We found the overall family stigma in the community to be of moderate level. Furthermore, living in rural place, explanations regarding the cause of mental illness, perceived signs of mental illness were associated with family stigma. However, living in rural place was the strongest predictor of high family stigma.
The moderate level of family stigma in the current study can be directly or indirectly associated with the public stigma against PWMI or due to low mental illness information as found in the current study. A previous study in the same study area reported that there was high public sigma against PWMI . Nonetheless, the current score was lower compared to the stigma against PWMI score reported in the previous study .
Rural residents have shown significantly high stigma than urban residents which may be due to low mental health literacy and rural respondents may be disadvantaged of other underlying causes such as high illiteracy, low media and mental health service access which implies that reducing the gap on such determinants may enhance reducing of stigma against family members of PWMI.
One of the reasons of stigma development is lack of explanation and fear about a given illness [1, 31]. Similarly, in the current study both high perceived supernatural and psychosocial and biological explanations of mental illness were significantly correlated with lower stigma against family members of PWMI. This indicates that there is high need for programs targeted at increasing the public awareness about the causes and nature of mental illness to reduce stigma against family members of PWMI.
High supernatural explanation of mental illness was associated with lower stigma at lower level of exposure to PWMI. This can be related to the type of explanation and sympathy that people with high supernatural but lower exposure to PWMI might have i.e. they may be less likely to blame the family for the relatives’ mental illness. Similarly, significantly lower stigma was obtained when individuals scored high on exposure to people with mental illness at high education level. This may be due to the combination of high education level which can facilitate exposure to diverse media on mental illness and enhance the ability to understand messages related to PWMI.
High number of reported signs of mental illness by the public was significantly correlated with lower stigma against family members of PWMI. Similarly, stigma against PWMI was lower among people who were familiar to the illness, and those who had previous contact to persons with mental illness [19–21, 25, 32–34]. People who are aware of many signs of mental illness may have better general information about mental illness through formal and informal means. Thus, they may have also less stereotyped beliefs and prejudices.
Respondents who had high income but low education showed significantly high family stigma. Such type of respondents may be in a disadvantage to get more information about mental illness from other sources like print and visual media. In addition, they may also have limited opportunity to get awareness and knowledge about mental illness from the school environment.
Generally, in the current study there was a high tendency of blaming family members for the illness of the patients. The belief among the public for the need to restrict the patients by the family members to avoid contact to the community may be associated with the type of explanation of mental illness and perceived dangerousness of people with mental illness. On the other hand, a low score was observed on restricting family members from being a member of social gatherings. In the multivariate analysis, no significant correlation was scored between many socio-demographic characteristics (i.e., age, sex, marital status, religion, ethnicity and occupation) and stigma against PWMI. Exposure to PWMI was very low in the current study which calls for mental health awareness interventions in the study community.
This study is the first of its kind exploring family stigma in Ethiopia. The relatively large randomly selected community sample representing diverse social and economic background adds to the robustness of our data. Although we have achieved semantic equivalence of the measurement, the lack of other aspects of validation could be potential limitation. In addition, the face-to-face interviews, which were most appropriate in the context of high level of illiteracy, may have resulted in social desirability bias while responding stigma items. Nevertheless, our findings contribute to the existing body of knowledge regarding the correlates of family stigma in low-income setting.