Mental health literacy of schizophrenia: a community sample representative Open Access

Mental health literacy is the knowledge about symptoms, causes, treatment and prevention of mental disorders related to their recognition, management and prevention (Altuncu et al., 2023; Doll et al., 2022; Jorm et al., 1997). Kometsi (2016) highlights that there will be various conceptualisations about mental illness among different cultures and that mental health literacy should be defined as a spectrum of cerebral and social skills that promote mental health. For example, the belief that mental illness results from demonic possession or withdrawal of the ancestors’ protection from an individual exists among many Indigenous Africans (Lloyd and Hutchinson, 2022; Lloyd and Panagopoulos, 2023; Maah, 2023). Maah (2023) views schizophrenia as a complex disease and further describes the complexity as related to the assumption that its causes are conceptualised differently across cultures, further adopting a double-sided conceptualisation, both psychiatric and socio-cultural. While in South Africa, schizophrenia was conceptualised in psychological and social terms and using supernatural descriptions such as bewitchment and ukuthwasa (an isiXhosa word for psychological disruption on the path to becoming an Indigenous healer) (Kometsi et al., 2020; van der Zeijst et al., 2022).

One important factor influencing mental health stigma is how mental illness is conceptualised (Schomerus et al., 2016). Gangi (2021) emphasises the importance of fundamentally understanding the rudiments and intricacies of mental illness stigma, which may be a major deterrent to seeking and or adhering to treatment for mental health problems (Perlick, 2014). Like other aspects of mental health literacy, knowledge and beliefs about the causes of mental illness have an association with different demographics.

Working in a South African community, Kometsi (2016) found that there were no differences between genders in terms of aetiological beliefs about schizophrenia, with both genders believing that psychological issues were the cause of the condition, while among the Lebanese population, females showed better understanding than males (Tantawi et al., 2024). Farrer et al. (2008) in a study in Australia, like the Lebanese population (Tantawi et al., 2024) found that young adults have better knowledge about schizophrenia than older participants. While these Lebanese participants with primary school education also had lower knowledge levels than those with higher education degrees.

The current study attempted to assess the communities’ conceptualisation of schizophrenia and the association of these with their demographic variables.

We conducted this study in the North West Province of South Africa, one of the poorer provinces in a very unequal middle-income country. Almost no research has been conducted on serious mental disorders in this province, with the exception being the work of Brooke-Sumner et al. (2015). There is no data on the epidemiology of serious mental disorders in the region.

The current study attempted to assess the individuals’ conceptualisation of schizophrenia and the association of these with their demographic variables.

The overarching research question of the study, adopted from the five components of mental health literacy, is as follows:

The aim of this study was to investigate mental health literacy among a community sample of Ngaka Modiri Molema District Municipality in the North West province, with a particular focus on the conceptualisation of schizophrenia. This study focused on the following objective:

This objective was investigated by compiling two fictitious vignettes (one male and one female character who meets the diagnostic criteria for schizophrenia). The vignettes used in this study did not present a full psychiatric history of the characters but only a scenario depicting salient features relevant to making a diagnosis. The materials were all designed and presented in both English and Setswana, the language spoken in the region, and by the researchers with specific questions regarding their conceptualisations of schizophrenia. Where participants were insufficiently literate to read the vignettes, these were read to them by researchers.

The research objective was addressed through a quantitative approach using a non-experimental and correlational research design (Stangor, 2015). A quantitative approach relies on the ontological positions and epistemological framework of positivism. A correlational design seeks to find relationships between two or more variables (Edmonds and Kennedy, 2017). In the current study, a correlational design was used to determine whether any relationship existed and the strength and direction of the relationship between variables (Yilmaz, 2013).

The study was conducted in the Ngaka Modiri Molema District Municipality in the North West province of South Africa, and the sample was drawn from two towns, Ditsobotla and Mahikeng local municipalities. The population size of Ditsobotla and Mahikeng local municipalities combined consists of a population of 496,259 (Municipalities of South Africa, 2016a, 2016b). Based on the population size and using a confidence level of 95% and a margin of error of 5%, a number of 384 participants was arrived at with the aid of a sample calculator. This was then divided into two parts to account for the two municipalities. To recruit a sample size of 192 participants, both Ditsobotla and Mahikeng, a randomised communal recruitment strategy was adopted using a sample size calculator. Data were collected between January and June 2022. Participants in this study were randomly recruited in communal spaces with large numbers of people, such as taxi ranks and bus stations, on township streets and in their homes. Only Black African participants over the age of 18 were included in the study. No participants were excluded based on their education level, gender, financial situation or marital status. A combination of systematic and convenient sampling as integers was used with the aid of a computer program called randomizer to select participants. On each day of data collection, a randomizer was calculated to create new sets of unique integers, with each integer having a different value (Saghaei, 2004). When the researchers went into the field, they counted whoever they met, and once they reached the integer number, that person was used as a participant.

Ethical clearance for this study was sought from the Research Ethics Committee (HSD2020/1585/21). Written permission was also obtained from Ngaka Modiri Molema District Municipality to conduct the study in the local municipalities. The participants were informed about the aim of the study, the estimated time required to fill out the mental health literacy questionnaire, the right to withdraw from the study and other ethical considerations as they appeared on the consent form. Consent to take part in the study was totally voluntary, and participants signed a consent form prior to participation. It took about 20 to 30 min to complete the questionnaires. The researcher was always available to answer and clarify participants’ questions.

Another ethical principle followed in this study was open co-operation. In this study, no form of deception was used with participants. Participants were also informed that they should not write identifiable information on the questionnaire to ensure anonymity. As soon as the data had been captured, all the hard copies of the questionnaires were archived for a period of five years in a locked cabinet in the supervisor’s office. In addition, the electronic data were password-protected on the researcher’s personal laptop. As a precaution, electronic data was also stored on a password-protected external hard drive in the event of theft.

The biographical variables involved in the study were age, gender, marital status, level of education, religious affiliation and home language. Apart from age, all the other biographical variables were measured on a nominal scale. The average age of the participants was 31.06 years with a standard deviation of 9.4 years. Their ages ranged from 18 to 62 years. Information on the variables that were measured on the nominal scale is presented in Table 1.

Of the total group, for the male vignette, 62.1% was female and 37.9% was male, while for the female vignette, 70.8% was female and 29.2% was male. Concerning age, most participants (65.8%) were between the ages of 18 and 34. Regarding marital status, the majority were single for both vignettes, at 72.6% for the male vignette and 77.1% for the female vignette. Furthermore, most of the population, 66.3%, had tertiary education for the male vignette and 70.8% for the female vignette. In addition, concerning religion, most of the population reported being Christian – 81.1% for the male vignette and 92.7% for the female vignette. Finally, the language spoken in the area was Setswana, with 85.3% for the male vignette and 72.9% for the female vignette.

The instrument used to collect data was the shorter version of The Attitudes and Beliefs about Mental Health Problems: Professional and Public Views Questionnaire (Jorm et al., 1999). The questionnaire comprised three sections. The first section encompassed information explaining the survey and the questions asking participants to consent to take part in the study. This section also required simple demographic information such as age, gender, marital status, education level, religion and language spoken at home. The second section of the questionnaire contained the two vignettes which depicted a male or female character who meets the diagnostic criteria for schizophrenia contained in the Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5; American Psychiatric Association, 2013). Both vignettes contained information on the onset, duration, frequency, progression and acuteness of the individual behaviours displayed by each person per vignette. In addition, the vignettes included the impact(s) of the mental illness on the character of the vignette. The third part of the questionnaire contained questions to elicit an understanding of attitudes towards and perceptions of mental illness. The first questions were open-ended, allowing participants to freely express their opinions and use personal phrases, which allowed the researcher to gain an empathetic understanding and comprehensive information from their answers. Most of the questions in the questionnaire consisted of Likert-type scales, and some required a “yes” or “no” response. In summary, these questions captured participants’ perceptions of the aetiology of mental illness, their thoughts regarding various professionals and helpful treatments, the likelihood of the characters’ recovery, social distance and attitudes towards people with mental illness, likely sources of knowledge about mental illness and knowledge of available treatment facilities. Kometsi’s (2016) study, drawing on a sample of South African adults, used a Pearson correlation coefficient to test the questionnaire with a result of 0.85, indicating the questionnaire’s reliability. However, for the current study, the Pearson correlation coefficient was not calculated because not all the items in the measuring instrument were used. English questionnaires of the male and female vignettes depicting a diagnosis of schizophrenia and the consent forms were translated into Setswana, adhering to the translation guidelines proposed by Rosnow and Rosenthal (1996, as cited in Kometsi et al., 2020). This accommodated participants who would have liked to complete the questionnaire in their mother tongue. As pointed out by Tsotetsi et al. (2021), Setswana is the preferred language of translation as the North West is dominated by Tswana-speaking people. However, no participants answered the translated questionnaire in this study.

A pilot study was conducted to test the suitability of the research instrument for this study. Stopher (2012) suggests that the sample size of a pilot should be 3% to 7% of the main sample of the study. Accordingly, a total of 17 participants were randomly selected in Matlosana Municipality to test the appropriateness of the measure for this study. The purpose of the pilot study and issues relating to consent were explained to the participants; after that, the questionnaires were distributed to them. The participants were able to understand and respond to the questionnaire without any constraints. This indicated that the questionnaire was suitable for the present study.

The pilot sample of 17 people (8.8% of the main study), of whom 41.1% were male and 58.9% female, was recruited in Klerksdorp, a town of 186,515 people in North West province. The Black African population comprised 74.0%, with Setswana being the predominant first language at 43.0% (Klerksdorp, 2022). Klerksdorp is in Dr Kenneth Kaunda District, part of the City of Matlosana Local Municipality. The town was chosen because the researcher believed the population was like that of the Ditsobotla and Mahikeng municipalities, where the main study occurred. The steps used to recruit the pilot sample were the same as those used to recruit the sample for the main study. However, the data collected during the pilot study were not included in the results of this study, as the aim of the pilot study was merely to test the data collection methods.

The IBM SPSS Statistics (Version 28.0) was used to analyse the results of the study. The chi-squared test of homogeneity (Howell, 2010) was used to investigate the three formulated research objectives because all the items were measured on a nominal scale.

Before a chi-squared test may be applied, it is important to take note of two assumptions, namely, the assumption of independence and the inclusion of non-occurrences. The first assumption refers to the independence of observations, which means that one participant’s choice has no effect on another participant’s choice. The second assumption of the inclusion of non-occurrences refers to the fact that the responses of all participants (positive or negative) should be included in the analysis. Both these assumptions were met before the analysis started.

Both the 1% and 5% levels of significance were used in the analysis. The effect sizes were also calculated for those findings with statistically significant values to determine whether the statistical findings were of practical significance. In the case of the chi-square test, the effect sizes were indicated as w and the following guidelines were used to interpret the effect sizes: 0.1 = small, 0.3 = medium and 0.5 = large. Results with medium to large effect sizes (2.5 or higher) will be discussed in more detail.

It is important to note that when assessing participants’ conceptions of mental illness, the focus was not on the correct usage of diagnostic labels but on their conceptualisations of mental disorders. Participants were asked what they would say is wrong with or is happening to someone depicted in any of the two vignettes. Their responses varied too greatly to permit meaningful statistical analysis and were therefore assigned to five different categories:

• medical (e.g. chemical imbalance in the brain, generic or inherited problems);

• psychological (e.g. own bad character, stressful circumstances in his life);

• social (e.g. the way he was raised); and

• traditional (e.g. God’s will, failure to perform certain cultural rituals, ancestral anger, evil spirits/sorcery, punishment of sins he committed).

Table 2 presents a list of these examples.

As indicated in the introduction, two vignettes (one of a male and another of a female character who meets the diagnostic criteria for schizophrenia) were compiled and presented to the participants with specific questions regarding the conceptualisation of the specific mental health problem. To investigate the research objective, the researcher decided to keep participants’ responses separate for the male and female schizophrenia vignettes to investigate whether differences could be identified when the person is of a different sex.

The study’s objective was to investigate the impact of demographic variables on participants’ conceptualisation of schizophrenia for a male and a female vignette. The chi-square test for homogeneity was used. The participant’s conceptualisation of schizophrenia will be discussed.

The relationships between the independent variables and the participants’ conceptualisations of a male who was diagnosed with schizophrenia are presented in Table 3. The following question with 10 different items was put to the participants: Q8: “In your opinion, how likely is it that the situation of a person who was diagnosed with schizophrenia might have been caused by the following” (see items in Table 3). Participants answered this question by referring to the 10 different items using the following scale: Scale 1 = Very likely; 2 = Don’t know; 3 = Less likely. The results are presented in Table 3.

According to Table 3, no statistically significant differences could be identified for two independent variables, namely, gender and level of education. This is surprising, given the association often mentioned in the literature between education level and attribution of biological causes. However, with reference to age, statistically significant differences, on the 1% and 5% levels, in proportions were found for two of the items [Item 6: God’s will; p = 0.003 (older participants more likely to endorse this); w = 0.35 and Item 9: Evil spirits/sorcery; p = 0.015; w = 0.30]. The corresponding effect sizes also indicate that the results are of medium practical significance and Table 4 indicates the frequencies and row percentages for Item 6 for the two age groups (≤30 and ≥31 years).

From Table 4, in comparison with the older participants (17.0%), a much greater proportion of the younger participants (50.0%) indicated that they did not know if the condition of a male diagnosed with schizophrenia was the result of God’s will. A greater proportion of the older participants (57.4%), in comparison to the younger participants (35.4%), believed it was less likely that the condition of a male diagnosed with schizophrenia could be attributed to God’s will. Table 5 indicates the frequencies and row percentages for Item 9 for the two age groups (≤30 and ≥31 years).

Table 5 indicates that, in comparison with the younger participants (14.6%), a greater proportion of older participants (40.4%) believed it is very likely that the condition of a male diagnosed with schizophrenia could be attributed to evil spirits/sorcery. As in the case of Table 4, a much greater proportion of the younger participants (50.0%) than the older participants (29.8%) indicated that they did not know if the condition of a male diagnosed with schizophrenia could be attributed to evil spirits/sorcery.

In Table 6, the relationships between the independent variables and the participants’ conceptualisations of a female who was diagnosed with schizophrenia are presented.

Question 8, with 10 items, was used, and the results are presented in Table 6.

According to Table 6, no statistically significant differences were identified for two of the independent variables, namely, the gender and age of the participants. However, regarding level of education, statistically significant differences, on the 5% level, in proportions were found for three of the items (Item 2: Chemical imbalance in the brain; p = 0.040; w = 0.26; Item 5: Generic or inherited problems; p = 0.029; w = 0.27 and Item 9: Evil spirits/sorcery; p = 0.049; w = 0.26). The corresponding effect sizes also indicate that the results are of practical significance. Table 7 indicates the frequencies and row percentages for Item 2 for the two levels of education groups (primary/secondary school and tertiary).

From Table 7, in comparison with the participants with tertiary education (44.1%), a much greater proportion of participants with primary/secondary school education (71.4%) believed it is very likely that the condition of a female person diagnosed with schizophrenia could be attributed to a chemical imbalance in the brain. Table 8 indicates the frequencies and row percentages for Item 5 for the two levels of education groups (primary/secondary school and tertiary).

Table 8 indicates that in comparison with the participants with tertiary education (32.4%), a much greater proportion of participants with primary/secondary school education (60.7%) believed it is very likely that the condition of a female diagnosed with schizophrenia could be attributed to genetic or inherited problems. In addition, a greater proportion of participants with tertiary education (45.6%) than participants with primary/secondary schooling (21.4%) indicated that they did not know if the condition of a female person diagnosed with schizophrenia could be attributed to genetic or inherited problems. Table 8 indicates the frequencies and row percentages for Item 9 for the two levels of education groups (primary/secondary school and tertiary).

Table 9 shows that in comparison with the participants with tertiary education (22.1%), a much greater proportion of participants with primary/secondary school education (46.4%) believed it is less likely that the condition of a female diagnosed with schizophrenia could be attributed to evil spirits/sorcery. A greater proportion of participants with tertiary education (48.5%) than participants with primary/secondary schooling (28.6%) indicated that they did not know if the condition of a female diagnosed with schizophrenia could be attributed to God’s will/sorcery.

This study explored participants’ conceptualisations of schizophrenia. Two vignettes (one of a male and another of a female character who meets the diagnostic criteria for schizophrenia) were compiled and presented to the participants with specific questions regarding the conceptualisation of schizophrenia. The demographic variables were age, gender and level of education. These variables were further compared with the conceptualisation of schizophrenia to generate any association/connection. The responses provided were categorised as psychological, medical, traditional and social to generate meaningful data.

The findings in this research study revealed an adoption of a more non-western explanation, which is traditional (spiritual) when a male is diagnosed with schizophrenia. This demonstrates the accommodation of a single explanatory model of schizophrenia, thus signifying an embracement of non-Western influence. Traditional (spiritual) causes were the most dominant response, particularly age being an important variable. This contradicts Kometsi (2016), whose findings depicted older participants attributing schizophrenia to medical, social and spiritual factors (multiple explanatory models). Furthermore, a general trend in the literature identifies genetics as a cause of schizophrenia (Legge et al., 2021).

When a female is diagnosed with schizophrenia, Western influences (medical) are embraced. Education becomes an important variable, with participants with low education embracing medical causes more than traditional ones. This is in contradiction with Furnham’s et al. (2011) findings, where they found that participants who had studied psychology and medicine scored higher on the causes of schizophrenia than their counterparts.

Gender did not have any impact on the conceptualisation of schizophrenia for both a male and a female diagnosed with schizophrenia.

While this study can be regarded as achieving its goal to a certain extent, certain limitations must be pointed out. In view of these limitations, the following recommendations can be made for future research:

• In this study, it is acknowledged that the findings may not be universally applicable, as they focused only on three demographic variables: age, gender and level of education. Therefore, future studies should explore other demographic variables such as home language and marital status. It is also important to recognise that different variables that affect mental health literacy may operate in intersectional ways and that issues of stigma may fundamentally affect how people with schizophrenia are treated. Future studies must take these issues into account.

• In addition to these concerns, the study cannot fully account for complex social and economic factors that influence the experiences of different racial or age groups. It is therefore suggested that future research be broadened to include more diverse racial and age groups to improve the generalisability of findings and capture more wide-ranging experiences. In addition, larger-scale studies on the epidemiology of mental disorders in relation to issues of social class and economic exclusion will be important, especially as it is well-established that there is an association between poverty and mental ill-health (Lund et al., 2010).

In conclusion, the study unveils intriguing patterns in the attribution of causes for schizophrenia among different demographic groups. Older participants tend to attribute the cause of schizophrenia to traditional factors when a male is diagnosed, reflecting the possible influence of cultural and generational perspectives on perceptions of mental health. Conversely, participants with lower education levels show a tendency to associate the cause of schizophrenia with medical reasons when a female is diagnosed, underscoring the impact of educational background on the interpretation of mental health issues. These findings emphasise the complexity of attributions surrounding schizophrenia, indicating that age and education play distinct roles in shaping perceptions. Recognising and understanding these variations is crucial for developing targeted interventions and educational initiatives that consider the diverse perspectives within our society. Further research is warranted to delve deeper into the underlying factors influencing these attributions and to explore potential implications for mental health awareness and support initiatives.

Funding: The author(s) received no financial support for this article's research, authorship and/or publication.

Declaration of conflicting interests: The author(s) declared no potential conflicts of interest to this article's research, authorship and/or publication.