Public Stigma against People with Mental Illness in Jimma Town, Southwest Ethiopia

Background Stigma towards people with mental illness (PWMI) can result in low self-esteem and isolation and threaten employment. Therefore, this study aimed to assess the magnitude of public stigma against PWMI and factors associated with it among Jimma town residents. Methods A community-based, cross-sectional, descriptive study was conducted in adult residents of Jimma town. Data were collected among 820 randomly selected residents with the interviewer-administered Community Attitudes toward the Mentally Ill (CAMI) scale. Linear regression analyses were performed to identify predictors of stigma against PWMI. Result A total of 444 (54%) of the 820 respondents were females, and the mean (SD) age was 35 (8.5) years. The minimum and maximum possible values on each CAMI subscale were 10 and 50, respectively. The respondents had high scores for a stigmatizing attitude towards PWMI across all the subscales, as indicated by the mean (SD) scores: authoritarianism, 27.17 (4.96); social restrictiveness, 32.41 (4.20); benevolence, 35.34 (4.42); and community-based mental health ideology, 33.95 (5.82). Compared to housewives, private organization employees showed more autocratic and socially restrictive views (std. β = 1.12, P<0.01). Single people had a lower social restrictiveness stigma score than married people (std. β = -0.20, P<0.001), and participants’ academic levels correlated inversely with the stigma score (std. β = -0.12, P<0.001). A higher benevolence stigma score was observed among participants with no relationship with PWMI than among those with PWMI in their neighborhood (std. β = 0.08, P< 0.046). Conclusion The study revealed that a negative attitude towards PWMI is widespread. Therefore, there is a need to develop strategies to fight the stigma attached to PWMI at the community level.


Result
A total of 444 (54%) of the 820 respondents were females, and the mean (SD) age was 35 (8.5) years. The minimum and maximum possible values on each CAMI subscale were 10 and 50, respectively. The respondents had high scores for a stigmatizing attitude towards PWMI across all the subscales, as indicated by the mean (SD) scores: authoritarianism, 27.17 (4.96); social restrictiveness, 32.41 (4.20); benevolence, 35.34 (4.42); and community-based mental health ideology, 33.95 (5.82). Compared to housewives, private organization employees showed more autocratic and socially restrictive views (std. β = 1.12, P<0.01). Single people had a lower social restrictiveness stigma score than married people (std. β = -0.20, P<0.001), and participants' academic levels correlated inversely with the stigma score (std. β = -0.12, P<0.001). A higher benevolence stigma score was observed among participants with no relationship with PWMI than among those with PWMI in their neighborhood (std. β = 0.08, P< 0.046).

Conclusion
The study revealed that a negative attitude towards PWMI is widespread. Therefore, there is a need to develop strategies to fight the stigma attached to PWMI at the community level.

Ethical statement
Ethical clearance was obtained from the Jimma University Ethical Review Board, and written consent was obtained from the Jimma Town Health Office. Before the actual interview, written informed consent was obtained from all study participants. Data collectors read the consent statement to those study participants who could not read and then obtained their fingerprints as indication of consent. To ensure confidentiality, all interviews were conducted in private places, and data were recorded anonymously.

Study design and setting
We performed a community-based, cross-sectional study in Jimma town, Southwest Ethiopia, from October to November 2012. Jimma town has 3 districts (woreda) and 13 sub-districts (kebeles). The number of households reported in the town at the time of the study was about 32,191. The 2007 Central Statistical Agency (CSA) census reported the total population of Jimma town to be 120,960 [34], and with a projected growth rate of 4.7 the estimated total population in 2012 was around 152,186. The town has 2 hospitals: a general hospital and a university hospital. Inpatient mental health services are available only at the latter, Jimma University Specialized Hospital.

Participants and sample size
To determine the sample size, we used the formula for a single population proportion. Because there were no published data on the specific research area, the total sample size was calculated with an assumption of a 50% public stigma level. The minimal difference was taken as 5%, with a 95% confidence interval. Therefore, the estimated sample size was 384. After considering the design effect, the calculated sample size increased to 768. We assumed a 10% nonresponse rate, so that the final total sample size was 845.
We used multistage probability sampling. From the 13 kebeles (smallest administrative unit) in the town, 5 were selected by a lottery, then the total sample size was proportionally allocated to each kebele. Each sampling unit (household) was selected by systematic random sampling. The first study unit was selected randomly, and if there was more than one study participant (individuals age 18 or above) in a household, a simple random sampling method was used to select one person to participate. If no illegible respondent was found in the selected household, the next household was selected for the study.

Measurement
The outcome variable was public stigma against PWMI. The exposure variables included sociodemographic and economic factors such as age, sex, religion, ethnicity, level of education, type of occupation, household income, and number of individuals in the household. Other variables such as history of contact with PWMI, history of mental illness, mental health information, and threats from PWMI were included in the study.
Two questionnaires were used. The first comprised questions about sociodemographic variables and other independent variables, and the second was a Community Attitudes toward the Mentally Ill (CAMI) scale to measure the level of public stigma. The CAMI scale includes 40 items rated on a five-point Likert scale from 1 to 5 (1 = strongly agree, 2 = agree, 3 = neither, 4 = disagree, and 5 = strongly disagree) and divided into four subscales: Authoritarianism (AU), Benevolence (BE), Social Restrictiveness (SR), and Community Mental Health Ideology (CMHI) [30]. Authoritarianism refers to a "view of the mentally ill person as someone who is inferior and requires supervision and coercion"; Benevolence is "a humanistic and sympathetic view of mentally ill persons"; Social Restrictiveness covers "the belief that mentally ill patients are a threat to society and should be avoided"; and Community Mental Health Ideology corresponds to "the acceptance of mental health services and the integration of mentally ill patients in the community." Each subscale consists of ten items and has a minimum score of 10 and maximum score of 50.
The original English version of the CAMI questionnaire was translated into Amharic and Afan Oromo languages and then back translated into English by native speakers and experienced professionals in order to check its consistency. Before the start of the study, consensus on the compatibility of the forward translations to Amharic and Afan Oromo was assured. In this study, the reliabilities of the Amharic version of the CAMI subscales were as follows: AU, α = 0.54; BE, α = 0.46; SR, α = 0.60; CMHI (α = 0.79).
Before the actual data collection, the questionnaire was pre-tested on 5% of a similar population in another kebele, and both the principal investigator and supervisors assessed the clarity, understandability, flow, and completeness of the items and the time needed to complete the questionnaires. This step helped to correct systematic errors, ensured consistency in the flow of items and provided an estimate of the time needed to complete the questionnaire.

Statistical analysis
The data were checked for completeness and entered into EPI-DATA version 3.1. For analysis, data were exported to SPSS version 20.00 and then cleaned. Reverse coding was performed for negative statements before starting the analysis. Descriptive statistics (frequencies and percentages) were calculated for important variables. Before performing linear regression, the stigma values were checked for normality through histograms and kernel density plots. After performing assumption tests, correlation of continuous independent variables was tested, and an independent sample T-test and one-way ANOVA were used to determine the association of categorical variables with the outcome variable. Finally, variables found to be associated by Pearson correlation, T-test and one-way ANOVA were computed to multiple linear regression to control confounders by using a significance level of <0.05.

Exposure to PWMI, and history of mental illness among respondents
Among the total of 820 respondents, 703 (85.7%) had encountered PWMI, and of those a large proportion (47%) reported that PWMI lived in their neighborhood; 27.4% reported that they had no relation to PWMI (i.e. encountered PWMI only on the streets). One-third of the respondents (33%) reported that they had been threatened or attacked by PWMI, and 61% had witnessed others being threatened or attacked. A small proportion (6.2%) of the respondents reported that they themselves had suffered from mental illness at least once in their life time.

Mental health information
Of the total of 820 respondents, 461 (56.2%) reported that they had heard any information about mental health within the last year, and the remaining 359 (43.8%) reported that they had no information about mental health. Among the respondents who had heard about mental illness in the last year, the leading sources of information were reported to be television (40.1%) and a combination of different sources of information (31%); relatively fewer respondents reported that they had obtained information from health institutions (4.3%) or magazines (4.3%).

Stigma against PWMI
The internal consistency for the total CAMI scale was found to be α = 0.73; among the subscales, CMHI had the highest internal consistency (α = 0.79) and BE the lowest (α = 0.46). The highest stigma score was on the BE subscale (mean 35.34, SD 4.42), and the lowest stigma mean score was on the AU subscale (27.17 ± 4.96) Authoritarianism. The majority of the respondents (74.9%) believed that keeping PWMI behind locked doors is the best way of handling PWMI (A2). More than three-fourths of the participants (86.3%) stated that adults with mental illness needed the same kind of control and discipline as young children (A5), and more than half (56.3%) rejected ("disagree" or "strongly disagree") the view that virtually anyone can become mentally ill (A10). The majority (88.2%) also agreed with the statement "Mental hospitals are an outdated means of treating the mentally ill" (A9). Even though the mean stigma score was lower on this subscale than on the others, the majority still accepted the negative attitude statements ( Table 2).
The multiple linear regression showed that occupation significantly predicted authoritarianism: government employees had 0.13 higher stigma scores in this subscale than housewives (t (791) = 2.49, P = 0.013), and private organization employees had 0.10 higher scores than housewives (t (791) = 2.04, P = 0.042). These results show that both government employees and employees of private organizations have a more autocratic attitude towards PWMI than housewives. The results of the regression indicated that the predictors explained a significant proportion of variance in the AU score (R 2 = 0.081, F (26, 791) = 42.64, p < .001) ( Table 3).
Benevolence. The majority of respondents (68.4%) rejected the statement that describes the mentally ill as having been the subject of ridicule for too long (BE1). Although a large portion of the respondents (96%) agreed that they have a responsibility to provide the best possible care for PWMI (BE5), 84.0% believed that increased spending on mental health services is a waste of tax money (BE8). A total of 85.1% of respondents rejected the statement "We need to adopt a far more tolerant attitude toward the mentally ill in our society" (BE3), and more than three-fourths agreed with the statement "It is best to avoid anyone who has mental problems" (BE 10). In general, the majority of respondents agreed with the negative attitude statements ( Table 4).
The two most important variables that significantly predicted benevolence were relationship to PWMI and occupation: government employees had 0.10 higher stigma scores for benevolence than housewives (t (808) = 2.30, P = 0.022), and private organization employees had 0.11 higher scores (t (808) = 2.04, P = 0.016). Considering neighbors as a reference, those who had no relationship to PWMI (i.e. only encountered PWMI on the street) had a 0.08 higher benevolence stigma score than those with PWMI in their neighborhood (t (808) = 2.00, P = 0.046). These findings imply that those who only encountered PWMI on the street were more stigmatizing, i.e. they were less benevolent towards PWMI than those who had PWMI in the neighborhood. Also, both government and private employees are less benevolent towards  Table 3). Social Restrictiveness. The statement "The mentally ill should be isolated from the rest of the community" (SR2) was found acceptable by majority of the respondents (84.8%), and 93.0% of them did not want to live next door to someone who has been mentally ill (SR4). A total of 82.7% of the respondents believed that a woman would be foolish to marry a man who had suffered from mental illness, even though he seemed to be fully recovered (SR3). The majority of respondents (73.7%) believed that anyone with a history of mental problems should be excluded from taking public office (SR5). Only 7.3% of respondents believed that PWMI should not be denied their individual rights (SR6), and 78.3% of the respondents did not agree that PWMI should be encouraged to assume the responsibilities of normal life (SR7). A total of 89.6% of respondents rejected the statement "No one has the right to exclude the mentally ill from their neighborhood" (SR8). In general, most of the residents believed that PWMI are a threat to society and should be avoided, indicating that most of them had socially restrictive view towards the PWMI (Table 5).
Marital status, education, occupation, and encountering PWMI independently predicted social restrictiveness and explained a significant proportion of the variance in the SR stigma score (R 2 = 0.066, F (18, 801) = 4.24, P<0.001). Marital status was associated with social restrictiveness: the mean SR score among single people was 0.20 lower than among married ones (t (801) = -3.85 P<0.001), indicating that married people have more socially restrictive views towards PWMI than single people. Social restrictiveness increased with higher educational level: people with elementary school education had a 0.12 lower mean SR stigma score than people with a secondary school education (t (801) = -3.26, P = 0.001), and those who were at college or university had a 0.99 higher score than secondary school students (t (801) = 2.61, P = 0.009).
Occupation significantly predicted social restrictiveness: employees of private organizations had a 0.11 higher mean stigma score than housewives (t (801) = 2.57, P = 0.010), as did students (t (801) = 1.97, P = 0.049). Both students and employees of private organizations in general had a more socially restrictive view than housewives. The results of logistic regression showed that encountering PWMI independently had a significant effect on social restrictiveness, i.e. those who had not encountered PWMI had a 0.04 lower SR stigma score than those who had encountered PWMI (t (801) = -1.98, P<0.048) ( Table 3). Community Mental Health Ideology. The statement "Residents should accept the location of mental health facilities in their neighborhood to serve the needs of the local community" (CMHI1) was rejected by 90.9% of residents. Many respondents (85.0%) also did not agree with the statement "The therapy for many mental patients is to be part of a normal community" (CMHI2). A total of 74.5% of respondents had a negative view about providing mental health services through community-based facilities (CMHI3), and a large number (91.5%) rather believed that locating mental health facilities in a residential area downgrades the neighborhood (CMHI10). In this subscale, again the majority of participants tended to hold a negative attitude towards PWMI, but less so than in the AU subscale (Table 6).
In the logistic regression, only education had an independent effect on community mental health ideology; it explained 4.5% of the variance in the stigma score of the CMHI subscale (R 2 = 0.045, F (19, 800) = 3.033, P< 0.001). Those who were unable to read or write had a CMHI The best therapy for many mental patients is to be part of a normal community (CMHI2) 101 12. It is frightening to think of people with mental problems living in residential neighborhoods (CMHI9) 547 66.7 23 2 250 30.5 stigma score 0.08 lower than those who had attended secondary school (t (800) = -2.20, P< 0.028), and those who were only able to read and write had 0.10 lower score (t (800) = -2.78 P<0.006). Respondents with only primary school education have 0.09 less score of community mental health ideology than those of secondary school educational level (t (800) = -2.30, P = 0.022). In general, relatively least educated have better community mental health ideology than that of well educated. (Table 3).

Discussion
The current study attempted to determine the magnitude of public stigma against PWMI and factors associated with such stigma. Stigma was assessed by the Community Attitudes towards the Mentally Ill (CAMI) scale, which has four subscales: Authoritarian (AU), Benevolence (BE), Social Restrictiveness (SR), and Community Mental Health Ideology (CMHI). Similar to studies performed in Southern Ghana [17], Nigeria [18], India [31], and Taiwan [29], negative views were prevalent in all of the CAMI subscales; in contrast, however, our study found few relative variations among the mean scores of the subscales. A higher mean stigma score was found for BE and CMHI; this finding was similar to the study performed in Southern Ghana [17], but the highest stigma scores recorded in the other studies were for the subscales AU and SR [18,29,31]. The finding may differ from that of the Indian study [31] because of its very small sample size (n = 100). The difference from the Nigerian [18] and Taiwanese [29] studies could be a result of the different study population used (study participants in the Nigerian study were teaching hospital senior staff members, and the Taiwanese study used telephone interviews). Our finding may imply that residents of Jimma town are less benevolent and community mental health-oriented and also less authoritarian and socially restrictive than the participants in the other studies [18,31].
The present study found that residents of Jimma town have an autocratic attitude towards PWMI, shown by the findings that the majority of the respondents believe that keeping PWMI behind locked doors is the best way to handle them (74.9%) and that adults with mental illness need the same kind of control and discipline as young children (86.3%). This finding is similar to that of the Indian study mentioned above [31]. It implies that residents of Jimma town prefer aggressive treatment of PWMI, perhaps because the majority (85.7%) of them had encountered PWMI, 33% of them had been threatened or hurt by PWMI, and 61% had witnessed others being hurt or threatened. Thus, many had experienced aggression in one way or another, which may create fear. Participants may have preferred aggressive treatment of PWMI to protect themselves from impending danger.
More than half of the participants (56.3%) rejected the view that virtually anyone can become mentally ill, which is consistent with findings from Southern Ghana [17], South Africa [20], and India [31]. This may imply that the residents of Jimma town have poor knowledge on perceived causes of mental illness and treatment options; this hypothesis is supported by the finding that 43.8% of the respondents had no information about mental health. Regarding a benevolent view, almost all respondents were of the opinion that they have a responsibility to provide the best possible care for PWMI, a finding that is similar to the study performed in Southern Ghana [17]. Similar to the Indian study [31], more than three-fourths of the respondents considered that it is best to avoid anyone who has mental problems and believe that PWMI do not deserve their sympathy [31]. This implies that respondents place less emphasis on care and choose rather to avoid PWMI than to provide care. This could be a result of poor information on the nature of mental illness and might be also a reaction to the fear of aggression from PWMI. The majority of respondents had a socially restrictive view towards PWMI. Most respondents said that they do not want to live next door to someone who has been mentally ill and are against a woman marrying a man who has fully recovered from mental illness. This finding is similar to that of two studies performed in Nigeria [16,18]. Only a minority of the respondents believe that PWMI should not be denied their individual rights, and they also believe that excluding a mentally ill person from the community is not wrong and is permitted. Similar to the studies performed in India [31] and Southern Ghana [17], rather than helping PWMI to assume a normal life and public responsibility, the majority of the respondents would rather exclude and discourage PWMI. Overall, the residents are likely to deny patients their rights, not allow them to take on responsibility, and prohibit marriage and living together with PWMI. These findings may be due to the fact that the majority of the respondents had encountered PWMI on the street, which may have made them think that mentally ill people are unable to look after themselves and are dangerous.
The location of mental health services is challenged by the majority of the respondents in that the statement "Residents should accept the location of mental health facilities in their neighborhood to serve the needs of the local community" was rejected by 90.9% residents. A total of 85% of respondents do not believe that the therapy for patients with mental illness should be part of a normal community. This implies that the residents do not accept that it is helpful for patients to stay in the community, which might also be an additional reason why they prefer to keep the PWMI out of the community. Nearly three-fourths of respondents have a negative view about the provision of mental health services through community-based facilities, which may represent an obstacle when attempting to integrate mental health services at the community level. A large percentage of respondents (91.5%) believe that locating mental health facilities in a residential area downgrades the neighborhood; this finding is consistent with the Nigerian study [18]. This can be explained in a similar way to the previous finding, i.e. respondents prefer to exclude PWMI from the community. The observed agreement with two apparently contradictory statements in different parts of the CAMI scale could have been due to the acquiescent response style of "yay-saying" in the local culture [35].
The study found that highly educated people are more socially restrictive than less educated people. This might mean that better educated people have higher expectations of social responsibility than illiterate people and therefore do not trust PWMI, a finding that is consistent with that of the study performed in Taiwan [29]. Those respondents who attended only primary school and below were less community mental health oriented than those who had attended secondary school, i.e. the latter approved the integration of PWMI and mental health services at the community level more than people with less education; this finding is consistent with the study performed in Southern Ghana [17].
Occupation was associated with scores on all subscales except for CMHI, in which government and private organization employees were less benevolent and more authoritarian than housewives. We found that housewives held more humanistic and sympathetic views towards PWMI; this may be because most housewives are mothers, and their maternal instinct may contribute to their caring and sympathetic tendencies. The participants who were students and private organization employees were more socially restrictive than the housewives, which can be explained in the same manner as the issue of less authoritarianism and benevolence. The studies in Nigeria [18] and India [31] mentioned above did not find that occupation was associated with any of the subscales. This may be because of the different study participants: almost all of the participants in the Nigerian study were medical workers, and the Indian study included only 100 rural residents. In contrast, our study participants were from an urban setting.
Respondents who had PWMI in their neighborhood were more benevolent than those who had no relationship to PWMI, which may mean that those who are more familiar with PWMI are more caring and sympathetic towards them than those who are not familiar with them. This finding is similar to that of a study performed in Toronto, which showed that familiarity with PWMI is associated with a low level of stigma [30].
Having information on mental health was not found to be significant in any of the subscales. However, this finding does not imply that having such information could not reduce the level of public stigma, because our study only asked whether participants had ever heard any information on mental health during the last 12 months and did not ask for any details (e.g. type of information, duration, frequency etc.). Therefore, this point requires further investigation before any conclusions can be reached.
When compared with the Ethiopia Demographic and Health Survey report of 2011, the educational status of our study participants was higher than that of the average Ethiopian community. This is probably because our study participants were recruited from an urban setting (Jimma town), whose community has much better access to education. This could limit the generalizability of our findings to the entire Ethiopian population, who are mainly rural and consequently tend to have a lower education status [36].

Conclusions
The findings of this study indicate that residents of Jimma town extensively hold a stigmatizing attitude towards PWMI. Overall, the residents are likely to deny PWMI their individual rights, prevent them from taking on various responsibilities and forbid people from marrying and living together with PWMI. In general, few respondents have information on mental health. In situations like this, mental health facilities are expected to play a great role in enhancing mental health awareness, but in contrast our findings show that health service institutions contribute less to delivering mental health information than other sources.
Occupation was the most important predictor strongly associated with authoritarianism, benevolence, and social restrictiveness: all associated subscales showed that housewives have a positive attitude. A higher level of education is related to more social restrictiveness but a better attitude towards accepting the integration of PWMI and mental health services on the community level. Respondents who had encountered someone with mental illness were more socially restrictive, and those who had PWMI in their neighborhood were more benevolent than those who had no relationship with PWMI. The results as a whole indicated that there is an urgent need to develop strategies to enlighten the public regarding the stigma of mental illness and to foster acceptance of PWMI by the community.