This systematic review and meta-analysis was aimed at comparing mental illness stigma associated with CMDs based on racial background. In line with our expectations, the results suggest that racial minorities have more mental illness stigma for CMDs when compared with racial majorities. Sensitivity analyses showed the robustness of these results. The multi-variate meta-regression analyses indicated that studies of poor quality had higher effect sizes than the studies with high quality.
Higher mental illness stigma in racial minorities is in line with the growing literature highlighting the variations in mental illness stigma based on ethnicity and/or race (11-12,18-22-23,57). This finding could be explained in relation to the social identity theory (58). In collectivistic cultures, group harmony and cohesion are of central importance (24). In this context, CMDs might be seen to fall outside of societal expectations (13,58-59) and this would precipitate shame and subsequent mental-illness stigma (58-59). Thus, what is defined as “them” not “us”, would reinforce the public opinion that individuals with CMDs are dangerous and must be segregated (60). Consistent with this notion, there is research evidence indicating that cultural beliefs of CMDs are related with the extent of the impact of mental illness stigma (i.e. fear of someone with CMDs and/or desire for social distance from an individual with CMDs) at individual and societal levels (24,60-63).
In extent, the consequences of mental illness stigma are more harmful especially when the preceding cultural and/or personal attributions coincide with social adversities such as migration, poverty, gender, diagnosis of a CMD, ethnic and/or sexual minority statuses (7-9,63-64). This phenomenon is known as the intersectional impact of stigma and it accounts for the underutilisation of mental health services among those who are affected (10,63-67). For instance, a large scale study from America (N=15,383) showed that, Black women with low socio-economic background reported more stigma related concerns and were less likely to utilise mental health services compared to locally born women from White backgrounds (8). Furthermore, the negative impact of the intersection of different types of stigma on utilisation of health services is well-documented in the HIV literature investigating (HIV-related stigma, sexism, racism, and homo/transphobia) on individuals` well-being (65-67).
Studies comparing public beliefs and attitudes regarding CMDs point to the importance of migration status and/or ethnicity in shaping mental illness stigma (23). For instance, in a large scale vignette study (N=2013) by Makowski and colleagues in Germany, there was an indication for the differences in mental illness stigma between locally born minorities, migrants and non-minorities (23). When the vignette concerned an individual from a migrant background with depression, participants from migrant background expressed greater stigmatizing attitudes (e.g. negative stereotypes, emotional reactions and desire for social distance) compared to the locally born minorities and majorities (23).
Strengths and limitations
This study has various innovations. This is the first meta-analysis in the stigma field utilizing a unified conceptual framework to pool the studies. This is an important step to advance the current understanding of the mental illness stigma and how racial/ethnic differences impact people`s experiences of mental illness stigma.
Based on the current results an important message for the public health field is to tailor the existing anti-stigma interventions according to the specific racial and/or ethnic groups (25). At present, recommended practices include psycho-educational campaigns aiming to improve the public knowledge on CMDs and individuals with CMDs (25,68). Such practices however, fail to demonstrate changes in attitudes among stigmatized and stigmatizer (69). Often, educational campaigns do not provide evidence for the effectiveness of anti-stigma campaigns in increasing help-seeking for CMDs (69). Our results suggest that the development and the implementation of such campaigns can be improved if the messages of these campaigns are adapted according to the socio-cultural and political contexts of the countries that the individuals live in (69-70).
Even though our results suggest that all racial minorities, regardless, have higher mental illness stigma, various limitations need to be taken into account whilst interpreting our findings. We found no variability in mental illness stigma based on specific races. This could be explained with the small number of studies representing each racial minority group in sub-group analyses. Further, we were also unable to investigate if there were differences in mental illness stigma between racial minorities based on the migration status since this was rarely investigated in the included studies (8,23). It may be that there are other potential moderators such as the degree of acculturation which are overlooked in the current stigma literature and could not be taken into consideration in our study.
Further, the multi-variate analyses pointed to the possible moderating role of the number of studies with high risk of bias on the effect size of stigma. It could be that if there were more studies with low risk of bias, we might have found no racial differences in stigma for CMDs. Given these reasons, the results of this study must be treated with caution.
The studies from North America were overrepresented in the current study. This limits our understanding of how contextual differences between continents influence the impact of mental illness stigma. For instance, the experiences of discrimination related with stigma might be different in Europe compared to America (23).
Related issues were the racial classification and the definition of minority and/or majority statuses in this study. Since there were many ethnic and religious groups in the included studies, we decided that the mutually exclusive category approach (28), re-allocating individuals in existing categories defined by the previous research (29), would be the most suitable. We have also defined minority and majority statuses based on the country of the included studies. Since the definition of race-ethnicity includes history, religion, language and socio-political dynamics (28,70), we oversimplified the term by using broad racial categories. Moreover, our definitions do not capture the changes in ethnicity and minority statuses over time depending on the changing socio-political circumstances (28). An alternative to the preceding is the multiple characteristics approach (28) taking into account the various aspects of ethnicity such as language, country of birth, nationality and religiosity (28). Even though the latter offers more effective approach to the measurement of ethnicity in detail, the former is more pragmatic and facilitates the comparability of the studies across countries (28). In the light of these caveats, our results are restricted with the broad racial categories which are defined at a specific point in time. It is recommended that ethnicity and minority statuses are measured at different time points in prospective studies (28).
Another limitation restricting the representativeness of the sample in our study was the exclusion of clinical populations that had comorbidities. One reason could be that we have excluded HIV, and other comorbidities and yet most clinical studies include people who often have CMDs and other conditions. Some studies indicate that diagnosis of a CMD would reinforce the label indicating that the individual falls out of the societal expectations and therefore is unpredictable and/or dangerous (8,52,71). Alternatively, diagnosis of a CMD might create opportunities for more psycho-education about CMDs, and more contact with the others with CMDs and this might reduce the mental illness stigma in return (71). Since clinical populations are not represented in our study, we were not able to assess if there was a variation in mental illness stigma among racial minorities depending on the presence of a formal diagnosis. Thus, the following questions are important and yet they remain to be answered by the future studies: Among those who have received a diagnosis of a CMD, are there any differences in the experience of mental illness stigma between ethnic minorities and majorities? Among ethnic minorities, are there differences in mental illness stigma between those who have been diagnosed with CMDs and those who have not?
Further, some types of mental-illness stigma are under investigated. There is an indication that stigma is associated with greater burden when it is accepted and internalized by those who have a mental illness (11). In our meta-analysis, many studies investigated stereotypes or prejudices attributed to those who have a mental illness. Conversely, how stigma was anticipated or experienced by ethnic minorities when compared with majorities were rarely investigated (20-21). Given this limitation, we could not investigate whether there was a variation in different perspectives (e.g. anticipated, internalised) and/or components (e.g. cognitive, affective, behavioural) of mental illness stigma between racial minorities.
It is worth to emphasize that there were considerable number of the questionnaires which were developed for the purpose of the studies by the authors themselves. In our sensitivity analyses, exclusion of these questionnaires and limiting the analyses with the validated questionnaires such as CAMI, reduced the heterogeneity. The revision of the existing self-report questionnaires by the authors was another limitation. Often, these questionnaires were revised according to the research question and/or the sample and their psychometric properties were not tested. Additionally, the self-report questionnaires often measured multiple stigma mechanisms within the same scale (11). This might conflate the results as stigma mechanisms might differently relate to the outcomes.
Poor quality of the vignette studies also worth mentioning. The vignettes were developed based on the DSM criteria and the authors` clinical information of the CMDs. Often, there was no information about the psychometric properties of the vignettes with one exception. Subramaniam and colleagues followed a systematic approach by revising the vignettes with clinical experts, piloting them with participants and revising the relevance and acceptability afterwards (23). Since the vignettes are often developed by the researchers themselves, it could be that they are more suggestive when compared with self-report measures and this would confound the results (see appendix B).
Our results are also limited with the MISF as other frameworks exist such as the Framework Integrating Normative Influences on Stigma (FINIS) (72). We have chosen the MISF as to our knowledge, it identifies specific outcome measures capturing specific stigma mechanism (11).