Ethnic differences in psychosis—Lay epidemiology explanations

Abstract Background Much research attention has been given to the high rates of psychosis diagnosed in the Black community. However, little has been heard about possible reasons for this from Black African and Caribbean mental health service users themselves. Aims To determine how Black African and Caribbean service users perceive and explain these apparent differences. Methods We conducted four focus groups between 2014 and 2015 with 35 participants from the Black African and Black Caribbean community in Lambeth and Southwark, South East London, diagnosed with a psychotic illness. Recruitment was through a local voluntary sector organization and other community contacts. Results Each group described an elevated risk of psychosis in their community and explanations followed the following themes, with increased rates due to: (a) an accumulation of stressors due to disadvantaged ethnic minority status, (b) further disadvantage due to inequitable experiences of mental health services, (c) an absence of community support and (d) a double stigma: as a result of external discrimination, due to ethnicity, and internal stigma about mental illness from within the Black community itself. Conclusions Black mental health service users attributed an elevated risk of psychosis in their community to an accumulation of stressors directly related to ethnic minority status.

In recent years, the number of studies on this theme has increased exponentially, partly as a result of the increasing availability of large datasets of psychiatric records in which ethnicity has been recorded. Research in this area is at a critical point where a number of theories have been proposed, along with some supporting evidence, but a convincing explanation has yet to be established. A range of possible causal factors have been proposed, including disadvantaged socio-economic status, repeated experiences of discrimination and racism, social isolation and lack of access to social capital. While supporting evidence has been given for each, studies often lack a clear overall theoretical framework to explain why migrant and ethnic minority groups are at an increased risk. [4][5][6][7] It has been argued that for research in this field to now progress we need to look in greater detail at the processes behind these differences and, to this end, research would therefore benefit from qualitative study designs. [8][9][10] In this way, phenomena that might otherwise be missed using administrative or survey data could be explored in much greater depth, as could the specific social and geographical contexts in which they occur.
We propose that one useful starting point would be the accounts of ethnic minority service users themselves, investigating those explanatory factors they think are most relevant. This would also have important secondary benefits; this is a politically sensitive topic, and the way findings have been disseminated has in the past met with a hostile reception, in part, because authors may have failed to recognize how their work might be interpreted by the Black community. [ 11,12 ] There is, however, a paucity of explanatory accounts from Black and minority ethnic minority service users. While there are some relevant studies, they have not approached mental health service users directly about their explanations for the increased rates of severe mental illness in their community. [ 9,13 ] We argue that it is important to understand how members of Black and minority ethnic (BME) groups themselves perceive issues of concern to them, not only to inform research, but also to better communicate the results of research and how we translate these into interventions. The value of lay epidemiology has already been demonstrated in the field of cardiovascular disease health promotion, where lay accounts of disease aetiology have helped inform how health promotion initiatives are communicated to patient groups. [ 14 ] Furthermore, lay accounts may be tied to clinical outcomes by shedding light on perceptions and attitudes to treatment [ 15 ] and narrowing the gulf between professional and service user discourses. [ 16,17

| Focus group procedure
Each participant was given an information sheet, and the purpose of the study and the implications of taking part were explained to them before obtaining consent to participate. The focus groups were chaired by a BME former mental health service user with extensive experience working with this client group in a therapeutic environment. PS introduced the project and provided an overview of the study aims at the beginning of each group and helped with their facilitation. PS took notes during the discussion, which then helped to facilitate the interpretation and analysis of the focus group transcripts.
Each group lasted between 60 and 90 minutes, and discussions were semi-structured with discussion topics and prompt questions outlined in the study topic guide. The study topic guide was intentionally broad to avoid framing the discussions according to pre-defined academic concepts, and a group facilitator was deliberately chosen who did not have an academic background. Also, as the groups progressed, the topic guide evolved to better reflect the priorities of participants. All meetings were audio-recorded, recorded with participants consent, and transcribed at which point any identifying details were removed.

| Data analysis
Data were synthesized through a comprehensive process of thematic analysis, aiming to elicit the salient themes arising from the focus group discussions and subsequent feedback. A hybrid approach of inductive and deductive analysis was used. [ 22 ] Transcripts and focus group notes were managed using NVivo. Data were coded through detailed reading by two researchers (MK, PS), in close collaboration with the other members of the research team (DA, EP), using an inductive approach.
The research team met regularly to reflect on the data and identify and discuss emerging themes in an iterative process. These themes were cross-checked across the focus group data and notes. An overarching conceptual framework was refined by the research team to enable the data to be synthesized to examine relationships between themes and develop explanatory accounts for the data until a thematic saturation point was achieved, whereby no new themes emerged.
Member checking of emergent findings was carried out by asking the participants, either by email or phone to comment on the accuracy of the themes. All participants were sent a copy of a brief outline of preliminary themes to comment on. Four participants responded at this stage of the analysis, and their feedback was incorporated in the analysis.

| RE SULTS
A range of views were expressed in the groups, and several prominent themes emerged. Some participants felt that mental health issues were an individual matter and did not agree that being in a particular group, because of ethnicity or any other characteristic, was relevant. The majority, however, felt that members of the Black community were more likely to be diagnosed with a psychotic illness and that this reflected a greater underlying risk. There was much discussion about how this in turn was related to wider disadvantages.
The overarching themes are presented below along with anonymous illustrative quotes.

| Within-and between-group comparisons
The characteristics of each focus group, including the way they were recruited, had a bearing on the study results (see Table 1). The first three groups were recruited in the same way, initially drawing on members of a Black mental health support group and also advertising to day centre attendees at a local MIND branch. We then encouraged those who were interested to invite friends and contacts. Group 1 comprised both Black Caribbean and Black African participants, and there were important differences in the accounts of both ethnic groups, for example around the question of stigma within their respective communities. We therefore decided to run subsequent groups for each ethnic group separately. Focus groups 1 (mixed) and 3 (Black Caribbean) were similar in size, and the tone of the discussion was also similar. Participants were highly engaged with the topic, and there were often heated exchanges of views.
For example, many felt strongly that racism and discrimination were central to explanations for increased rates of mental illness. Others, however, felt that views expressed were overly negative and hostile to the majority White population. This was a particular cause of friction for some participants in group 3. Focus group 2 (Black African) was much smaller and more cohesive, with participants tending to reinforce the points others in the group had made. This group were particularly concerned about the role stigma plays in the Black African community and how this can exacerbate the experience of mental distress. They placed less weight on discrimination and racism as contributory factors (compared to groups 1 and 3). Focus group 4 (Black African) were recruited through a local GP practice and were more divided than previous groups, with participants split between those who saw social factors, such as discrimination, as relevant and those who felt that mental health should be primarily an individual responsibility. The latter were therefore more inclined to look for individual explanations rather than examine factors operating at the level of the wider BME community, and therefore saw debates about ethnic differences as largely unhelpful and consequently engaged least with the research question.

| Theme 1: Accumulation of stressors
Much of the groups' explanatory framework centred on the accumulation of life stresses resulting from multiple social disadvantage, both directly and indirectly related to ethnicity: We're dealing with things like poverty, isolation, a lot of people are coming from other countries, asylum seekers … They have stress where they are, they've got stress when they come over here.

| Theme 2: Experience of discrimination within mental health services
While many thought the underlying risk of psychosis was greater in the Black community, it was also a commonly held view that they were more likely to be diagnosed:

| Theme 3: Absence of community support
The absence of a community support system was also highlighted as a particular risk factor for mental illness in the Black community.
For instance, where previously the extended family might have played an important role this was no longer seen to be the case: We just don't have support systems. I mean my par-

| Theme 4: Stigma within the Black community
In contrast, there was also the commonly expressed view that Black communities, particularly the African community, could be more prone to stigmatize people with mental health problems:

| D ISCUSS I ON
We set out to investigate Black mental health service users' views about and explanations for the elevated rates of psychosis reported among ethnic minority groups. This was perceived by almost all participants as a problem in their community, and a range of explanations were given. Most often this was attributed to an accumulation of stressors that were specific to members of the Black community.
Stresses attributed to socio-economic deprivation and discrimination were seen to be further compounded by inequitable treatment from mental health services and a tendency to over-diagnose ethnic minorities with a psychotic illness. The absence of community and support networks was also highlighted while some felt that stigma from within the Black community itself was an important factor.
Many of these accounts closely parallel explanations put forward by the academic community. Accumulated social adversity including unemployment, social isolation and poor education has been shown to contribute to the increased risk of psychosis among ethnic minority groups, [ 7 ] and it is the cumulative nature of these stressors that is most often attributed a causal role. [ 7,24,25 ] For many, the underlying, or fundamental, cause behind these experiences of adversity is racial discrimination [ 8,26,27 ] and this is as much a unifying theme in the academic literature as it was in the focus group accounts. The role that mental health services play, in terms of both poorer quality and more coercive treatment for ethnic minorities, is also well documented.

| Study strengths and limitations
There is now an extensive literature on patients' explanatory models, that is the way patients make sense of the illnesses they suffer in contrast to formal medical "scientific" explanations. However, this investigation was more concerned with how patients explained illness at a community or population level. The contrast here is between how an individual patient accounted for their illness and how a group of patients explains the patterning of illness across their community-hence the descriptor of "lay epidemiology" and the use of focus group methods to elicit these population perspectives.
Inevitably, some of the focus group discussion did relate to individual experiences but in the main all groups managed to give voice to their collective experience and discuss issues they saw as contributing to the increased rates of psychosis in their community. We also acknowledge a potential bias where coders, already familiar with the academic literature, might frame lay accounts according to already established themes. However, we made sure to use a second coder, from outside of the field of psychosis research, whose independent analysis arrived at a very similar set of themes to the first coder. At the analysis stage, we also involved respondents themselves, circulating summaries of the main themes and incorporating subsequent comments. It is, however, difficult to make a claim for these groups to be truly representative as participants were selfselected having been motivated to participate in the first place.
The recruitment strategy used will also have influenced the kind of responses received. Given the well-known difficulties recruiting mental health service users, especially from BME communities [ 40,41 ] we decided to take a pragmatic approach by recruiting participants through a local voluntary sector organization with strong links with the target population. To test whether links with this one organization affected the views expressed, we also recruited one group (focus group 4) from mental health users attending a GP practice. It is notable that this was the least cohesive of the four groups and also the least likely to engage with the research question.
While we were restricted to English speaking participants only for this study, we were able to achieve a balance of Black Caribbean and Black African participants, broadly representative by age and gender. However, we acknowledge that this is a small exploratory study of specific groups of people at a specific time. Despite this, we argue participants were able to speak on behalf of their community to convey something of the universality of their experience and, in this way, contribute to our understanding of ethnic differences in mental health.

| CON CLUS ION
Epidemiological research identifies risk factors for mental illness; many of these factors, such as socio-economic status or gender or age, that seem causally related to mental illness, might not be recognized by patients themselves. Lay epidemiology attempts to explore how lay people account for the experience of illness in their community. As such, it tends to capture reasons rather than causes but it has the potential to reveal the processes by which mental illness comes about. This qualitative study of lay accounts has offered some support for the epidemiologist's risk factors.
Deprivation, here presented in a richer and more nuanced way, is clearly recognized as a significant factor in both determining and maintaining mental illness in the Black community. But what is particularly salient, and often missing from epidemiological accounts, is the experience of stigma internal to the Black community. These Black mental health service users report stigma as operating indirectly in exacerbating their state of deprivation but also directly through the reactions of others to their behaviour. On the one hand, these "others" are the mental health services that for some-though not all-continue to show institutional racism, but on the other hand the Black community itself can be a source of stigma given its understanding of mental illness. These lay epidemiology accounts therefore provide insights into questions of aetiology that may otherwise be missed by more conventional risk factor models.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflicts of interest.

E TH I C A L A PPROVA L
The study received ethical approval from the Biomedical Sciences,

Dentistry, Medicine and Natural and Mathematical Sciences
Research Ethics Subcommittee (BDM RESC), Kings College London (Reference BDM/14 /15-21). Participants provided written informed consent before taking part in the focus groups.

DATA AVA I L A B I L I T Y
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to ethical restrictions.