care: Preventive health care is a range of activities with the goal to reduce the risk of ill-health or disability within an identified population [28].

Search strategy

Currently there is no Medical Subject Heading (MeSH) specifically for ‘shared care’, and the term is frequently used interchangeably with other terms such as ‘collaborative care’ and ‘integrated care’. Through testing and review of literature in the field, we identified several MeSH terms and additional key words for each of the major concepts within the research question (Table 1). The search strategy was refined within Medline prior to it being adapted for use in each database searched. A university librarian reviewed the search strategies prior to their application.

Table 1 Search strategy
Full size table

A search log documented the name of each database, the number of articles retrieved, the date of coverage and date searched (Table 2). Each search was stored within the platform so it could be re-run. Search results were imported to the Endnote reference manager and duplicate citations were removed and stored in an Endnote duplicates library.

Table 2 Databases searched
Full size table

Table 3 outlines the inclusion and exclusion criteria applied. Included were published data from the last 10 years (2011–2021) regardless of study design. Ten years was chosen as a suitable timeframe as it coincided with substantial Australian mental health reform[29,30,31], the introduction of shared care between mental health and primary care services for the management of consumers on clozapine [32, 33] and the establishment of shared care models in aspects of healthcare where non-GP specialists and GPs need to work together such as antenatal shared care, cancer, and diabetes [20]. Ten years was therefore seen as timely in which to find relevant data that we could incorporate into the (SHAReD) trial. Inclusion was limited to publications in English and those originating within Organisation for Economic Cooperation and Development (OECD) countries (as we wanted comparative health systems to that of Australia). Excluded were editorials, letters, opinion pieces, and protocols for trials, as well as studies of children and adolescents.

Table 3 Inclusion and exclusion criteria
Full size table

Study selection

Studies were excluded if a good proportion (75% or more) of the study sample did not have a mental health condition consistent with the persistent and enduring nature of SMI. We also excluded studies that did not fully describe their population, or those that used vague or general ways to define their population (e.g., those with ‘mental health issues’ or ‘behavioural health issues’).

Because of our interest in the interaction between primary care and mental health services that are typically separate entities within the Australian health care system, we excluded studies where the primary care service and the mental health service did not operate independently of each other; that is, where the two services were already combined in some way (e.g., shared systems, shared staff, shared funding). Services could be located close to each other if there was evidence from the publication of independence. We also excluded studies in which the interaction was between primary care and a service that did not align with our experience of a public community mental health service (i.e., a service providing assessment and support to people with mental illness living in the community). This excluded hospital based, outpatient or outreach services, and specialist services such as private psychiatrists or mental health in-patient units.

To be included in our review, the interaction described in the study needed to be specifically focused on the prevention or management of a physical health condition for PWSMI. We excluded studies of cancer care due to the specialised nature of this management.

Two authors (SP and KP) reviewed the titles and abstracts of all citations identified from the database search. Full-text articles were retrieved where it was not possible to determine inclusion/exclusion requirements from the title and abstract alone. Where there was disagreement or uncertainty, this was resolved through discussion or a third assessment (MH or CS). Reference lists of the most relevant citations were reviewed for additional literature that might have been missed through searching. We contacted the author of one publication where the final outcomes publication could not be sourced.

Data extraction and reporting

A data extraction template was developed in Excel. Two authors (SP and KP) independently extracted data and collaborated to finalise the extraction of included studies. A narrative synthesis with tabular mapping was used to describe and summarise the data identified through the database search. As this review was primarily interested in identifying the barriers to and facilitators of the engagement of primary care to shared-care arrangements, all reported barriers and facilitators were extracted from the included studies and organised under specific headings for the results.

The Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR, 2018) provided as a Supplementary file was used to guide reporting (1) [34]. The checklist contains 20 essential reporting items and two optional items (http://www.prisma-statement.org/Extensions/ScopingReviews).

Results

Study selection

As depicted in the Prisma flow diagram [35] (Fig. 1), our search identified 295 citations from 6 bibliographic databases, with 54 duplicates subsequently removed. We screened 244 citations using titles and abstracts which included these citations plus three additional citations obtained through handsearching. A further 195 citations were excluded at this point. We retrieved 49 full text articles of which 42 were excluded, leaving 7 included studies.

Fig. 1
figure 1

PRISMA 2020 Flow Diagram

Full size image

Description of the included studies

The seven included studies (Table 4) originated from Australia, the United Kingdom (UK), and the United States (US). A narrative synthesis was used to describe the studies and to summarise the barriers and facilitators identified within the studies.

Table 4 Description of the included studies
Full size table

Perkins 2010 [40] and Fitzpatrick 2017 [36] reported on an Australian rural general practice clinic providing assessment and management of physical health conditions for PWSMI managed by a mental health service. A designated consultation day was set aside due to shared recognition that access to comprehensive primary care was lacking. Perkins conducted a mixed-methods evaluation of the clinic using two years of service data (2007–2009) and 15 qualitative interviews with health care staff and others associated with the clinic. Fitzpatrick used qualitative data from a different date set (16 of the interviews conducted with healthcare staff during 2015/2016).

Hunt 2016 [37] described a ‘boundary spanner’ role initiated in England in which two members of the community mental health team were seconded as Community Physical Health Coordinators to span the primary care and community mental health service with a focus on improving communication and collaboration. Key informant interviews, a focus group, surveys, and quantitative data were collected, and the study reported on aspects of the role itself, outcomes of the multidisciplinary meetings, and changes in reporting from GP records.

Pastore 2013 [39] was an ‘easy-to-implement’ practice adaptation within a primary care setting in the US involving a Behavioural Health Liaison Officer, who was a registered nurse (RN) from the family practice who took on a coordination role and served as a single point of contact for patients with SMI. This study used pre/post data and a review of medical record data to evaluate the practice enhancement and use of the service.

Rossom 2020 [41] was a randomised controlled trial (RCT) from the US that provided interim study results on the implementation of a clinical decision support (CDS) tool that alerted both primary care and mental health clinicians for a given patient with SMI about increased cardiovascular (CV) risk.

Two publications from the US were included that described early efforts to set up integrated primary and mental health services to improve the physical health care of PWSMI. The Nover pilot program 2014 [38] used a partnership model to set up a clinic serving rural low-income individuals. A series of Plan-Do-Study-Act (PDSA) cycles were used by the clinic to set initiatives resulted in new forms for referral, assessment and treatment, and a shared care plan, all for use by the treatment team. Scharf 2013 [42] described a program which supported the integration of primary care services into community behavioural health settings for adults with SMI. This publication describes the grantee organisations (agencies that were the recipients of grants), their integrated care programs and implementation plans, and the challenges associated with this approach at start-up and after one year of program implementation.

Reporting of physical health outcomes

There was minimal reporting of physical health outcomes for patients within the included studies (Table 4). The Perkins study did not report on outcomes but described the physical health status of attendees. Over half (52%) of the clients assessed by the GP did not have a physical health illness and most (79%) were referred to other health and community services. The physical conditions reported among the participants were vascular disease/diabetes (10%), GI disorders (7%) and other (28%) [40]. In the Hunt study, although no physical health outcomes were reported, the study highlights that as a result of the multidisciplinary team meetings and audit of GP systems, disease reviews and actions related to physical health checks, and medication reviews were the most frequently discussed issues. The study reports improvement in recording of CV risk and improvement in the recording of systolic BP, HDL, smoking status, and BMI which was reported to improve from 38% of service users (n = 115) to 58% (n = 79) of users post project [37]. The Pastore study also did not report physical health outcomes but reported that at baseline 12.5% of patients went to the ER for a physical health problem, and 6% were hospitalised for a physical health condition [39].

Contribution of the literature to the research question.

To meet our definition of shared care, all included studies involved a general practice or equivalent. The studies reported a range of outcomes including feasibility, service operation and acceptability, provider viewpoint, provider reporting mechanisms and quality of life (Table 4). This review however was interested in the barriers and facilitators to the engagement of primary care in shared care for people with SMI. Therefore, to address our specific research question, we have separated our results as such. The barriers and facilitators reported tended to refer to the program/intervention as a whole rather than by individual services or health professionals. Therefore, while the discussion that follows is relevant to primary care and GPs, it may also be relevant to other health professionals involved in shared-care arrangements.

Facilitators to the participation and engagement of primary care in shared-care arrangements with community mental health services for preventive care of PWSMI

Although the yield from the database search was modest, we identified some items from the literature that had the potential to enable or enhance the participation and engagement of primary care in shared-care arrangements around the physical health needs of PWSMI. These included:

  1. a)

    Service models

Perkins [40] and Fitzpatrick [36] reported on the establishment of a GP clinic for consumers of the local mental health service in a small rural town of NSW with about 8,000 residents, in which designated appointments were set aside on a specified day to see patients with SMI. The authors report that there was little deviation from usual practice and no requirement to invest in new systems or to change service systems or the way staff work. While this clinic aimed to improve access to a GP rather than shared care, this model could also provide an opportunity for shared care to be established because of the increased collaboration between the practice and the mental health service based around individual clients. The authors argued that the key to success was the simplicity of the model; inherently no requirement by GPs to drastically change the way they practiced, no additional cost to services and a perceived mutual benefit by the partners involved [40]. They also acknowledged that a similar model might not be generalisable to other settings and was reliant on the ‘nimbleness, flexibility, resourcefulness, and persistence’ of individual health professionals to both set up and maintain the clinic [36].

  1. b)

    Specific roles to promote communication and coordination within existing systems

Most studies incorporated a specific role designed to provide active coordination [36,37,38,39,40]. These roles were not specifically based in general practice, and the outcomes assessed when these roles were utilised, tended to be how the role enhanced working relationships or cross service processes, or impacted patient outcomes. Generally speaking, coordination roles were seen as valuable because they improve the quality of care for patients [36]. When specifically designed to promote sharing or collaboration around certain tasks, the clinical and coordination ambiguity is removed and accountability for actions applied to specific individuals [39], resulting in better patient monitoring and subsequently better outcomes. Within this review we identified a variety of roles built into larger programs or operating as standalone interventions including Community and Physical Health Co-ordinators [37], Behavioural Health Liaison officers [39] and nurse-led coordinators [38]. Although these roles varied in name and function, at their core, they aimed to generate greater engagement in patient care and participation by all services. Engagement was enhanced because the roles fostered trust, created a link between services and improved respect for professional roles across the services [37]. At their simplest, the roles allocated a designated person to make and monitor appointments and transport people to appointments [36, 40]. The more complex arrangements used various staff who, through a variety of functions, could ‘span the boundaries’ between the primary care and mental health service [37, 39]. Both existing staff members [37, 40] and new staff [38] were allocated to take on these roles. Some shared arrangements utilised a prior position with designated liaison tasks with, or knowledge of, the collaborating services [36, 37, 39, 40]. These therefore drew on existing working relationships to promote effective mediation or liaison between the services.

While not a specific focus of this review, better engagement of patients was also associated with these roles because they incorporated a navigation function through secondary referrals and specialty care [39]. Coordinators also provided patient education to reconcile complicated medication lists, resulting in a higher rate of kept appointments and therefore improved care continuity [39]. In the study by Nover [38] for instance, nurse-led care coordination for medication support, illness management and referral/follow up was the most frequently requested service by patients.

  1. c)

    Improved collaboration and teamwork

In the Nover [38] study outcomes in relation to teamwork were more notable than the service quality and patient outcomes. This study reported that the psychiatric and primary care providers worked more closely together on patient health outcomes, that there was greater provider adherence to established standards of care, and a renewed emphasis on promoting healthy behaviours as a result of the partnership [38]. Providing better outcomes for patients is the core business of all health-related services, therefore achieving measurable benefits for patients engages health providers in the process because they can see the results that this brings. The fostering of clear roles and expectations creates a working environment where collaboration can be mutually beneficial, and there are realistic aims and expectations [36].

Joint care planning is acknowledged to promote engagement and partnership of health services around the needs of individual patients without making dramatic structural change [14]. In the Hunt study [37] the multidisciplinary team (MDT) meetings provided an avenue for multiple perspectives to be incorporated into an effective management plan for patients. Identifying key objectives and actions related to these allowed the team to ‘co-evolve’, develop new ways of communicating and working together, and to share and integrate knowledge to improve the physical health care of service users [37]. When shared care planning could be achieved more efficiently through the use of decision support systems, shared Electronic Health Records (EHRs), or other electronic means, there was greater provider adherence to established standards of care for chronic illness, and a renewed emphasis on promoting healthy behaviours [38]. Identifying major risk factors in PWSMI improved the opportunity for health professionals to intervene and to improve risk factor control in a timely way [41].

Enhancing communication between service providers, whether it be through MDT meetings, boundary roles or system changes acts to increase the willingness to collaborate [37]. Increased communication allows each provider to better understand both the medical and psychiatric needs of the patient [39]. This was identified as a key facilitator for designing and delivering appropriate management plans because it provided an opportunity to discuss individual cases, secure clinical consultation or specialist advice, provided new avenues for professional learning and opportunities for additional skill development [37, 39].

Barriers to the participation and engagement of primary care in shared-care arrangements with community mental health services for preventive care of PWSMI

The literature provided some indication of specific issues or factors that appeared to create a barrier or blockage to the participation/engagement of health services and health professionals in shared-care arrangements around the physical health needs of PWSMI.

  1. a)

    Willingness, confidence, and capacity

Shared care will only be successful if services and individuals have willingness to participate, confidence to carry out the required tasks and access to the physical structures that produce capacity (e.g., staffing, space). Experiencing problems in these areas will automatically impact service and individual engagement as it increases the time required to participate, the level of frustration experienced and ultimately impacts the perceived value of the activity.

The commitment and willingness of individuals was a resounding factor in the Perkins and Fitzpatrick studies [36, 40]. The idea to start a GP clinic for SMI was a joint decision by an individual GP and the community mental health service. In effect this GP volunteered his/her practice and time to provide the clinic, although he/she was still able to be remunerated through Medicare. Fitzpatrick [36] also noted widespread concerns by GPs about their ability to care for clients with complex SMI pointing to a perceived lack of experience, confidence, or knowledge. Providing education to staff involved in shared-care arrangements was offered to help them engage in the process and to feel confident in participating [38, 39]. This was deemed necessary for primary care staff so they could understand the behaviours they might experience and be able to respond accordingly [39], to increase the understanding around the importance of treating mental and physical health simultaneously [38] and also to increase the knowledge and understanding about the separate environments in which all partners work [39].

Larger projects in the US that reported on early integration across multiple sites also reported experiencing challenges related to staffing, organisational resistance, and internal capacity issues [42]. These included recruiting and retaining staff, general turnover and having sufficient space to provide an integrated service [38, 42]. Being able to adequately provide recovery-oriented mental health care also presented barriers and frustration among service providers. Within one Australian study, the limited capacity of the clinic to collaborate with other community and social services was noted [36]. Despite incorporating key aspects of recovery-oriented practice such as advocacy, social inclusion, and the support of individuals and their families to self-manage their physical and mental health, the study found less evidence for the necessary infrastructure and supportive relationships required to strengthen the recovery orientation of service delivery to include housing, employment, and income and vocational training support, all of which are recognised key elements to achieving positive recovery outcomes.

  1. b)

    Funding and incentives

Overwhelmingly, the need for adequate funding, funding restructure and monetary incentives to appropriately support shared care came through in this review. Poor alignment of incentives and payments with patient needs were seen as reinforcing previously established practices which effectively siloed services and excluded PWSMI [36]. This study (Fitzpatrick 2017) also noted that the GP clinic appeared constantly ‘under siege’ by the political and economic understandings that determined what should be considered as cost-effective and efficient practice. For example, having multiple people in a consultation for one person can be perceived as time wasting or inefficient from a payment/funding standpoint.

Delays in payments for services through bureaucratic processes [42] and missed provider payments if patients with SMI failed to turn up for appointments [36] all proved problematic in this review. General practice in Australia runs on a fee for service/small business model in which funding delays or inadequate funding may impact their engagement in shared care. Even programs with secured funding for the integration of primary care and mental health, noted a deficiency in adequate resources to promote the program more widely in the community and among allied services such as dieticians [38].

  1. c)

    Inability to share information

A common stumbling block for shared-care arrangements is the inability to share information about patients between services due to incompatible IT systems, or approval to access these systems. The inability to access and update information in real time adds pressure to work practices and frequently introduces repetition, particularly if information must be updated manually, is in multiple places, or is not accurate or up to date when it is required [37]. The Rossom study tested a clinical decision support system that could link with the EHR and provide alerts for all treating parties as a way of negotiating these split systems. The study found that this resulted in reduced disruption, and real time updates resulted in increased accuracy [41].

Across multiple sites within the Scharf study, EHRs, registries, and data sharing challenges were present at both baseline and follow-up. Although some software options were available to be bought that might provide a workaround, there were serious concerns from services regarding the quality and capacity of available packages, and electronic systems that efficiently and securely integrated behavioural and primary health information were simply not available [42].

  1. d)

    Challenges in engagement with PWSMI

PWSMI are a hard-to-reach group, and this makes it challenging for health services to provide the required care even with good, shared care systems in place [37]. This has traditionally caused problems with engagement of patients [37, 40] and providers alike since there are many points at which care can be derailed. The need for frequent in-patient care can further disrupt, particularly if the persons social structures (family/carers) and housing are impacted.

Within the Scharf study approximately one in four grantee organisations providing integrated care for PWSMI reported difficulty recruiting and retaining consumers. They reported (Page 663) “Many clients [are] referred to the program, but then they will either avoid our case manager’s attempts to contact them, or they will not show up for their intake” [42]. This study recommended the development of care managers and the use of peer specialists who may have the time and skills to engage with clients and ensure they attend appointments [42]. Hunt [37] suggested that more work be done to address access for difficult to reach groups. Studies that used tools to aid engagement of patients like health passports and support groups found they were poorly utilized [37, 39].

Discussion

Seven studies met our criteria for this review. We identified variable contribution to the research question of barriers and facilitators to the engagement of GPs in shared-care arrangements for PWSMI from this literature. This may relate in part to the parameters used within the review, and a broader focus on integrated services within the literature, as opposed to studies where GPs and the mental health service share responsibilities for care but otherwise operate as separate entities, as within the Australian context. This suggests the need for more research which evaluates GP participation or engagement more fully, particularly around the physical health needs/preventive care for PWSMI when primary and specialist mental health services are equally involved in patient care.

The studies identified used a variety of strategies to support shared care including special clinic sessions, boundary-spanner workers, decision-support tools, multidisciplinary shared-care plans, and shared health records. There was limited reporting of patient physical health outcomes in the studies as a result of these strategies. In some cases, studies highlighted that the strategy brought more focus to assessing physical health issues for PWSMI, or that clinical documentation or risk factor reporting improved. This did not translate to obvious improvements in physical health. Key facilitators of engagement included having a good fit with current organisation and practice, providing a dedicated role to coordinate the parties and help patients to navigate appointments, utilising multidisciplinary teams and facilitating shared health records in real time. Barriers included lack of motivation and confidence of primary care practitioners, poor alignment of funding/incentives, inability to share patient information and other demands on the life and care of PWSMI.

These findings are broadly consistent with other general research about shared care between specialist and primary care for people with long-term conditions. A recent review on collaborative care for depression found several barriers including lack of commitment, limited resources, poorly integrated information and communication systems, poor coordination of finances and care pathways and conflicting objectives. This review also stressed the importance of agreeing on guidelines for care, roles and responsibilities and “willingness to co-work and co-learn” [43]. A review of GP engagement with specialist palliative care services identified the importance of communication and information sharing, clear roles and pathways for referral and time and capacity within primary care [44]. An overview of reviews of interprofessional collaboration in primary care settings identified lack of time and training, lack of clear roles, fears relating to professional identity and poor communication as the main barriers [45]. Facilitators included co-location and a shared understanding of roles.

A feature of shared-care for PWSMI, that is not always present for shared-care with other populations, is the level of difficulty engaging consumers in the process. This was reported as a major recruitment and retention issue within the Scharf study [42] and Hunt reported that despite improvements in knowledge sharing among health care providers, access and engagement remained a problem for some ‘difficult to reach’ users [37]. There is a significant body of research on how PWSMI do not access primary care for preventive care for a range of reasons including patient factors (e.g., cognitive challenges, psychopathology) and provider factors (e.g., provider bias, skills and confidence in working with PWSMI) [8, 10]. The difficulties with shared care described above, sit on top of the difficulties in engaging PWSMI in primary care in the first place.

The review findings broadly align with Relational Coordination Theory. This framework assesses teamwork, in which having shared goals, shared knowledge, and mutual respect is seen to foster timely, accurate, problem-solving communication, enabling stakeholders to effectively coordinate their work across boundaries. These attributes are mutually reinforcing and are underpinned and supported by cross-cutting organisational structures and processes that encourage and increase stakeholder collaboration (Fig. 2). The theory further contends that strong networks of relational coordination facilitate the achievement of quality, efficiency, worker and learning outcomes [46, 47].

Fig. 2
figure 2

Relational Coordination Theory. Based on the figure from Bolton et al. 2021. Revisiting Relational Coordination: A Systematic Review. Journal of Applied behavioural Science 57:002188632199159. Permission to use this figure was obtained from Sage Publishing 23/01/2023

Full size image

Within this review there was evidence to support a number of relational components (Table 5) particularly shared goals, willingness to participate, increased communication and shared knowledge. Communication and relationships among healthcare professionals are particularly important when services operate as separate entities and essential if they are to improve the care of the patients they share. Within this review there was a focus on providing education to health care providers, not only to improve knowledge, confidence, and capability but also to improve their understanding about the services that they work closely with. Having shared accountability is particularly important in shared care arrangements to streamline the work and also to ensure that individual tasks are achieved. When services work together and experience these benefits, this acts as a mutual reward, which also comes from knowing patients are receiving appropriate care and continuity in their care. Organisationally, services benefited from ‘boundary spanner’, ‘hinge’ and case-coordination roles that reach out into primary care, having shared multidisciplinary meetings and communication systems and being able to share medical information in a timely way. Having guidelines and standards of care that were mutually agreeable and relevant was also desirable.

Table 5 Relational coordination theory domains identified within the included studies
Full size table

This review provides some future considerations for the design of shared care programs between mental health services and GPs providing preventive care for PWSMI. The traditional focus on education and guidelines, while useful, is likely to be insufficient to achieve shared care even for programs that involve minimal changes to general practice. Joint development and co-design of evidenced-based guidelines that have the input and support of both general practice and mental health and which clearly delineate responsibilities [48] may be beneficial however, effective shared care requires considerable changes to systems and structures and achieving these changes and establishing new systems can take considerable time. The ability to share patient information that is accurate, updated in real time and readily accessible by all parties within health information systems seems crucial and is currently a major barrier [49]. Measures to improve the connectivity of these systems with shared or accessible clinical records without adding additional processes that take extra time or lead to treatment burden for patients [49] would be highly valuable. Hence, commitment to providing appropriate and sustainable funding, incentives and support would be required across the participating services. Additional roles that provide coordination and promote collaboration such as boundary spanners and care coordinators can be valuable because they promote the working interface between services, they take on those tasks that often cannot be routinely achieved in clinical settings, and they develop the structures that encourage individual services to better communicate and work together. Therefore, providing nurse navigator and peer worker roles may assist PWSMI to navigate the complexity and also help provide coordination between services.

In any shared arrangement, contextual factors also need to be considered such as the stressors operating on the workforce. This implies that shared care relationships need to be underpinned by policy to provide sufficient capacity within general practice systems to constructively engage with shared care. This could include MBS funding for health assessments, QI activities or case conferences, and additional roles such as nurse navigators [50].

The positive aspects of the Relational Coordination model (shared goals, willingness to participate, increased communication and shared knowledge), although ideal, may not be easy to achieve in practice. Services are not static entities and the cultural and organisational patterns within services are often deeply embedded. This has considerable implications for practice. As indicated in the model, some experimentation to ‘close the gap’ may be required. Achieving better working relationships across services may need to happen slowly, through testing of strategies to see what works locally, and by making relevant and incremental changes. More research would be required however to identify the optimal way to achieve this in practice.

Strengths and limitations

The strength of this review lies in the systematic approach used to identify relevant studies and the use of recognised reporting frameworks. This review was limited by the lack of research specifically dedicated to answer the question of how to improve the engagement of primary care and general practitioners in shared care with PWSMI and their mental health services. A broader focus on integrated services within the literature may have resulted in missed opportunities to explore different collaborations, however this review was specifically looking at separate services and how they might work better in a shared arrangement. There was also limited reporting of physical health outcomes as a result of the interventions, therefore we cannot draw any conclusions about the evidence for changes in health outcomes resulting from changes to shared care arrangements. Study evaluations were often conducted over relatively short time frames, and this may have been insufficient time to identify system changes, as this can take considerable time. Research was not identified to provide evidence-based guidance on improving engagement in shared care in diverse settings (e.g., rural settings). As such any recommendations made within this review are not generalisable.

Conclusion

Consistency of these results with other research on shared care between specialists and primary care for people with other long-term conditions (e.g., cancer) suggests much of the broader literature on shared care is likely to be applicable to the context of GP-MHS shared care. However, PWSMI present particular challenges for recruitment and retention to a shared-care program. Our findings were consistent with relational coordination theory, which posits that performance outcomes rely upon providers sharing “goals and knowledge, mutual respect” and engaging in “frequent, timely, accurate, problem-solving communication”, supported by a range of structures such as shared information systems and roles that span services. These factors may be more important in engaging primary care in shared care arrangements than the traditional focus on incentives, education, and guidelines.