We are grateful to the commentators for their constructive observations on our review. We agree with Kwame McKenzie (Reference McKenzie2009) that consensus needs to be built; the key point we attempted to make is that, to gain such a consensus, the problem of high rates of psychosis in migrant and minority ethnic populations needs to be de-coupled from the no less important issue of service provision for minority ethnic patients. In the same way that improving customer services for insurance claimants following an accident is irrelevant to reducing the rate at which such accidents occur, so reforming mental health services (important as this no doubt is) will have no impact on population rates of disorder.

There are a number of points raised in the commentaries that we would like to address further. We were careful not to go beyond the available evidence in making the case that cumulative ‘social adversity’ across the life course is important in explaining the high rates. In noting these ‘adversities’ we were as specific (and vague) as the evidence allows. Here, we agree with March et al. (Reference March, Hatch and Susser2009) and Selten & Cantor-Graae (Reference Selten and Cantor-Graae2009) that future research needs to be significantly more sophisticated in how it captures the social structures and lived experiences that impact, over the course of development, on risk of psychosis. In this, socioeconomic disadvantage, usually based on some measure of ‘social class’, is no more of a conceptual advance than social adversity. What the evidence currently suggests is that a range of contexts and experiences are likely to be relevant (from hostile neighbourhood environments to childhood trauma to experiences of discrimination, and so on). What March et al. rightly allude to is the complexities involved and the inordinate methodological challenges that we face in disentangling these.

From this, much as ‘social defeat’ is a succinct and appealing hypothesis, we think it doubtful that the range of adversities that appear relevant over the life course can be so readily collapsed into a single exposure. In humans, the kinds of experiences that might comprise social defeat (entrapment, loss, humiliation) have been more consistently linked with depression (Harris, Reference Harris2001) and indeed the ‘social defeat’ paradigm was originally developed as a model of depression (Bjorkqvist, Reference Bjorkqvist2001). In contrast, exposure to intrusive and anxiety-provoking events and contexts may be particularly relevant to psychosis, i.e. not a resulting state of defeat and learned helplessness, but a state of heightened sensitivity to the external environment and its perceived hazards. On this, there are indeed studies in humans that suggest the dopaminergic system is altered in response to adverse and stressful early environments (De Bellis, Reference De Bellis, Lefter, Trickett and Putnam1994; Pruessner, Reference Pruessner, Champagne, Meaney and Dagher2004). That this is the case further supports the proposition that a much broader range of experiences are potentially relevant – it is not only ‘social defeat’ that has the potential to sensitize the dopaminergic system. This reminds us that, while animal research can provide clues, we need to be cautious in borrowing terminology and applying findings to humans. There is much that can be lost in translation.

All of this said, as much as greater sophistication in research will further our understanding, we need not sit back and wait for this to accrue before advocating change at policy and service delivery levels. As Brian Cooper (Cooper, Reference Cooper1992) commented nearly 20 years ago in relation to the same issue: ‘the history of public-health epidemiology, from cholera to bronchial carcinoma, has repeatedly demonstrated that effective preventive measures can precede the full causal elucidation of a disease’ (p. 597).

Of course, achieving significant policy change that may ameliorate the impact of social disadvantage, particularly in high-risk groups, is extremely difficult. But we do not agree, as Selten & Cantor-Graae suggest, that this is primarily a job for politicians. Psychiatry has a role – perhaps even a moral responsibility – in advocating for, and contributing where possible to, the implementation of change. On a daily basis, psychiatry encounters the consequences of social inequalities, trauma, and negative life experiences in the private miseries for which patients and their relatives seek help. As we have seen in the UK recently, left to themselves politicians are as likely to tend their homes and gardens at the public's expense as they are to take seriously the needs of those they purport to represent. It is a failure for us all that, despite over 40 years of evidence, the high rates of psychosis in the Black Caribbean and other migrant populations remain of no concern to our governments.