Unhealthy scepticism: Time for a fair go for best available evidence

Scepticism and innovation are complementary and equally important contributors to progress in society, and hence in mental health care. Innovation has an evil twin: dogma. Examples litter the history of psychiatry, from aspects of psychoanalysis to leucotomy and even in relation to key aspects of the process of de-institutionalisation and mainstreaming of psychiatric care. The antidote to dogma is the scientific method and an evidence-informed approach to change. The evil twin of scepticism is denialism, in which the scientific approach is corrupted or misused, and unjustified doubt deliberately created or exaggerated within a field of inquiry or reform, for reasons that are non-transparent (Manne, 2012; Oreskes and Conway, 2010). The antidote to denialism is greater transparency. Scepticism plays the ball, focusing objectively on the message and the evidence for and against. Denialism plays the man, taking deliberate aim at the messengers in order to destroy the message. Scepticism operates from a position of strength, denialism from one of implicitly acknowledged weakness. Hence the tactics employed.

From the early 1990s, the international early intervention field (and more recently its close relative the youth mental health field) has been led by academic clinicians who have demonstrated a collective commitment to an evidence-informed approach to mental health reform. The exponential growth in peer-reviewed publications in early psychosis, and more recently early intervention for other disorders, reflects this commitment, as does the record of eight large-scale international scientific meetings on early intervention held across the globe since 1996. We contend that few if any other substantial reform processes in mental health care have embraced or been informed by evidence to this extent. Evidence is rarely conclusive in health care reform, and hence critiques, which are an essential and healthy phenomenon, can always be mounted and should be welcomed. Opinions change at different rates and few reforms achieve perfect consensus. Critiques have always been respected and taken seriously by the international leadership of the early intervention field, who have responded promptly and conscientiously. Critical review helps us all to think clearly and make up our minds. An evidence base takes time to be assembled, and for the overall indicative pattern to become clear. A succession of impartial observers and policy makers around the world and in Australia have now acknowledged this pattern (e.g. the reports from the National Health and Hospitals Reform Commission in 2009 in Australia (National Health and Hospitals Reform Commission, 2009) and the recent Schizophrenia Commission in the UK (The Schizophrenia Commission, 2012) chaired by Professor Robin Murray) and have endorsed early intervention as a best buy in mental health reform, valued highly by consumers and families. The mainstream view, strongly supported by the community, is that action is needed. This growing international consensus and government action in many countries is not, as Raven (2013) implies, the result of special pleading. Change is rarely that simple and certainly not sustainable on such a basis. Need and evidence are better drivers. It is striking that no credible alternative is ever proposed to the early intervention models that are so relentlessly critiqued by a small number. Dr Raven claims to be worried about the opportunity costs, and alleges biased resource allocation, arguing that the precious and still modest funds allocated to early intervention could be better spent on other programs. She is completely silent on specifics and on the comparative evidence and cost effectiveness of any new alternatives, and on the opportunity cost of persisting with the status quo. Yet the impact of delay and inaction, and the status quo has actually been measured around the world, and young people and families are counting the cost. Those struggling with emerging psychosis would really like to know what alternatives would be of greater value than scaling up of specialised early psychosis programs.

Of course this does not mean that early intervention is the only area to invest in with mental health reform. And, indeed, the vast majority of the Australian Government’s 2011 investment in reform was allocated to other domains and stages of illness. However, while unmet need abounds, all these investments should also be as evidence-informed as possible, and held to the same high standards of proof as early intervention. Given the ubiquitous unmet need and the tight fiscal environment, we simply cannot afford to invest in non-solutions and programs that are not ‘shovel-ready’. Neutral observers may wish to contrast the evidence in support of the range of new programs, notably ‘Partners in Recovery’, which, with AUD$549m, attracted the lion’s share of the mental health reform budget in 2011, with the evidence and international experience underpinning models of early intervention, ‘housing first’ programs and assertive community treatment.

Turning to the specifics of Dr Raven’s piece calling into question the two cost-effectiveness studies of the Early Psychosis Prevention and Intervention Centre (EPPIC), it is crucial to note that it is not merely methodology that is questioned but also the integrity of the authors. Dr Raven asserts that the authors have sought to deceive and mislead the scientific and wider audiences. This may help to explain why she is so exclusively focused on the EPPIC studies, when there is a raft of international evidence pointing to the same essential conclusion. While there are always methodological issues with any study, there is international expert consensus that all the available economic evidence supports early intervention as a best buy (Mihalopoulos et al., 2012). In fact, Dr Raven’s only ‘new’ issue concerns her demonstrably false assertion that we have inaccurately represented the control group used in the effectiveness and cost-effectiveness studies of EPPIC (called the pre-EPPIC group) in the 2009 study (Mihalopoulos et al., 2009) with an intention to mislead. Her main argument is that as the control group received their initial episode of care in the Aubrey Lewis Unit (ALU) of Royal Park Hospital – a specialist inpatient unit for first-episode psychosis – they cannot be characterised as receiving care in mainstream mental health services. She alleges that the follow-up paper to the original cost-effectiveness analysis of EPPIC (Mihalopoulos et al., 1999) somehow tried to conceal this fact and then implies improper use of these data in relation to advice to government. The truth is, as we will show, that nothing has been concealed, and ironically the nature of the control group strengthens rather than undermines the conclusions drawn in favour of early intervention.

In the 2009 paper (Mihalopoulos et al., 2009), on the long-term follow-up of first-episode samples, we state clearly that the comparison group received ‘high-quality inpatient care but ‘treatment-as-usual’ community care’ (Introduction, p. 910) and ‘The historical controls received initial treatment for psychosis from a specialist inpatient research ward with a focus on early psychosis. Following hospital discharge, follow-up community-based care was provided by local generic community psychiatric services’ (Methods, p. 910). Hence, the 2009 paper describes on two occasions the precise nature of care received by the control group, which was very largely within generic mental health services, except for the first hospitalisation. The key difference in care being tested was that of the availability of early detection and of more intensive and streamed community-based care over the first 2 years post diagnosis. The fact that both received what was enhanced, specialist, first-episode inpatient care had the potential to cloud the issue, making it more difficult to identify any advantage of community-based early psychosis streaming either in outcome or cost. Could Dr Raven have missed this obvious point by a wider margin?

Dr Raven has pursued this issue intensely for some time. She has cold called and interrogated our research staff on this and a number of other issues linked to early intervention research. In February 2011, she contacted one of us (CM) by email, questioning this specific issue of labelling of the comparator group in the 2009 study. The facts outlined above were patiently explained and clarified by CM via email as follows:

This label best captured, in a very brief way (which is required in abstracts which have impossibly short word count requirements), the type of care received by this group during the study time frame. The group were initially admitted to the ALU at Royal Park Hospital for their first (usually brief) hospitalisation and then discharged to generic mental health services where they received their care during the study time horizon (whether it involved subsequent rehospitalisation or not). The EPPIC group in comparison received all their care for the first two years from the EPPIC service.

We also pointed out to her at the time that the similar initial inpatient care received by both groups actually weakened her critique of the results. We can add that both cost-effectiveness studies (1999 and 2009) found that the results were very robust to cost assumptions; for example, the 1999 evaluation reported that even if the costs of the control group were halved and the costs of the EPPIC group were doubled, EPPIC was still highly cost effective. These conclusions have since been buttressed by a number of other health economic studies in early psychosis; for example, Valmaggia et al. (2009), McCrone et al. (2010) and Hastrup et al (2013).

Dr Raven goes on to say she found the report on optimal early psychosis care commissioned by the National Advisory Council on Mental Health (NACMH) in 2009–10 ‘troubling’, and (without saying why) again alleges that this report was ‘misleading’. Orygen Youth Health Research Centre won a nationally competitive tender before being commissioned by the NACMH to conduct this seminal piece of work, which was purely expert and independent technical advice to policy makers. It is important to note that the current first author (PM) was never a member of the NACMH, so no conflict of interest existed. The consultation report and proposal of a detailed and optimal model of care for early psychosis was based on a systematic literature review, a distillation of the 20-year international experience in early psychosis, and a series of in-depth national and international consultations with experts in the field. The whole process was conducted within the probity requirements of the Department of Health and Ageing.

A decade ago, there was scepticism and robust debate, which did not cross the line into denialism, in which one of us (PM) readily participated, when the early intervention reforms were advancing across the UK. Interestingly, in Canada, Western Europe, Asia and some parts of the USA, where these reforms have also occurred, there has been rather less scepticism and, to our knowledge, no public questioning of the reputations of the researchers and clinicians working in this field. There has also been very extensive debate in 2011–12 in this Journal as national investment in early intervention belatedly became a reality in Australia. However, this is fast becoming reminiscent of ‘Groundhog Day’, and is rapidly losing its status as debate at all, with increasingly weak and idiosyncratic scientific arguments, and growing ad hominem and reputational themes. Having given the issue of early intervention a very extensive airing across multiple issues, perhaps the Journal should heed John McGrath’s advice (McGrath, 2012) and restore some balance by devoting equal scrutiny to other emerging issues such as activity-based funding, the erosion of community mental health care and non-evidence-based approaches in mental health care, including much of the status quo.