The emergence of diagnostic criteria for mental disorders led to the development in the 1980s of standardised diagnostic interviews that could be administered by lay interviewers. When these interviews were used in community prevalence studies, they showed mismatches between diagnostic status and self-reported use of services. Much attention has been given to the group that met diagnostic criteria, but were not receiving services. These individuals were regarded as a having an ‘unmet need’ for treatment (Andrews, Reference Andrews, Andrews and Henderson2000). At the population level, these individuals were part of the ‘treatment gap’, which has been defined as ‘the absolute difference between the true prevalence of a disorder and the treated proportion of individuals affected by the disorder’ (Kohn et al. Reference Kohn, Saxena, Levav and Saraceno2004).

The flip-side of ‘unmet need’ is individuals who receive services but do not meet criteria for a diagnosis. This group has received far less research attention, but has been described as having ‘met un-need’ (Andrews, Reference Andrews, Andrews and Henderson2000). An obvious strategy for reducing the treatment gap would be to try to reallocate treatment resources away from people with ‘met un-need’ and towards those with ‘unmet need’.

The paper by Bobevski et al. (Reference Bobevski, Rosen and Meadows2017, this issue) explores the issue of ‘met un-need’ using data from the 2007 Australian National Survey of Mental Health and Wellbeing. They found that 43% of mental health service users did not fit criteria for a diagnosis in the previous 12 months. On the surface, this finding indicates substantial ‘met un-need’. However, when other indicators of need were included, such as a lifetime diagnosis, a lifetime suicide attempt, current psychological distress and disability, most of these individuals were found to have some indicator of need. Only 3.5% were found to have used mental health services but had no diagnosis and no need indicators, indicating that the problem of ‘met un-need’ is small. These findings are consistent with an independent analysis by another research group using the same survey data (Harris et al. Reference Harris, Diminic, Burgess, Carstensen, Stewart, Pirkis and Whiteford2014).

What Bobevski et al. (Reference Bobevski, Rosen and Meadows2017, this issue) have essentially done is to convert a categorical indicator of need (meeting diagnostic criteria in the past 12 months) to a more graded indicator. They classified participants into five groups in a hierarchy of need as follows:

1. 1. 12-month psychotic symptoms.

2. 2. Diagnosis of a 12-month disorder.

3. 3. Diagnosis of a lifetime disorder without 12-month symptoms.

4. 4. Possible need indicators in the absence of a 12-month and lifetime diagnosis.

5. 5. No indicator of need.

When mental health problems are viewed in this way, there is a corresponding gradient of service use, with use being highest in categories 1 and 2, intermediate in categories 3 and 4 and low in category 5.

The graded approach used by Bobevski et al. (Reference Bobevski, Rosen and Meadows2017, this issue) implicitly calls into question the use of a categorical diagnostic classification to indicate need for services and proposes instead what looks more like a continuum. Indeed, the evidence from taxometric studies supports a conceptualisation of most mental health problems as dimensional rather than categorical (Haslam et al. Reference Haslam, Holland and Kuppens2012). As I have pointed out elsewhere, the concepts of ‘unmet need’ and ‘met un-need’ are in part the result of the imposition of categorical diagnoses on to underlying continua of symptoms and disability (Jorm, Reference Jorm2006). Take as an example, a diagnosis of Major Depressive Disorder by DSM-5, which requires the presence of 5+ depressive symptoms over the previous 2 weeks. If a person has 5 symptoms and does not receive services then they could be seen as an example of ‘unmet need’. On the other hand, if they have four symptoms and do receive services, then they could be regarded as having ‘met un-need’. However, the only difference is that they fall on different sides of a somewhat arbitrary cut-off point on symptom count. In principle, it would be possible to reduce or increase the treatment gap in a population simply by moving the cut-off points used in diagnostic criteria for mental disorders.

While a categorical approach to diagnosis can be useful for clinical purposes, it is not the best way to think about the mental health intervention needs of a population. As Rose (Reference Rose1993) pointed out many years ago, we need to stop thinking in terms of reducing psychiatric caseness and instead think about shifting the distribution of mental health problems in the population as a whole. The dominance of categorical thinking when studying the mental health of populations has led to a focus on reducing the treatment gap through the expansion of treatment services (Chisholm et al. Reference Chisholm, Sweeny, Sheehan, Rasmussen, Smit, Cuijpers and Saxena2016). However, this strategy has been a failure so far. Despite substantial increases in the provision of treatment in a number of high-income countries in recent decades, there has not been any improvement in population mental health (Jorm et al. Reference Jorm, Patten, Brugha and Mojtabai2017).

To move distributions of mental health problems in the population may require consideration of a broader range of intervention than treatment services. Clinical services have an important role to play for people with severe, persistent and complex mental health problems, but we need to complement these with interventions appropriate to the rest of the continuum, including prevention, promotion of self-help, e-mental health and mental health first aid. In 2014, Australia's National Mental Health Commission carried out a review of the nation's mental health services (National Mental Health Commission, 2014). This review clearly recognised the continuum of need and proposed reallocation of resources accordingly. However, this vision remains a work in progress.