Improving the quality of global mental health care requires universal agreement on minimum national investment

Kilbourne et al1 provide an informative review of current theory and approaches to the measurement of quality of mental health services from a number of higher income countries across the world. A welcome emphasis is given to social outcomes, and the authors note that quality of life, personal recovery and community tenure are as relevant as more traditional outcomes such as symptoms and functioning.

The authors also acknowledge that any outcomes framework needs to take into account variables such as morbidity and socioeconomic factors, to avoid “cherry picking” and gaming by providers. This is an important point and one that is as relevant to social outcomes as “clinical” outcomes, but more difficult to adjust for. Quality of life is a notoriously slippery concept which has a complex relationship with relative expectations of what a “good” quality of life comprises2. Similarly, personal recovery is, by definition, a subjective concept and it is no surprise that the development of valid recovery measures has been hampered by a lack of consensus amongst providers, researchers and service users. This is unlikely to be solved by further investment in tool development.

Apart from the problems of actually measuring relevant outcomes in mental health systems, a major issue with many existing measurement based schemes is that they only focus on the simpler parts of the system. The three examples from mental health systems given by Kilbourne et al (the UK's Improving Access to Psychological Therapy, the Dutch Depression Initiative and the Australian TrueBlue model) are all primary care facing models that aim to address common mental disorders. These services deliver specific, time limited, evidence based interventions and are ideally suited to straightforward monitoring of their structures, processes and clinical outcomes. This has led to greater investment and their being embedded into national service models in the UK and the Netherlands.

However, people with more complex mental health problems tend to require multiple interventions from multiple services, often spanning statutory health, social care and non‐governmental organization providers. The problem of identifying standard, universal metrics and measures that can capture the impact of these complex arrangements in order to assess whether “quality” is being delivered has, unsurprisingly, proved insoluble, not least because social outcomes are often more relevant to this group than clinical outcomes such as symptoms.

The abandonment of the outcome based reimbursement system for mental health in the UK probably had more to do with this issue than with administrative burden or risk of gaming. The main clinician rated outcome measure that was under consideration, the Health of the National Outcome Scale (HoNOS)3, is one of the most widely used mental health outcome assessment tools worldwide, but there are concerns about its appropriateness and sensitivity to change for people with longer term and more complex mental health problems. As such, it cannot reliably indicate whether a service is providing effective care and should be reimbursed.

In Australia, universal routine outcome data (including HoNOS) have been collected systematically for around 20 years, but this has not stopped the gradual disinvestment in statutory mental health services for those with the most complex needs, and concerns are now being raised about the quality of care provided by other sectors for this group4.

A bigger issue, something of an elephant in the room, is that there is not such good evidence that improving quality of care actually leads to better clinical outcomes, particularly when we consider longer term, complex conditions. Evaluation of the impact of the national Quality and Outcomes Framework for diabetes care in the UK found no clear association with improved clinical outcomes over the three years before and after its introduction5.

Nevertheless, it would clearly be counterproductive not to attempt to understand how to organize services to be as safe, effective and efficient as possible. The difficulty in identifying robust universal measures for mental health services that can do this may explain why, as Kilbourne et al point out, most “outcome” measures are actually process measures. In complex systems such as these, it is much easier to describe what you are doing than to assess whether it has had an impact. Perhaps New Zealand has adopted the most pragmatic approach: to focus on monitoring key indicators that can be agreed on as universal markers of basic service quality, such as the minimization of seclusion and restraint, and suicide reduction6.

Indeed, the increased support for “pay‐for‐performance” or “activity” rather than “payment for results” models probably reflects a growing acceptance that there is no simple way to assess outcomes in most mental health services. Consequently, comparative benchmarking that uses various process metrics has become increasingly popular in England and Wales through the voluntary National Health Service benchmarking network. However, this can only work within a publicly funded system where sharing data does not potentially threaten an organization through competitive market forces.

Finally, the biggest issue (an even larger elephant) is resourcing. Across the world, most countries lack even basic mental health care. The nuances of different approaches to quality assessment in higher income countries pale into insignificance when considering the appalling consequences of this. Globally, most people with serious mental health problems are in long‐term institutions, often living in unacceptable, inhumane conditions7. Taylor Salisbury et al8 recently showed that, across Europe, the proportion of the national health budget spent on mental health was positively correlated with the quality of the country's longer term facilities.

It seems that adoption of a universal national minimum percentage investment in mental health care should be the first crucial step in any global quality improvement initiative.

Helen Killaspy
Rehabilitation Psychiatry, Division of Psychiatry, University College London, London, UK