To our knowledge this is the first report of suicidal behaviour among community dwelling elderly persons in sub-Saharan Africa. We sought to identify the prevalence of three behaviours which are known to occur in a continuum between normality and actual suicide. In this study about 1 in 20 elderly persons had experienced suicidal thoughts since they attained the age of 65 years while about 1 in 100 had made suicidal plans during this period. Given that the mean age of the sample at assessment was about 75 years, the suggestion is that these events had occurred over a period of about 10 years. The three reported behaviours provided a pattern that is not uncommon: thus while about 20% of persons with ideation in old age had made a plan, the proportion who had made an attempt among ideators during the period was 6%. A somewhat similar pattern has also been reported in an earlier study in the general population of Nigerians above the age of 18 years [24
Estimates of suicidal ideation vary widely across studies depending on the methodology of the study, the time frame considered, or the comprehensiveness of the assessment instrument. Thus, studies of community dwelling older adults from other parts of the world report that suicidal ideation occur at rates that range between 2.16% and 17% [25
]. Direct comparison with other large surveys is difficult since our assessment of suicidal behaviour was specifically made on the period since the respondents attained the age of 65 years, rather than over the entire lifetime. Also, we cannot be sure of the extent to which reluctance to report these behaviours might have affected our rates. Suicidal behaviours still carry a significant level of stigma in the cultural setting of our study. However, our results show a trend similar to that reported among community dwelling older people in previous studies conducted in the West [26
]. However, we note that suicidal ideation in this elderly sample is much higher than in the general Nigerian adult population in which a figure of 3.2% has been reported over the lifetime in a community sample with a mean age of about 35 years [24
]. This finding suggests that old age, at least in the country setting of our study, is a significantly stressful period of life and a risk period for suicidal behaviour. The observation is in consonance with a previous report from our group suggesting one of the highest rates of major depression, both current and lifetime, among elderly persons anywhere [17
]. Although most people with suicidal ideation will not progress to suicidal attempt, the older adult is however more likely to make determined effort at progressing through to actual suicide using the most lethal methods compared to younger subjects [12
]. The fact that only 0.2% of our sample admitted to have made a lifetime suicidal attempt does not contradict this evidence. Since we do not have information about those who might have committed suicide in old age, the possibility exists that perhaps our observation only reflects the fact that only those who survived an attempt were alive to tell the story of their attempt. Furthermore, cultural issues about admitting to previous suicidal attempts may have affected the rates of attempted suicide reported in this study.
We also sought to identify the predictors of suicidal behaviours in this population. Living in a rural location predicted the occurrence of suicidal ideation, while being separated either through death or divorce predicted suicidal planning. In the period before the actual suicide, older people are known to experience various kinds of losses that heavily burden them. The loss of a partner has been particularly linked to an increased risk of suicide, especially in men [28
]. Spousal loss reduces the amount of social support available to the elderly and thus increases the likelihood of social isolation. It has been argued that increased likelihood of social isolation among the elderly compared to younger persons is one of the reasons for the greater risk of suicide in the former [31
]. Social isolation also predisposes to depression, a major risk factor for suicide [27
]. Social isolation as well as exclusion may also explain why, in this study, living in a rural location predicted suicidal ideation. Although it has previously been argued that social isolation is more common in large populations [32
], the circumstances leading to social isolation in the elderly in contemporary Africa may be slightly different. For instance, in Nigeria as in many low income countries, the young and able bodied individual drift from rural areas to urban centres seeking better economic opportunities, leaving behind a population of older persons in the rural areas. The result is a reduction of the social network available to older persons. Furthermore, in Nigeria as in other sub-Saharan countries going through social changes, there is a gradual but continuous erosion of the traditional extended family system. Typically, rural areas are deprived of social amenities and opportunities for recreation. When social amenities are available, they are often not tailored to the needs of the older person. The pattern of the relationship between rural dwelling and suicidality has been reported in the general population of many countries [33
]. This appears to be in keeping with the theory of Durkheim in which poor social integration leads to increased suicide in the population [7
]. Depression has often been reported in the literature as the most important predictor of suicidal behaviour in older adults, especially given its relationship with other risk factors such as social isolation, disability and illness [27
]. However lifetime depression was not a significant predictor of suicidal behaviour in this study. One reason for this could be due to the use of different timeframes for the two measures: lifetime for depression and old age for suicidal behaviour. Perhaps a different pattern might have been observed if depression since age 65 years was examined. The possibility that recall bias might have affected the assessment of lifetime depression is also a plausible factor. Also, because of the cross-sectional nature of the data, we were unable to clearly differentiate depressive episodes occurring before suicidal behaviour, the variable of interest, from those occurring after. The effect of this would be to introduce a lot of “noise” to the examination of the association with the resultant loss of statistical power. For these reasons, we could not rule out the potential role of depression in predicting suicidality in our sample. Nevertheless, some similarly designed studies have reported that only a minority of persons with suicidal thoughts have depression symptoms sufficient for the diagnosis of a major depressive disorder even though they may have depressed mood and some other associated symptoms [11
]. This finding reflects the importance of several other variables apart from depression in the complex process of suicidality.
In interpreting the results of our study, it is important to bear in mind its limitations. These include its reliance on a retrospective measure of suicidality and depression, as well as the cross-sectional nature of the study. As noted earlier, stigma around the topic of suicide is a potent factor in the population in which our study was conducted and may have resulted in conservative estimates of suicidal behaviour. Nevertheless, the findings are novel in so far as they present the picture of an important public health issue in this under-studied population.