Mental health service system improvement: Translating evidence into policy

Introduction

Mental disorders (including alcohol and illicit drug use disorders) are prevalent and a leading cause of health-related disability. Improving access to, and the quality of, treatments for mental disorders is a global health challenge (Collins et al., 2011) with calls to use the evidence on effective interventions to scale up services to reduce the burden of mental disorders and develop more strategic and coordinated approaches to mental health planning and service delivery. In Australia nationally coordinated mental health reform commenced in 1992 and mental health was made a national health priority area in 1996. Despite over 20 years of reform, considerable public, professional and political concerns about deficiencies in mental health services remain (Senate Community Affairs Committee, 2011).

Mental health reform has almost always been incremental and cyclical. Windows of opportunity open from time to time which allow change to occur (Kingdon, 1995). Often these windows open when there is a political imperative to ‘do something’, driven by a sense of crisis and an urgency to address a problem, especially when the problem and its consequences are being repeatedly highlighted in the media. A real or perceived crisis needing urgent attention has often been a driver of government decisions to undertake mental health reform in Australia. In the lead up to the 1992 launch of the National Mental Health Policy and the 1993 Commonwealth budget, the Human Rights and Equal Opportunity Commission inquiry (the Burdekin inquiry) was identifying and publicising major shortcomings in services. The 1998 Coalition reforms were preceded by the Port Arthur shootings and, although not found to be a result of a deficiency in mental health services, the tragedy focused attention on mental health issues in political circles and a need for services in the community. Prior to the Council of Australian Governments (COAG) National Action Plan on Mental Health in 2006, the findings of the Not for Service report by the Mental Health Council of Australia, Senate inquiries and the failures that occurred in the treatment of Cornelia Rau (an Australian with a severe mental illness who was unlawfully detained for 10 months as a suspected illegal immigrant) were all very prominent. There have been similar, well-publicised examples of service failure and personal tragedy, forcing a response from government at State and Territory levels.

Incremental changes, adopted in response to a crisis, will lead to piecemeal, ad hoc reform. We have never used our understanding of the needs of the population in Australia to develop a comprehensive concept of what an efficient mental health service system, from prevention through to recovery, should look like. And when service models have been developed, we have never really articulated what the policy settings of governments needed to be, in order to bring that model into existence. There is a need for an approach to mental health reform in Australia that takes a long-term perspective, informed by evidence and that does not rely on ‘crisis’-driven, ad hoc changes to service delivery.

Let’s start with how big the burden is and identify what will reduce it

A first step in taking this approach is to determine how big the burden is and to identify what will reduce it. There are a number of ways to measure the burden resulting from diseases and injuries such as mental disorders. A commonly used measure is the disability adjusted life year (DALY). The DALY is a generic metric combining years of life lost and years lived with disability. While it has limitations, it does provide a comprehensive set of numbers, comparably determined across diseases. DALYs are an accepted and frequently used measure of disease burden in Australia and have been used in a series of priority-setting exercises endorsed by government decision-makers and clinical stakeholders (e.g. Mihalopoulos et al., 2011; Vos et al., 2005).

Reduction in disease burden can be achieved by: (1) decreasing disorder incidence (primary prevention); (2) decreasing the duration of a disorder; (3) reducing disorder severity (resulting in a reduction in disability); and (4) reducing premature mortality. A service system should provide interventions targeting these areas, on the continuum from primary prevention delivered in population health programs; self-help, including internet-based therapies; early intervention, primary care and specialist treatment; and rehabilitation and recovery programs. Resource allocation to these areas should be informed by the cost, effectiveness and equity of the interventions in reducing burden.

There has been a series of studies to establish priorities for burden reduction in the treatment and prevention of mental disorders in Australia. The Assessing Cost-Effectiveness (ACE) approach to priority-setting identified both cost-effective and less cost-effective interventions for mental disorders, but noted a more efficient service system would require reallocation of funds between interventions, between disorders and between service providers with different funding mechanisms (Vos et al., 2005). The Tolkien II project (Andrews and Tolkien II Team, 2007) used data on the prevalence and disability associated with 15 mental and substance use disorders to estimate the burden reduced by a mental health service, using a stepped care model to provide optimal care to the people currently in contact with the system by reorganising the system and using the current budgets. Tolkien II estimated that a 30% increase in budget could treat 60% more people and produce a 90% increase in health gain (DALYs averted).

While we need to know the size and distribution of the burden and the most cost-effective interventions to reduce it, this information has had limited success in promoting significant change in the service system, despite some studies engaging policymakers from an early stage. Clearly, translation of evidence into policy requires more than information on the burden and cost-effectiveness of treating disorders and more than policymaker engagement in this research. When changes in service systems occur, they are usually in response to a change in policy settings such as regulatory requirements or the introduction of new incentives, and/or from the allocation of new resources. The evidence used to inform the changes in policy settings can be fragmented or even irrelevant to the policy goals. Information on the epidemiology of disorders across the lifespan is often not considered (McGorry et al., 2007), cost-effectiveness alone rarely influences priority setting (Sabik and Lie, 2008), and integrative interventions have not been able to compete when funding decisions are made within service silos (e.g. within, rather than between, prevention programs and clinical services) (Segal et al., 2010). Those who plan services are also constrained by the political demand to respond to vocal stakeholders and media stories about service deficiencies. So, how do we make the evidence more relevant and compelling for policymakers?

Next, organise interventions into a service delivery framework

Organising cost-effective interventions into a service delivery framework will require expanding the evidence base beyond disorder-level interventions to service-level interventions. While clinicians and many cost-effectiveness studies typically focus on interventions for individual disorders, policymakers and service planners are more engaged in issues of access to, and the effectiveness of, service systems. These service ‘platforms’ include internet-supported therapeutic interventions, primary care, drug and alcohol services, specialist mental health services and community mental health support services. They are administered by Commonwealth, State and Territory governments, the private sector and not-for-profit, non-government agencies. Policymakers and service planners are therefore likely to be more receptive to information about the evidence for the cost-effectiveness of service platforms than to evidence for the cost-effectiveness of disorder-level interventions. The concept of interventions by service platform is being adopted by international agencies promoting service reform (e.g. the Disease Control Priorities Network funded by the Bill and Melinda Gates Foundation; http://globalhealth.washington.edu/project/4236).

Changing the delivery of health services requires not only workforce changes but changes in the way the services are organised and funded (Brooks et al., 2011). To design a service system that delivers optimal interventions matched to population need and distribution requires a service taxonomy, a description of the work of each service component or platform and the burden it averts, and specification of how each platform in the system could be changed to avert more burden at the same or lower cost. Once the service system is developed, empirically derived planning targets can be obtained to quantify the required changes in service delivery and to monitor progress towards agreed goals. To date, there have been few examples of such empirically derived, government-endorsed planning targets (Pirkis et al., 2007).

Finally, adapt policy settings to implement the desired service framework

Improved service planning requires more than the best information about cost-effective interventions, allocated in an ideal service system with empirically derived planning targets. It requires accurate knowledge of the policy instruments through which the desired reforms would be implemented, as these are the way in which governments do their work (Salamon, 2002). Service planning also requires engagement with key policymakers to identify the relevant policy instruments and an understanding of, and an ability to influence, government decision-making (Russell et al., 2008).

The policy instruments available to government include standards, licensing, regulation, incentives, coordination, education and information provision. Relevant economic instruments include financing and funding system design. In the health system, these instruments come from fields such as evidence-based medicine (e.g. clinical practice guidelines), public management (e.g. accreditation standards and performance indicators) and financing (e.g. activity-based funding). The critical importance of these instruments in the health sector for health policy research and reform has been emphasised by a number of Australian researchers (e.g. Gillespie et al., 2009).

The interrelationship between different policy instruments determines their overall effectiveness (Eliadis et al., 2005). A lack of coordination between instruments has contributed to poor or failed reform in areas such as health insurance (van de Ven et al., 2003), pharmaceutical policy (Paraponaris et al., 2004), coordinated care (Segal et al., 2004), tobacco control (Givel, 2006) and obesity (Swinburn, 2008). The same has been true for mental health, with the service system failing to optimally reduce the burden of mental disorders and unable to strengthen coordination linkages to other support services such as housing and disability (Betts and Thornicroft, 2001).

While government policy instruments are critical to implementing mental health reform at the system level, translation of evidence into service delivery may also occur by influencing service managers, clinicians, consumers, the general public and academic researchers. Clinicians, for example, are critical to engage in changing intervention guidelines. While it is important to engage all of these groups in service reform, government policy instruments remain key for implementing changes in financing, coordination and regulation.

Is this being done?

The impact of the small number of evidence-based planning models that have been developed in Australia and elsewhere (Pirkis et al., 2007) has been constrained by one or more of the following factors: incomplete epidemiological and burden of disease estimates; limited coverage of the population across the lifespan; limited coverage of the intervention spectrum; and limited coverage of the service sectors that comprise the mental health service system. Moreover, none of those models has attempted to identify and influence the government policy instruments needed to translate the model into successful service reform.

To address this, the National Health and Medical Research Council has funded a Centre of Research Excellence in Mental Health Systems Improvement to research and advise on an optimal mental health service system for Australia, to maximally avert the burden of mental disorders at the lowest cost, based on the approach outlined in this paper. This Centre of Research Excellence is designed to address many of the gaps in previous evidence-based mental health planning efforts, build upon current efforts (e.g. the National Mental Health Service Planning Framework project; http://www.health.gov.au/internet/main/publishing.nsf/Content/mental-nmhspf) and highlight evidence gaps that need further research to inform evidence-based service planning.

While research such as this is underway, immediate improvements can be made. Three examples are given below. The first is by increasing coverage where the burden of disease is high. For example, treatment rates for major childhood mental disorders are low, despite mental and behavioural disorders being the leading cause of DALYs in Australians aged 5–14 years (Institute for Health Metrics and Evaluation, 2013). Improving service coverage for these disorders is vital given that early intervention can reduce the burden presented by these disorders both in childhood and later in adulthood.

Second, improvements can be made by applying current knowledge to better align the mental health needs of the population to existing service platforms. For example, there is now strong evidence for the efficacy and real-world effectiveness of both self-guided and clinician-mediated internet-supported interventions for a range of depressive and anxiety disorders, achieving similar outcomes to conventional face-to-face therapies (Griffiths et al., 2010).

Third, services may be improved by integrating treatment provided by different parts of the system. For example, there is a substantial life-expectancy gap between people with long-term, severe mental disorders and the general population (Lawrence et al., 2013). Integrating physical and mental health treatment for these people may significantly improve their overall health and life expectancy.