Systems levers for commissioning primary mental healthcare: a rapid review

Primary Health Networks (PHNs) began operation in Australia on 1 July 2015. PHNs have been tasked with increasing the efficiency and effectiveness of medical services for patients, particularly those at risk of poor health outcomes, and improving coordination of care to ensure that patients receive the right care in the right place, at the right time (Australian Government Department of Health 2015a). PHNs are to play a key role in mental health reforms, particularly through the planning and commissioning of primary mental health services at a regional level (Australian Government Department of Health 2015b). This will be supported by a flexible funding pool for mental health and suicide prevention services (Australian Government Department of Health 2015b).

The Australian Government expects PHNs to undertake capacity building in, and implementation of, systems and processes that will support effective commissioning of services in the future. Commissioning is defined by the Department of Health as: ‘a strategic approach to procurement that is informed by the baseline needs assessment and associated market analysis […]. Commissioning is further characterised by ongoing assessment to monitor the quality of services and ensure that relevant contractual standards are fulfilled’ (Australian Government Department of Health 2016, p. 10). Dawda et al. (2016, p. 5) elaborate on this definition and emphasise that commissioning includes ‘a range of actions including strategic needs assessment; prioritisation; planning and designing services; options appraisal; sourcing; delivery; and monitoring and review’.

With respect to primary mental healthcare, the ultimate aims of reform are to alleviate the burden of mental illness and improve the quality of life for those who experience mental ill-health. With respect to the implementation of reforms, the Australian Government has identified that PHNs will play a role in implementing a stepped care approach to delivering primary mental healthcare services (National Mental Health Commission 2014; Australian Government Department of Health 2015b). Stepped care involves a continuum of mental health services that facilitates the delivery of least intensive, most appropriate, evidence-based treatments to patients according to their level of need (National Institute for Health and Care Excellence 2011; Australian Government Department of Health 2015b). In addition, following recommendations from the Primary Health Care Advisory Group in 2016, and as part of broader primary healthcare system reforms, a healthcare homes approach is being trialled in selected practices across 10 PHNs (Health Care Homes Implementation Advisory Group 2016; Primary Health Care Advisory Group 2016). Healthcare homes are a model of integrated, technologically enhanced, primary healthcare service, where co-located providers take responsibility for coordinating care and providing multidisciplinary services for eligible patients (Cummins et al. 2007; Ferrante 2010; Moore et al. 2014; Wilks et al. 2015; Primary Health Care Advisory Group 2016).

Systems levers are a means by which PHNs can fulfil the imperative to influence the Australian health system (Australian Healthcare and Hospital Association 2015a, 2015b). No bespoke definitions or frameworks exist to guide PHNs in effectively using levers to achieve health system change.

One framework for understanding the levers that governing bodies have at their disposal is provided by Roberts et al. (2009, p. 127), who defined policy levers as: ‘features of the system and strategies that governments can use in each arena to improve health sector performance’. They proposed a typology of five levers that can be applied by governments to deliver reform. The framework proposed by Roberts et al. (2009) has been successfully adapted for analysing Australian mental health reforms at a national level (Whiteford 2011; Grace et al. 2015; Meurk et al. 2016), and is potentially extensible to describing levers that can be employed by local health organisations (Table 1).

Table 1.  Definition of five policy lever types relevant to national level mental health reforms
Source: Grace et al. 2015

The primary aim of this rapid review was to identify and describe the systems levers that could be used to commission primary mental healthcare in a way that supports stepped care and healthcare homes approaches. The secondary aim was to adapt typology of policy levers proposed by Roberts et al. (2009) to describe systems levers relevant to regional level health governance.

The search comprised an academic database search and a grey literature search. Academic databases searched were: ProQuest Social Science, and Medline and PsycINFO (by Ovid). Search strings using Boolean operators took the form of ‘lever terms’ AND ‘primary mental healthcare model terms’ AND ‘country terms’ (Table 2). Lists of terms were internally connected by an OR operator. For the grey literature search, source documents were obtained from a pre-defined list of websites of government health departments, think tanks, online repositories and mental health professional bodies in countries with health systems comparable to those in Australia (Appendix 1). Health-specific websites were searched for the terms lever(s). Generalist sites were searched by adding the term ‘health’ to the search string, in order to restrict search results to health topics only. Reports that could be downloaded in PDF or DOC or DOCX formats were included for screening. Additional sources were identified through reference lists, citation searches and materials known to the authors.

Table 2.  Search terms used
Wildcard keywords were used as appropriate in order to capture variations in suffixes as well as plural and possessive forms of search terms

Following initial scoping of the literature available, we decided to take an inclusive approach. Documents were included if they explicitly or implicitly discussed levers, as defined by Roberts et al. (2009), and related to the commissioning of any healthcare services in a way that was relevant to primary mental healthcare. Inclusion criteria were: (1) the document was in English and about a high-income country or countries with health systems that are comparable to Australia’s in terms of governance, financing and the organisational role played by primary care, and are known to have implemented or piloted primary mental healthcare models at a national or state or provincial level. These included, but were not limited to, the United Kingdom, Canada, New Zealand and the Netherlands (Bywood et al. 2015; Mossialos et al. 2016); (2) the document was published between 2006 and 2016; (3) the document described, implicitly or explicitly, the application of any levers in the context of commissioning any healthcare services, except as identified in the exclusion criteria below; and (4) the document presented an analysis of primary data or described or reviewed previous or current approaches. Exclusion criteria were: (1) analysis of levers focused on national level reform or levers outside the sphere of PHN influence (e.g. local health networks as a lever of national reform, national responses to workforce issues); and (2) documents focused on levers relevant to laboratory testing, aged care homes, end-of-life care, prison care, population-level prevention and promotion, secondary care, integration between health and social services, or prescribing and pharmacy.

Documents were initially excluded through title and abstract screening (title and executive summary screening for reports). Further exclusions occurred as analysis proceeded. Basic characteristics of each document were extracted from each source, specifically: Author/Year; Study type; Area of health; Levers identified; and Brief summary of the document, including information about country of focus and specifics of models or programs where levers were discussed. Classifications for area of health were: Primary mental healthcare; Primary healthcare; Mental health, other; and Health, other. Evidence regarding the outcomes of applying levers was extracted, if reported.

Documents were classified according to their methods: qualitative, quantitative, mixed methods, review or position paper (categories adapted from Gardner et al. 2016). One descriptor was allocated per document. Given that the aim of this article was to identify levers, we did not apply quality criteria or bias assessment tools.

A deductive–inductive approach to classifying lever types and sub-types was undertaken, as follows: C. Meurk sorted literature into categories according to each of the five lever types given in Table 1 (deductive analysis). This typology was adapted as new categories of lever became apparent and initial ones were identified as irrelevant to regional health governance (inductive analysis). As this process unfolded, previously categorised literature was re-read and re-sorted as necessary. This process was interpretive and occurred in tandem with the extraction, analysis and synthesis of lever descriptions, and descriptions of their significance and use. Where levers could be multiply assigned, they were classified according to a primary mechanism – what has been termed a ‘leading edge’ approach to classification (Grace et al. 2015). Supporting levers were documented. A sample of 10 references was cross-checked by E. Wright and discrepancies in the classification of levers were resolved through negotiation between C. Meurk and E. Wright.

A narrative synthesis of results included an appraisal of the relevance of levers with respect to key challenges for implementing stepped care, such as service integration, team-based care and continuity of care, where possible. Priority in interpretation was given to documents that evaluated the use of levers most relevant to the primary mental healthcare context and to an Australian context.

The search resulted in 56 documents being included in the narrative synthesis (Fig. 1).

Fig. 1.  PRISMA flowchart. Adapted with permission from Moher et al. (2009).

In total, 39 of the 56 documents focussed on primary healthcare, 12 on primary mental healthcare, 1 focussed on other areas of mental health, and 4 focussed on other areas of health (Table 3). Twenty-seven documents were classified as using qualitative methods, nine as using quantitative methods and five as using mixed methods. A further 13 documents were classified as reviews and 2 as position papers.

Table 3.  Characteristics of documents reviewed
PMHC, Primary Mental Health Care; PHC, Primary Health Care; MH, Mental Health, Other; H, Health, Other; Org, Organisation; Eng, Engagement; Enf, Enforcement; Info, Information; Tech, Technology; Fin&Pay, Finance and Payment

Six lever types and twenty-six different levers were identified. Organisation levers were identified in 19 documents, engagement levers in 17 documents, enforcement levers in 9 documents, information levers in 18 documents, technology levers in 11 documents, and finance and payment levers in 14 documents. Table 4 defines and identifies these lever types, the correspondence between the new typology and the typology proposed by Roberts et al. (2009), a justification for the changes made, and a list of specific levers that were mapped to each lever type.

Table 4.  Lever types identified in the included documents
Source for National level lever types is Grace et al. (2015), adapted from Roberts et al. (2009)

For each lever, Table 5 summarises its application(s) in the context of primary mental healthcare, its significance and use, and supporting levers (i.e. levers that are used in conjunction with the principle lever). Additional detail is given below.

Table 5.  Summary of levers for commissioning primary mental healthcare
En-dashes indicate that descriptions of key findings are intentionally left blank

Organisation levers aim to directly influence the structure of healthcare organisations, the workforce, the organisation of help seekers with respect to available services in time and space, and features that facilitate effective organisational cultures. Several of the organisation levers that were identified in the review (namely, leadership, nursing and contracting external support) were considered to be important to health systems functioning and primary care, as opposed to being only relevant to mental healthcare.

Referral management was identified as being particularly important to implementing a stepped care approach to mental healthcare (Bywood et al. 2015). Within Australia, GP referrals and mental health treatment plans are the most common referral pathways used to facilitate primary mental healthcare (Harris et al. 2010; Bassilios et al. 2013). GP mental health plans followed by vouchers and third party brokers have been the most common way to facilitate referrals to psychologists (Bassilios et al. 2013; Bywood et al. 2015). More recent innovations are technological and include web-based portals that can assist GPs to assess, manage and appropriately refer patients (WentWest Ltd 2015).

Enrolment with a primary care provider was identified as relevant to implementing a healthcare homes model, and a potential facilitator of achieving improved outcomes for people with complex needs (e.g. people with serious mental illness) (Douglas et al. 2009; Bywood et al. 2015).

Engagement refers to the activities of fostering and maintaining relationships, and dialogue, among the actors that comprise the health system. Engagement may refer to the interactions within commissioning organisations and between commissioning organisations and service providers, clinicians and public and patient groups (Naik et al. 2013; Perkins et al. 2014). Engagement was described as a means by which commissioners could influence provider behaviour in order to improve performance (Grant et al. 2015). Franx et al. (2013) identified power sharing, mutual respect and two-way capacity building as key elements of engagement. Bywood et al. (2015) describe the key goals of engagement as identifying shared goals and achieving mutual benefit. Effective engagement was noted as important for enabling health professionals to act autonomously; conversely, poor relationships and communication may undermine this, creating uncertainty, frustration and inefficiency (Checkland et al. 2009; McCafferty et al. 2012; Checkland et al. 2013; Zachariadis et al. 2013).

Successful engagement depends on stakeholder buy-in and trust, and is undermined where time pressures preclude stakeholder involvement (Naik et al. 2013; Ashman and Willcocks 2014; Smiddy et al. 2015). Financial incentives or contractual obligation may be required to facilitate engagement (Box 2009; Smiddy et al. 2015). Engaging people with lived experience of mental illness was highlighted as being particularly important to implementing integrated care (Bywood et al. 2015). Effectively engaging referrers (e.g. GPs, emergency departments) was a key facilitating factor in the successful operation of Australia’s Access to Allied Psychological Services (ATAPS), a major primary mental healthcare program implemented from 2001 to 2016 (Bassilios et al. 2013).

Engagement may be viewed as dovetailing with and offsetting the excesses of the so-called ‘hard’ levers (i.e. enforcement levers) (Chambers et al. 2013; Grant et al. 2015). In contrast to ‘soft’ levers like engagement – that are used to improve performance but do not possess sufficient power to hold underperforming organisations to account – ‘hard’ levers provide a sanction-backed means of compelling action by incentivising compliance and deterring deviance (Addicott 2014; Grant et al. 2015; Schwartzkoff and Sturgess 2015; WentWest Ltd 2015). Hence, enforcement is also linked with finance and payment levers. Although legally backed agreements are necessary in enabling commissioning organisations to hold underperforming organisations to account, competitive approaches (defined as ‘rivalry in the market, striving for custom between those who have the same commodities to dispose of’ (OED Online 2017)) or overly legalistic approaches to management were seen to run the risk of undermining engagement, goodwill and cooperation (Addicott 2014; Grant et al. 2015; Schwartzkoff and Sturgess 2015). Having an emphasis on contestability (‘the credible threat or possibility of competition’; Schwartzkoff and Sturgess 2015, p. 31), rather than competition, was described as a means to offset these negative effects (Grant et al. 2015; Schwartzkoff and Sturgess 2015).

An evaluation of the ATAPS program found that retention of mental health professionals by contractual arrangements was viewed as disadvantageous, relative to retention by employment, from the point of view of mental health professionals (Bassilios et al. 2013). In particular, retention by employment was believed to offer greater continuity of care in comparison to contracting (Bassilios et al. 2013). Retention by contracting was deemed to be cheaper for commissioners of services as it allowed them to externalise costs to mental health professionals and was seen to be advantageous in offering greater choice of provider to referrers and patients (Bassilios et al. 2013).

The information lever encompasses the collection, analysis, dissemination and exchange of information. Information can influence a system in multiple ways and is critical to evidence-based practice, performance management, accountability and quality improvement (Perera et al. 2013; Ashton 2015; Ham et al. 2015). Information is also important to making resource allocation decisions, fostering inter-organisation engagement, and engaging and fostering trust among stakeholders and the broader community (Franx et al. 2013; Perkins et al. 2014; Bywood et al. 2015; Ham et al. 2015). Information levers link with technology levers and payment levers, especially pay-for-performance approaches.

Successful use of data to inform commissioning is dependent in equal measures on technical skill and stakeholder support of data collection activities (Perera et al. 2013; Ashton 2015; Samra et al. 2015; Ball et al. 2016); poorly selected performance indicators run the risk of negatively affecting performance (Jacobs et al. 2006). Tremendous effort has been expended in Australia to establish high-quality routine data collection of mental healthcare processes and outcomes, as well as to develop the mental health research workforce (Bywood et al. 2015).

Information overload was described as posing a key barrier to implementing evidence-based practice, including decommissioning ineffective practices (Shepperd et al. 2013). To address this issue, systems to effectively manage the quantity of evidence, and monitor and implement evidence-based practice are needed (Shepperd et al. 2013). Lack of knowledge and skills have disproportionately negative effects on addressing the health needs of marginalised groups and minorities (Salway et al. 2016). It was noted that dissemination activities that involve providing negative feedback may cause hostility, and should be approached with sensitivity (Ball et al. 2016).

Technology is a lever for improving the efficient and effective functioning of the health system. It is closely related to the information levers and a necessary pre-requisite for the successful use of several other levers (e.g. organisation, enforcement, and finance and payment levers). Technology includes the use of health information systems, web portals, data visualisation tools and decision support tools. Health information systems, including electronic medical records, are a tool for improving data collection and increasingly a part of facilitating effective referral management, thus they are important to promoting service integration and continuity of care (Franx et al. 2013; Bywood et al. 2015).

Effective use of health information systems was described as enabling GPs to occupy the centre of a healthcare home model, even when services are not co-located, thus facilitating patient choice of provider (WentWest Ltd 2015). Decision support tools include the technologies mandated for use by the Australian Government (e.g. needs assessment), but can include several other tools – such as health technology assessment – which, although less often used, may enable effective commissioning (Elston and Stein 2007; Robinson et al. 2012; Goodwin and Frew 2013).

Finance and payment act as levers by providing the resources necessary to fund healthcare activities. Several mechanisms for financing and payment were identified, including targeted marginal investment, block funding, case-based funding, capitation, fee-for-service and pay-for-performance (also referred to as payment by outcomes). Different models of payment reflect different trade-offs between administrative simplicity and target specificity, and engender different distributions of financial risk and responsibility between commissioning organisations and service providers (Marshall et al. 2014). In practice, blended arrangements are often used (Marshall et al. 2014). The pay-for-performance approach, which fuses financial and contractual mechanisms, was the most commonly discussed payment model in the literature reviewed, with several commentators expressing the need for caution in utilising this approach in terms of its possible deleterious effects on intrinsic motivation and continuity of care (Buetow 2008; Campbell et al. 2008; Marshall et al. 2014; Schwartzkoff and Sturgess 2015). The success of pay-for-performance approaches depends on the ability to select and monitor performance indicators that align with, and therefore incentivise, desirable activities and outcomes that are within a provider’s control and can be measured over the period of a contract (Schwartzkoff and Sturgess 2015). Implementation of payment by outcomes that result in delays in payment or present a significant financial risk to the contractor may be detrimental to the mental health system (Schwartzkoff and Sturgess 2015).

A Cochrane review of evidence on the effect of financial incentives on the behaviour of primary care practitioners found insufficient studies to judge the effectiveness of incentives (Scott et al. 2011). Ludwick et al. (2010) found that mode of remuneration did not affect the implementation of electronic medical records in Canada, whereas a review by McDonald et al. (2008) found inconclusive evidence linking mode of remuneration to improvements in team-based primary care.

Effective primary mental healthcare depends on creating a system that patients can easily navigate and that is flexible to the way that symptoms and needs fluctuate. Facilitating service integration, multidisciplinary team-based care, and continuity of care are of paramount importance to achieving this. Although all levers identified in this review have applicability to PHNs with respect to their roles in improving health system functioning and primary care in general, the use of referral management, contracts, tendering and health information systems have specific implications to improving the performance of mental health systems.

Effective referral management, which focusses on mechanisms for facilitating patient access to the most appropriate, least restrictive care for their symptoms, will likely be critical to achieving better system integration and continuity of care. Contractual and payment approaches, specifically those that focus on contestability rather than competition, may help foster constructive and supportive relationships across providers that are critical to achieving system stability, integration and continuity.

Health information systems can be supportive of both mental health services and a healthcare homes model of primary care. In particular, the adoption of single multi-agency care plans may be an important facilitator of integration and continuity of care (Department of Health 2016).

Importantly, lever use is context-specific and multiple levers often work together. This is important to recognise for their successful application. However, it also raises methodological challenges in terms of how levers can be evaluated in order to build a much-needed evidence base to support their use (Greenhalgh et al. 2009; Cartwright and Hardie 2012; Grace et al. 2017).

The key strength of this review is that it describes a range of levers from health systems that are similar to those in Australia. There are four key limitations to note. First, the classification of content according to levers relied on judgement and interpretation. Other researchers could have classified content differently. We do not think this potential variability undermines our analysis, given that it is based on an existing framework and that we undertook cross-checking of a sample of documents. Second, in the documents included in this review, only one profession (nursing) was considered as a ‘lever’ (see Table 5). Further work needs to be done to establish where and when health professionals and the workforce should be viewed as ‘levers’. Third, the literature was sparse and not necessarily based on robust evaluation methods. There is a need for further research to develop an evidence base for policy (Gardner et al. 2016). Finally, although this review focussed on the use of levers at a single level of governance (i.e. at a regional level), in practice, levers can be enacted across multiple levels of governance. Future research on levers that cross levels is warranted.

This review identified 26 different levers that could be used by Primary Health Networks to influence the mental health system. Although all of these levers play an important role in health system functioning in general, referral management, health information systems, contracts and tendering models may be uniquely relevant to achieving the appropriate streaming of help seekers, service integration and continuity of care – critical success factors for a stepped care model. Enrolment, location, health information systems and capitation may be particularly important for implementing a healthcare homes model. Further reviews and primary research are needed to evaluate the use of systems levers in commissioning primary mental healthcare by local health organisations. PHNs can help build an evidence base for levers by documenting the levers that they use, evaluating their use, and making this information available to researchers and other commissioning bodies for further analysis and synthesis.

The authors declare that they have no conflicts of interest.