Tedium vitae, death wishes, suicidal ideation and attempts in Kenya-prevalence and risk factors

This analysis of suicidal thoughts and attempts derived from a mental health epidemiological survey of a household population in Maseno district, Nyanza Province in Kenya, found that over 25 % of the participants had thought that life was not worth living (tiredness of life or tedium vitae) at some point in their lives, while over 20 % had experienced death wishes at some stage. One twelfth reported suicidal thoughts and around one fiftieth had made actual suicide attempts at some point in their lives. In the adjusted analysis of factors associated with suicidal thoughts, being female, having CMD, having a number of recent life events and having a large social group size were all associated with increased rates of life time suicidal thoughts, but psychotic symptoms were no longer significant after adjustment for the other variables. In the adjusted analysis of suicide attempts, having any psychotic symptoms was the only factor associated with suicide attempts after adjustment for other factors significant at the bivariate level.

Previous studies of suicidal ideation in Kenya have focussed on samples of psychiatric inpatients [51] and outpatients [52], general medical inpatients and outpatients [53], primary care [54] and college students [55], but there have been no household population studies before this one. A comparison between students from four African countries and the US found that the prevalence of suicidal ideation ranged between 23.1 % and 31.9 % in the four African countries and was 16.9 % in the USA [56].

The World Health Organization World Mental Health Survey Initiative which studied 17 countries (including Nigeria and South Africa but not Kenya) found that the cross national prevalence of suicidal ideation and attempts was 9.2 % and 2.7 % respectively which is broadly similar to our findings in Kenya [57]. The national comorbidity survey in the US found prevalence rates of 2.8 % and 3.3 % suicidal thoughts and 0.7 % and 0.8 % suicidal attempts (and suicidal gestures combined), in the surveys of 1990–2 and 2001–3 respectively [8]. A national household survey of adults in Britain [12], found that a fifth of the sample [20.3 %] had thought that life was not worth living (tedium vitae) at some point in their lives, while death wishes and suicidal thinking were reported by a sixth of the sample, and 4.4 % made suicide attempts. A national survey in Australia found a life time risk of 16 % for suicidal thoughts and 3.6 % for suicide attempts [58]. Thus we found that suicidal ideation in this Kenyan population is more common than suicidal ideation in the UK and Australia, while suicidal attempts in Kenya were more common than that found in the US but rare than found in the UK. The only recent US survey reporting prevalence of desire for death is that by Baca Garcia [13] which found lifetime prevalence of 11.3 % and 10.2 % in two large US surveys compared to 19.2 % found in our Kenya study.

Risk factors associated with suicidal thoughts and attempts show some consistency across studies and countries. The World Mental Health Survey found that risk factors for suicidal ideation which were consistent across the 17 countries included being female, younger age, less well educated, unmarried (single, separated, divorced or widowed) and having a mental disorder [57]. The US National Comorbidity Survey found that higher rates were associated with being female, young age, less well educated, unmarried (single, separated, widowed and divorced) and unemployed [8]. An analysis of suicidal ideation and suicide attempts among Hispanic Subgroups in the US found that 45–64 year olds were at high risk [59, 60].

Since many studies (eg [12, 57, 59, 60]) have found an association between suicidality and age, it is surprising that an association between age and either suicidal ideation or attempts was not found in this study. A detailed cohort study in the UK study found that the decrease in reported previous-year suicidal thoughts with increasing age [59, 60] was partly explained by lower rates of reported abuse in childhood (in those aged 75+), depression, and anxiety symptoms (in those aged 55+), factors which are all strongly associated with suicidal thoughts; and partly by higher rates of protective factors in people aged 35+, specifically homeownership and cohabitation. Rates of phobias, irritability and compulsions also decreased with age, and the association of these symptoms with suicidal thoughts was particularly strong in the youngest (16–34) age group. People who reported experiencing childhood abuse in all age groups reported more suicidal thoughts, suggesting abuse has lifelong negative effects on suicidal ideation (59). It is possible that in our Kenyan sample, the older age groups have accumulated less protective factors than would be found in a Western sample.

The relationships we found between life events and suicidal ideation and attempts respectively were to be expected, and are found elsewhere [61] but the inverse relationship of suicide ideation and attempts with social network size is counter-intuitive. However, in the same survey we have found a similar inverse relationship of CMD and psychosis with social network size and perceived social support (unpublished papers). Similarly a previous study in Tanzania found a relationship of CMD with three or more recent life events but no relationship with social network size or perceived social support [40]; and a relationship of psychosis with two or more recent life events, but again no relationship with social network size of perceived social support [44]. Disturbed interpersonal relationships are an important precipitant of suicidal behaviour in the African setting, especially in young females [62]. It could be that those with a large network of friends are likely to suffer more shame and guilt if they have adverse life events, for example, childlessness and divorce which the woman may perceive as her own fault, hence leading to suicidal thoughts and self-harm. The disadvantaged position that women often occupy may contribute to the risk of suicidal ideation and attempts.

The relationships we found in the adjusted analysis between CMD and lifetime suicidal ideation, and between psychotic symptoms and suicide attempts are to be expected and are found in the world mental health survey in 17 countries [57] and in the UK [63, 64]. However, the multivariate analysis did not demonstrate significant associations between psychotic symptoms and suicidal ideation, and between CMD and life time suicidal attempts.

The strengths of the study are the use of a health and demographic surveillance site for the random sample of households, the high response rate, and the systematic approach to the clinical and socio-demographic assessments. The population in the surveillance site is regularly monitored by field staff who visit each household bi-annually to capture health and demographic information (birth rates, death rates, causes of death, pregnancies, immunization status, in- and out-migrations, etc.). Various studies nested on the HDSS platform take advantage of the sampling frame inherent in the HDSS, whether at individual, household/compound or regional levels. This familiarity with survey procedures is likely to have been influential in the achievement of a high response rate.

The findings of the study are limited by the fact that the cross-sectional design of the study does not allow causality to be inferred in the risk factor associations. It is also important to note that all the data was collected by a standardised structured research interview and was not confirmed by an interview with an expert psychiatrist, or by clinical records. People may be less willing to confide suicidal thoughts to a research interviewer than to an expert psychiatrist in a clinical situation.

The fact that the question about attempted suicide was only asked of those who had suicidal ideation might have left out the impulsive suicide attempts that are not preceded by suicidal ideation. This might have led to an underestimation of the true rate of attempted suicide. Furthermore, the question about attempted suicide gives only one example of a method, namely overdose of tablets, to help the participant understand what is meant by a suicide attempt. It is possible that not giving other examples may have led to under-reporting of suicide attempts by other methods, but since the question also says “or in some other way?”, we think it unlikely that participants would not recall suicide attempts which used other methods such as pesticides or asphyxiation.

Our data was collected via interview, which was then input to the PDA by the interviewer, rather than collected via self-completion and put directly onto the PDA by the respondent. Studies in the West show that higher rates are reported by self-completion than by interviewer-completion (e.g., [65]). Suicidality is regarded as a sin in religious cultures and this would increase discomfort in revealing suicidality in an interview, even to a health worker. Our suicide questions and other instruments were all reviewed by a local expert Luo psychiatrist and considered culturally relevant to the context.

The implementation of the study was hampered by a number of logistical challenges including the difficult terrain which posed problems for local transport for research staff and continuing administrative difficulties which led to delays in the implementation of the project. The interviewing period, initially planned to last 3 months, took place over a period of 6 months and was temporarily halted for several weeks over the period of the 2013 election due to further fears of election unrest.