What mediates ethnic density associations?

The findings summarised above represent a good example of the ‘relativity’ of relative risks, ^{Reference Neeleman8} as the increase in risk associated with ethnic minority status is contingent on social context – a phenomenon known as ecological effect modification. The fact that the strength of the relative risk for psychotic disorder varies inversely with the prevalence of the risk factor in question – minority status – may be indicative of selection (e.g. individuals of ethnic minority most at risk form a specific cluster geographically) or causation (e.g. higher levels of own-group density confer a risk-decreasing buffering effect). It is difficult to distinguish between these two mechanisms, as this requires extensive and sophisticated cross-context comparisons that are difficult to conduct – particularly across cultural groups – given that schizophrenia represents a rare and variably defined mental disorder that typically is identified on the basis of service contact data. In this issue of the Journal, Das-Munshi and colleagues present fascinating evidence that density associations may in fact represent a combination of risk-increasing and protective mechanisms, using a subclinical psychosis outcome (defined as one or more psychotic experiences measured with the Psychosis Screening Questionnaire – PSQ) that was independent of service use and not prone to cultural bias in diagnosis. ^{Reference Das-Munshi, Bécares, Boydell, Dewey, Morgan and Stansfeld9} Their findings, summarised in Fig. 1, suggest that lower own-group density may result in greater exposure to psychosis-inducing social adversity, and that although a risk-buffering effect of higher own-group density exists, this may be moderated by the level of strains and supportive relationships a person has.

FIG. 1 Results presented by Das-Munshi et al. ^{Reference Das-Munshi, Bécares, Boydell, Dewey, Morgan and Stansfeld9}

Three associations (A1–A3) and one interaction (Int1) were demonstrated across areas with different levels of own-group ethnic density. First, for a given individual who is a member of a minority ethnic group, the level of subclincial psychotic experiences was greater when living in a neighbourhood with lower own-group density, independent of social class, education and area deprivation (A1). Second, with decreasing own-group density, ethnic minority groups generally were more likely to report greater discrimination, poorer social support and more chronic strains (A2). Third, greater levels of racism, discrimination, chronic strains and difficulties, and lower level of social support, were associated with previous year psychotic experiences (A3). Finally, the protective effect of higher own-group density may be reduced in people experiencing chronic strains, and enhanced in people reporting high levels of practical or confiding/emotional support (Int1).

What outcome do psychotic experiences represent?

The findings presented by Das-Munshi et al differ from previous work in one important aspect, namely the outcome under study. Psychotic experiences as measured by the PSQ are not ‘schizophrenia light’. Rather, they represent a dimensional phenotype indexing aberrant attribution of salience that is prevalent in the general population (prevalence 8%), ^{Reference Linscott and Van Os3} substantially more prevalent in common mental disorder (anxiety and depression; prevalence 30%) ^{Reference Wigman, van Nierop, Vollebergh, Lieb, Beesdo-Baum and Wittchen10} and universal in psychotic disorder. In the general population, psychotic experiences predict an increase in risk for transition to psychotic and (to a lesser degree) non-psychotic disorder at an annual rate of 0.6%; ^{Reference Kaymaz, Drukker, Lieb, Wittchen, Werbeloff and Weiser11} in common mental disorder, they have a negative impact on course and outcome (see Appendix). ^{Reference Wigman, van Nierop, Vollebergh, Lieb, Beesdo-Baum and Wittchen10,Reference Perlis, Uher, Ostacher, Goldberg, Trivedi and Rush12} Thus, the increase in risk for psychotic experiences associated with lower own-group density in part may mediate the association with schizophrenia demonstrated in previous work. ^{Reference Boydell, Van Os, McKenzie, Allardyce, Goel and McCreadie6,Reference Veling, Susser, van Os, Mackenbach, Selten and Hoek7} However, Das-Munshi et al's findings also suggest that the impact of lower own-group density extends dimensionally to an extended psychosis phenotype of aberrant salience with relevance for (a) behavioural expression of risk in the general population and (b) course and outcome of common mental disorder. The findings therefore concur with recent meta-analytic work, for example in the area of childhood victimisation, ^{Reference Varese, Smeets, Drukker, Lieverse, Lataster and Viechtbauer1} that indicates that the environment may have an impact on dimensional expression of a specific domain of psychopathology that extends across traditional disorder boundaries.

Appendix