Non-fatal suicidal behaviour in rural Ethiopia: a cross-sectional facility- and population-based study

Injury related to “self-harm” is one of the leading causes of global disease burden in many parts of the world [1]. The burden is particularly high in high-income Asia-Pacific countries and Eastern Europe, where the burden of disease caused by self-harm is ranked fifth and sixth, respectively [1]. Self-harm related disease burden is reported to be relatively lower in Western Africa (ranked 69th) and Eastern sub-Saharan Africa (ranked 32nd). However, given the overall scarcity of data and stigma surrounding self-harming behaviours, this apparently lower burden in Eastern sub-Saharan Africa may be due to under-reporting or the dearth of studies reporting on the subject. For example, in a large scale international study coordinated by the World Health Organization involving 11 “developing” and 10 “developed” countries, the 12 month prevalence of suicidal ideation, plans and attempts were slightly higher among participants from developing countries [2]. In fact, 85 % of the over 800,000 annual suicide deaths [3] are likely to occur in low and middle income countries (LMICs), where resources to prevent suicide deaths are severely limited. Suicide might indeed be a hidden cause of death in traditional societies because of the high level of stigma and associated religious and cultural condemnation [4, 5]. For example, in a study of 100 key informants, primarily consisting of community and religious leaders from Ethiopia [4], people who commit suicide were considered “condemned sinners” who do not deserve appropriate funeral rituals but should be buried in isolated unmarked place. Those who attempt suicide were “feared” and considered “cruel” and “untrustworthy”.

Yet data from LMICs remain sparse and, where available, the incidence of suicide and non-fatal suicidal behaviour shows marked variability [6]. The potential role of rural residence on rate of suicidal behaviour is also not adequately explored. An old study evaluating the rate of suicide in West Africa [7] reported a higher rate of suicide among rural residents; however, the risk of under-reporting in this study is very high.

In studies from Ethiopia, the reported lifetime prevalence of suicide attempts varied between 0.9 and 14.8 % [8, 9]. The rate of suicide attempts was particularly high among the homeless in Addis Ababa [10] and those with mental disorders [11‐13]. Studies from other African countries report a similarly variable prevalence of suicidal behaviour. In a study conducted in a primary care facility in Kenya, 12.7 % of participants were reported to be at risk of suicide, although the specific rates for suicide attempt or ideation were not provided [14]. In Morocco, the one month prevalence of suicidal ideation was 6.3 % [15]. A study from Nigeria reported a 3.2 % prevalence of suicidal ideation and 0.7 % for suicide attempt. [16] In South Africa, the lifetime prevalence of suicide ideation, plans and attempts were 9.1, 3.8 and 2.9 % respectively [17]. Despite the relatively large variation, these figures indicate that suicidal behaviour is an important problem in Africa. Several risk factors for suicidal behaviour have been reported at the health system, community and individual level. The system level risk factors focus on access to means of suicide and health care. The individual factors include demographic features (female gender, younger age, lower education level, unmarried status) while stigma and discrimination are reported to be relevant community factors [3]. Mental disorders are perhaps the most important risk factors. Mood disorders, such as depression and bipolar disorder, impulse control disorder, and substance abuse are the most important disorders [18]. Given the dearth of data, particularly among rural communities, and the inconsistencies in prevalence figures, further studies to establish the prevalence and potential risk factors of non-fatal suicidal behaviour, such as demographic factors, depressive symptoms and alcohol use disorder [18], are required.

Moreover, non-fatal suicidal behaviours share risk factors with completed suicide [6] and are strong predictors of suicide even many years after the attempt [19]. It is therefore essential to understand non-fatal suicidal behaviour in order to develop intervention strategies which would also contribute to the prevention of suicide. In this study, we determined the prevalence of non-fatal suicidal behaviour (i.e. suicidal ideation, suicidal plan and suicide attempt). As in other studies [20], we defined suicidal ideation as the engagement in thoughts about self-inflicted harm to end one’s life. Suicidal plan was defined as “the formulation of a specific method through which one intends to die”. Suicide attempt represented “engagement in potentially self-injurious behaviour in which there is at least some intent to die.”

The study was conducted as part of the formative phase of the Programme for Improving Mental Healthcare (PRIME) [21], a multi-country research programme consortium, which works with Ministries, local governments and communities to integrate mental health services in primary healthcare.

The data reported here were collected from a cross-sectional population-based survey and a facility-based survey of adults aged 18 years and above.

Non-fatal suicidal behaviour is a relatively common problem in this predominantly rural community in Ethiopia. The current findings taken together with our previous observations in this population [25] of high levels of common mental disorders, alcohol use, and life events, are indications of the public health relevance of mental disorders.

Suicide attempt is one of the most crucial predictors of completed suicide, even four decades after the attempt had occurred, making attempted suicide an important target for the prevention of suicide [38]. Suicidal ideation is also in the pathway to completed suicide and would constitute an important target for prevention of suicide [39]. Although attending health services following suicide attempt or suicidal ideation offers an opportunity for prevention, in this study setting, most people with suicidal behaviours did not use health services. Communication of suicidality with family members, providers or other people was also very low. The low service uptake may be explained partly by the lack of accessible services; however, the low communication of suicidality with family and providers is an important barrier for the implementation of prevention strategies. Population level interventions should target the high levels of associated stigma and condemnation in the population [40] to encourage openness on the subject. Intervention programmes being developed and delivered by PRIME in Ethiopia [26], which include packages for enhancing awareness, community inclusiveness, empowering of providers and supporting health system change, have the potential to be beneficial, if tailored to suicidal behaviours.

Nevertheless, there are key gaps unaddressed in our study, which need to be considered in planning or providing interventions. This study did not explore the nature of the suicidality. For example, better information about the severity of the suicidality, the nature of the plans and what the means of the attempts were, would usefully inform public health intervention. In most rural developing country settings, organophosphate poisoning is a common method of suicide attempt and is associated with increased mortality due to lack of accessible health services [41]. Relevant triggers, such as social conflicts, family problems and financial strains should have been explored to identify targets for non-health interventions. Thus understanding the methods and immediate contexts of the attempt would inform the form and content of intervention to be provided for those with suicidal behaviour.

The factors associated with suicidality in this study, young age, alcohol use disorder and depressive symptoms, are consistent with what has been reported previously [16, 17]. However, the study was cross-sectional and thus causal interpretation of these associations can only be drawn based on previous studies.

The prevalence of non-fatal suicidal behaviour was relatively higher in this population compared to that reported in other populations in Ethiopia although rates are higher among those with symptoms of mental disorder. For example, whereas the lifetime prevalence of suicide attempt in the neighbouring district of Butajira was 3.2 % [8], the prevalence among those with minor depression in the neighbouring population of the Zay and Butajira was 5 and 14 % respectively [12, 13]. During a 10 year follow up of people with severe mental disorder, the rate of suicide attempt varied from 13 % in schizophrenia to 26 % in major depressive disorder [11].

Reason for the comparatively higher rate of suicidality in this study population may relate to the relatively higher level of mental distress, alcohol use and life events against a lower level of social support shown in our previous report [25]. This finding is also consistent with the relatively high rate of persistent suicidal preoccupations and suicide attempt assessed over a 30 day period in the same district [25]. Therefore, this higher rate of suicidality may be partly a reflection of the overall level of unmet mental health need in this population. A more detailed exploration of the context and factors contributing to suicide would have been relevant in determining the content of intervention and this is an important limitation of our study. Follow-up of this sample would also help determine shorter term clinical implications of these findings and should be considered. Although the cross-sectional nature of the study has the known limitations of recall bias and makes attribution of risk factors difficult, the study is adequate for establishing the magnitude of the problem of suicidality. Combining the community and facility sample may also introduce bias. However, the two samples were drawn from the same population (the Sodo district population) and combining would give us more power to detect relevant risk factors, especially in the analysis of suicidal attempt, which is a relatively rare but of higher public health importance. We present data separately before combining, which would allow interpretation of data separately. We also adjust for the source of the participant population (facility vs community) while investigating the role of other risk factors with larger sample size. Despite significant differences in the background characteristics of the two samples, study group was not a significant predictor of the target outcomes in the multivariable model. Finally, given this is a cross-sectional survey, the concept of transition from ideation to attempt may be considered inappropriate. Theoretically the two could occur at the same time although reverse causality is virtually impossible. Nevertheless, what has been attempted here is more pragmatic and the assumption has been made that, even when an attempt happens without premeditation, there is a degree of thinking that precedes any attempt.

The recent emphasis of the World Health Assembly is an important milestone in the context of findings such as ours suggesting a higher burden of disease related to suicidal behaviour. The WHO mental health action plan 2013-2020, identifies broad factors relevant for suicide and proposes multi-sectoral approach for the prevention of suicide. Furthermore, the action plan commits governments to reducing suicide by 10 %. Our data, complementing the sparse data originating from low and middle income countries, reveals that the hidden burden of suicidal behaviour is substantial. In the broader context of stigma and condemnation that may keep these behaviours hidden, public interventions to improve stigma are important. This study also indicates that addressing the need of young people, improving access to care for depression and alcohol use disorder are important next steps. Furthermore, we recommend further studies to understand the nature of the behaviour and context in which the behaviour develops, as well as developing effective brief interventions for the behaviour.

N/A.

The data is part of an ongoing cross-country study, Programme for Improving Mental Healthcare (PRIME). Because of the ongoing and cross-country nature of the study, we are not able to make the data publicly available at present. However, data will be made available once the study is completed in the next 3 years.