Pickett & Wilkinson's paper is the latest in a series of persuasive publications on income inequality and health stretching back to the 1990s. ^{Reference Pickett and Wilkinson1} However, in relation to what they variously call ‘mental illness’, ‘mental health problems’ and ‘distress’, I wonder whether they are taking at face value the highly inflated prevalence findings they cite. They write that ‘one million British children – one in ten between the ages of 5 and 16 – are mentally ill’ and that ‘one in four adults have been mentally ill in the past year’ in both the USA and the UK. These mostly represent the so-called common mental disorders. These figures are preposterous – as much to the citizen as to the psychiatrist in me – and make an urgent case for the profession to go back to the drawing board to revisit the core question: just what exactly do we mean by a ‘mental disorder’? Quantitative surveys tend to recast the epiphenomenal features of situational distress as free-standing disorder, so cannot but recruit false positives on a systematic scale. ^{Reference Summerfield and Rosen2}

The authors have plotted 12 countries onto a graph of ‘Percentage with any mental illness’ v. ‘Income inequality’, and it seems noteworthy that the strongest correlations are, successively, USA, UK, Australia, New Zealand and Canada. Below is France and, much lower down, The Netherlands, Belgium, Spain, Germany, Japan and Italy. I wonder why there is this apparent split between English-speaking countries and the rest, and whether this reflects particularities in the Anglo-American world, both in psychiatric culture and in trends towards the psychiatrisation of everyday life, which may be less advanced elsewhere. Taking account of possible skews of this kind would be likely to make the graph line rather less steep. Could the authors comment?

The authors acknowledge that a possible confounder in comparing rates of mental illness between one society and another lies in differing recognition and interpretation of survey questions, but go on to make the point that ‘at least the same diagnostic interviews are used in each country’. I am afraid this is to restate the problem, which is one of validity, rather than to resolve it. Valid research methods must reflect the ‘nature of reality’ for participants and a standard questionnaire used across heterogeneous societies cannot do this.

Pickett & Wilkinson conclude that if the UK is to reverse the massive rise in inequality experienced during the 1980s, ‘we need to encourage all mechanisms that help to reduce income differences’. But are we not all stuck with an intractable feature of late capitalism, its structural tendency to stratify incomes rather than to level them out?