Background

Stigma towards mental illness is a worldwide phenomenon that leads to substantial psychological distress, inhibition of help seeking behavior, limited provision of treatment resources, and societal ostracism especially in low and middle income countries (LMICs) [1–5]. In many cases, individuals with mental health disorders are subjected to abuse, isolation, and discrimination [2,6,7]. In Uganda, only 15% of individuals with mental illnesses receive any mental health treatment [2]. This low rate of treatment adherence can be attributed to multiple factors, including economic constraints, stigma, and limited healthcare resources [8]. While it is noted that merely 1% of Uganda’s healthcare budget is dedicated to mental health services, this statistic is not an outlier, with 24.4% of countries out of 78 studied spending below 1% of health expenditure on mental health [9]. Uganda, like many other countries, suffers from a severe shortage of trained specialists and limited access to effective treatments [2,10].

Given the negative impacts of stigma on people with mental illness and their families, treatment availability and healthcare seeking behavior, understanding and reducing mental illness stigma is a public health priority [11,12]. However, despite considerably heterogeneous presentations of mental illness, studies of stigma often group different mental disorders together and design anti-stigma interventions towards people with mental illness in general, rather than for specific disorders [13]. Some research in high-income countries (HICs) has attempted to address this knowledge gap, showing that there is more stigma attached to schizophrenia compared to depression or anxiety, although there is significant variation across life domains and across countries [3,4,14,15].

Another facet of mental illness stigma that is poorly understood is the influence of gender on the stigma towards people with mental disorders. Research conducted in HICs has indicated that individuals with mental illnesses whose symptoms align with gender-based stereotypes are more susceptible to experiencing stigma [16]. For instance, a man with schizophrenia is portrayed as potentially violent and may encounter heightened stigmatization compared to a woman with the same condition [16]. A study examining mental health stigma and substance abuse disorders in the United States, which analyzed factors like gender, ethnicity, and education, found that women exhibit a lower tendency to endorse discriminatory behaviors towards individuals with psychiatric disorders when compared to men [6,17]. In India, a lower middle-income country with clearly demarcated gender roles, individuals with schizophrenia experienced stigma most strongly in relation to their failure to meet gendered norms–for example, men in regards to employment and women in relation to marriage and childbirth [18]. However, literature on the topic has been relatively scant, and no studies have assessed differential levels of mental illness stigma by gender in sub-Saharan Africa.

To address this gap in the literature, we investigated differences in the level of societal and public stigma across different mental disorders while considering the gender of the study participants as well as the gender of the people depicted in the vignettes. The four mental disorders addressed in this study include major depressive disorder, generalized anxiety disorder, alcohol use disorder, and schizophrenia, as these disorders account for some of the highest disability-adjusted life years among mental illnesses globally [19]. Findings from this study could add nuance to the understanding of mental health stigma in LMICs and suggest specific differences in focus for future anti-stigma interventions.

Methods

Results

Of the 380 participants approached, all were included in the study. Data from 1 participant was excluded from the analysis, due to missing data (Fig 1). Participants were identified and interviewed from July 2023 to September 2023. Across vignette groups, there were no significant differences in the distributions of sex, age, and other demographic factors, suggesting successful randomization (Table 1). Fig 2 displays the mean acceptance of PAS estimated across mental disorders and gender combinations by our random effects model accounting for village effects; Table 2 presents these data numerically. The corresponding BAS mean estimates are presented in Appendices C and D. After accounting for multiple testing and model-misspecification via permutation, significant differences in levels of stigma among the four conditions were found for PAS (Bonferroni-adjusted, empirical p-value: 0.00235) and BAS (Bonferroni-adjusted, empirical p-value 0.04757).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. CONSORT diagram showing flow of participants through each stage of the trial.

https://doi.org/10.1371/journal.pmen.0000069.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Personal acceptance scale by condition, gender of vignette, and sex of participant means and standard errors esimated under random effects model accounting for village effects.

Filled circles denote female participants hearing a female vignette. Hollow circles denote female participants hearing a male vingnette. Filled diamonds denote male participants hearing a male vignette. Hollow diamonds denote male participants hearing a female vignette. See Table 2 for a table of the means and standard errors plotted here.

https://doi.org/10.1371/journal.pmen.0000069.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Descriptive statistics.

https://doi.org/10.1371/journal.pmen.0000069.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Table of means for the personal acceptance scale by condition, gender of vignette, and sex of participant–means and standard errors estimated under random effects model accounting for village effects.

https://doi.org/10.1371/journal.pmen.0000069.t002

In our secondary pairwise analysis, we found that acceptance differed between anxiety and schizophrenia according to PAS (Bonferroni-adjusted, empirical p-value 0.008) and BAS (Bonferroni-adjusted, empirical p-value 0.007). Visual inspection of Figs 2 and S1 suggests that individuals with anxiety are more accepted than individuals with schizophrenia. None of the other pairwise comparisons reached statistical significance for PAS or BAS (Appendices E and F).

In our final secondary analysis examining differences in acceptance across gender combinations within mental disorders, we found that PAS varied across gender combinations for depression (Bonferroni-adjusted, empirical p-value 0.017). In the context of Fig 2, we interpret this result as a reflection of personal acceptance being higher for women with depression than men with depression. Significant gender effects were not observed for any of the other mental disorders on the PAS (S4 Table) or for any of the four disorders on the BAS (S5 Table).

Discussion

In this population-based, experimental vignette study conducted in a rural region of eastern Uganda, we found significant differences in levels of stigma attached to depression, anxiety, alcohol use disorder, and schizophrenia. The differences persisted after adjusting for the sex of the respondent and the sex of the individual depicted in the vignette. Schizophrenia was least accepted, or most highly stigmatized, among the four conditions. In pairwise comparisons after adjusting for multiple testing, people with anxiety disorder appeared to be more accepted than people with schizophrenia.

The finding that anxiety is less stigmatized than schizophrenia aligns with existing literature from HICs [30]. In HICs, research indicates that mental illnesses conforming to gender-based stereotypes are more likely to face stigma, although other studies have shown no association [16,17]. It is possible that with a greater sample size, our study would have been powered to detect differences in stigmatizing attitudes between depression and alcohol use disorder and between depression and schizophrenia, given that pairwise comparisons without the Bonferroni correction demonstrated significant differences. This would be in line with previous findings that alcohol use disorder is stigmatized at a similar level as schizophrenia, and that schizophrenia and alcohol use disorder are more stigmatized than other mental illnesses [34].

Our measurement of differential levels of stigma based on the specific mental disorder, adjusting for sex of the respondent and the sex of the individual depicted in the vignette, provided us with a unique ability to assess mental illness stigma in a general population sample from a low-income country. Our use of a general population sample is a strength given that community and family support are critical predictors of mental health outcomes, particularly in LMICs where mental healthcare may be more difficult to access and where formal safety nets may be more limited [1,35]. Most studies on mental illness stigma have been in HICs, whereas 80% of the burden of mental illness is in LMICs, and stigma appears to be an even stronger barrier to treatment access in low resource settings [36]. The Stigma in Global Context–Mental Health Study assessed stigmatizing attitudes toward people with depression and schizophrenia but did not include any low-income African countries [37].

The random assignment of exposure to vignettes strengthens our confidence that the observed differences were not the result of selection bias. The vignette approach towards describing mental illness has been utilized in some studies investigating mental illness stigma, but not universally [38]. Participants in the Stigma in Global Context–Mental Health Study were randomly assigned to vignettes about depression and schizophrenia [37]. Rasmussen et al. conducted an experimental vignette study with a population-based sample in Uganda to assess the extent to which portrayals of successful treatment could attenuate stigmatizing attitudes toward people with depression, schizophrenia, and bipolar disorder [21]. Our study further assesses stigmatizing attitudes toward people with generalized anxiety disorder and alcohol use disorder. Our vignettes were informed by the DSM-5 and modified by Ugandan experts to generate locally acceptable and understandable language and communication. A psychiatrist developed the criteria for the vignettes per DSM-5, and a local team developed the vignettes per the criteria in a manner that is relatable to the local audience.

Our data could suggest that mental disorders with behavioral manifestations deviating from gender norms, such as men with depression, face greater stigmatization. [Fig 2] This finding is potentially consistent with population-based epidemiological studies from Uganda suggesting that depression is much more prevalent among women than men [39,40]. These results could also align with previous studies demonstrating the widespread stigmatization of depression among men across different cultures, often attributed to perceptions of “weakness in character [41,42].” However, there needs to be further studies to investigate this hypothesis, as it was not the primary outcome of our study.

A limitation of the study is that although the vignettes were translated from Runyankore to Lusoga [21], modified by experts to ensure linguistic and cultural appropriateness to the Busogan context, and then back-translated to English to verify that they appropriately represented psychiatric concepts similar to how they would be understood in the U.S., such procedures can inadvertently affect the properties of the tools. Subtle shifts in meaning could influence respondent understanding and responses, possibly affecting the nature of the responses elicited. However, the stigma outcome scales have demonstrated good reliability and construct validity in the Busogan context [24,25] and elsewhere in Africa [22]. Other limitations of the study include that the study design relied upon responses to a questionnaire and was therefore vulnerable to social desirability bias. One might expect that social desirability bias affected all conditions similarly, but it is also possible that certain conditions were associated with greater social desirability bias than others and that this differential social desirability bias could have affected the findings. Observing behavior, instead of relying on a questionnaire, would have potentially addressed this problem. Additionally, the study sampled a limited geographical area and cannot be directly generalized to other areas.

Stigma surrounding mental illness continues to be a neglected issue in LMICs, with limited exploration into its nuanced aspects taking specific disorders and gender into consideration [43]. This study has contributed valuable insights by examining illness stigma in LMICs, specifically by dissecting stigma levels based on the gender of the individual affected, the participant’s sex, and the specific mental health condition involved. To ascertain the broader applicability of these findings, further research is essential in diverse LMIC contexts. Further research could also investigate how gender affects attitudes towards specific mental illnesses. The data from this study can serve as a foundation for targeted anti-stigma interventions, highlighting the importance of considering specific mental disorders and gender norms in shaping these stigmatizing attitudes.

Supporting information

Acknowledgments

The team is grateful for the contribution of Buyende District community health workers and the Empower Through Health team.