The importance of mental disorders in suicide

A little learning is a dangerous thing;

Drink deep, or taste not the Pierian spring

—Alexander Pope, 1688–1740

The paper ‘Mental disorder and suicide: a faulty connection’ (Pridmore, 2015) is a continuation of Pridmore’s concern about the generally accepted importance of mental disorders in relation to suicide. It appears to be the culmination of a series of anecdotal reports in which he has conveniently omitted any data which has not supported his views; for example, in relation to Tolstoy (Goldney and Schioldann, 2011; Pridmore and Pridmore, 2011). While Pridmore may acknowledge the danger of extrapolating from large epidemiological findings to an individual, there does not appear to be similar caution in the converse. Indeed, his present offering is not only devoid of balance and introspection, but is also somewhat idiosyncratic in his allegedly supportive arguments.

The issue is not new … for over 200 years there has been debate in the literature about this. This has been reviewed in more detail elsewhere (Goldney et al., 2008), but early contributors included the clergyman Charles Moore who stated in 1790 that ‘there is a sort of madness in “every” act of suicide, even when all idea of lunacy is excluded’, and the physician George Burrows, who in 1828 noted that ‘a doubt may naturally arise, whether it be not sometimes perpetrated by a sane mind’. Many others have also contributed to this literature, not least the influential authors Bucknill and Tuke, who, in their 1858 classic, A Manual of Psychological Medicine, wrote: ‘the question so often asked, “Is suicide the result of cerebro-mental disease?” must be answered both affirmatively and negatively. That the act may be committed in a perfectly healthy state of mind cannot, for a moment, be disputed’.

And the debate has continued in the last 150 years. But, fortunately, modern statistical and investigative techniques have shed more light on this important issue than our forebears could have dreamed of. Before offering a sample of those, let us examine the validity of Pridmore’s thesis.

The first paragraph makes two points: that ‘Influential figures’ assert that suicide is ‘always or almost always the consequence of mental disorder’ and that funds have been predominantly directed to ‘medically orientated activities’.

With regard to the first point, he refers to a video clip of Insel (2013). However, Insel does not use the word ‘consequence’, but rather ‘contribution’, which, with respect, has a different meaning. Indeed, throughout the paper Pridmore uses no fewer than eight similes in prosecuting his thesis, including ‘consequence’, ‘cause’, ‘response’, ‘trigger’, ‘associated’, ‘predisposing’, ‘related’, and ‘result’ – several of which are reasonable, but, curiously, the arguably more accurate ‘contribution’ used by Insel is avoided.

The second point in relation to funding is not consistent with the data. A review by an experienced psychologist, Kerkhof (1999), of what was probably the first national program for suicide prevention in Finland, noted that: ‘The work has particularly influenced organizations and professionals in the social services sector, though not in the health sector as much as was initially hoped for’. Furthermore, a review of Australian initiatives by Robinson et al. (2006) concluded that: ‘Certain high-risk groups (particularly people with mental illness and people who have self-harmed) have been relatively neglected’.

Clearly the foundations of the paper are flawed, but what of the subsequent argument? It is curious that Pridmore should claim it to be an ‘advantage’ to have the views of a person ‘outside the medical system’. Such views should be respected, but if there may be an illness present, isn’t it logical to have the opinion of the best qualified person?

Unsubstantiated opinions abound, including those of Shahtahmasebi, whose polemical work is referenced twice. It is pertinent to reflect on this author, a graduate in statistics, who, after working in a number of UK hospitals, went to the antipodes and established his own foundation promoting a holistic approach to health. Indeed, his LinkedIn profile (2014) highlights his ownership of ‘Good Life Dressings’, which purport to lower cholesterol. This does not negate the potential importance of what may be written, but that information coupled with the hyperbolic nature of his writing should be considered carefully.

With regard to criticism of the psychological autopsy approach, of course there will be differences in approach and results. At one extreme there are even data which suggest that those who do not have sufficient symptoms to warrant a diagnosis prior to their death may be more similar to those with mental disorders than to a control group (Ernst et al., 2004), a paper understandably not referred to by Pridmore. At the other extreme is the review of Hjelmeland et al. (2012), quoted by Pridmore, which originates from a Department of Social Work, with none of the five authors being psychiatrically trained. Why not refer to the review of Hawton et al. (1998), which, with the benefit of a range of disciplines as authors, lends confidence to the procedure? It is also important to add that even if the figure were to be reduced markedly, that would not minimize the importance of delineating potentially treatable mental disorders.

Different suicide rates between different countries have long been recognized as related to socio-cultural factors … how could that not be the case? But the similarities between countries with similar socio-economic circumstances and analogous coronial systems suggest that there is some pervasive underlying contribution. This has been explored more fully elsewhere, with the suggestion that there may be a base rate of suicide predominantly related to psychiatric illness factors, with the marked variations observed between, and also within, different countries due to diverse psycho-social stressors (Goldney, 2013).

The statement that Durkheim was ‘the first sociologist’ will be a surprise to sociologists, as Durkheim was born 20 years after the discipline was mooted by Comte in 1838, when he adapted the meaning of the existing word ‘sociologie’ to describe what is the present understanding of the discipline (Wikipedia, 2014). Other mid-19th-century contributors have been reviewed elsewhere (Goldney et al., 2008), demonstrating that there were a number who addressed suicide in a sociological manner. It is of interest that the German Karl Marx wrote of the ‘critical descriptions of social conditions’ by French writers in relation to suicide, and the Norwegian Eilert Sundt noted of suicide that ‘if there is responsibility then it does not only rest on the individual who committed the act, but also on society’.

It is also evident that Pridmore’s concern about the purported uncritical presumption of the primacy of mental disorders in suicide does not extend to his own examples. How can a clinician’s experience of people with psychotic depression ‘prove’ that the mental disorder was ‘a sufficient cause of suicide’? And how can suicide be the result of ‘a single socioeconomic stressor’ in the absence of anything else? Perhaps use of the word ‘contribution’ may have been prudent.

But enough of a direct critique. What of evidence suggesting the importance of psychiatric illness?

Putting aside the psychological autopsy findings, although it is pertinent to re-iterate that even lower figures of mental illness mandate appropriate action, let us examine the implications from several other approaches. These include large population-attributable risk studies, genetic research, and work examining brain function.

The population risk statistic (PAR) is singularly suited to examining the relative importance of various factors associated with morbidity. This can be illustrated by considering its application to lung cancer. Although it is generally accepted that smoking causes lung cancer, the PAR shows that the elimination of all smoking would only prevent about 60–80% of lung cancer. Different illnesses and factors presumed to be associated with them can be examined in a similar fashion. For example, considering suicide as analogous to lung cancer, and calculating the relative importance of various factors that have been demonstrated to be associated with suicide.

Large population databases in Denmark have allowed Qin et al. (2003) to utilize the PAR statistic in an examination of the risk of suicide in relation to socioeconomic, demographic, psychiatric, and familial factors. They found the PAR for suicide of ever having had a mental disorder necessitating hospital admission to be 40.3%. This figure is overwhelmingly greater than the other statistically significant contributors. Thus, unemployment, having a sickness-related absence from work, being in the lowest quartile income and being on a disability or aged pension had PARs of 2.8%, 6.4%, 8.8%, 3.2%, and 10.2%, respectively. While it would be unrealistic to anticipate that our treatments would ever prevent all admissions, there is at least more likelihood that a clinical approach would be effective compared with attempting to influence the other factors. Nevertheless, they are important in their own right. There have also been a number of other PAR studies which have given a similar clear indication of where most leverage could be gained.

It has long been recognized that inherited factors can be associated with suicide. Indeed, the aforementioned Charles Moore wrote in 1790: ‘this extreme dejection of spirits, this melancholy, this lunacy and propensity to suicide, like many other disorders, is not confined to the unhappy object in the first instance, but by attacking successive generations of the same family proves itself to be hereditary’ (Goldney et al., 2008). Such early observations have been confirmed by modern research methodologies, even to the extent of estimating the relative importance of inherited as opposed to socio-cultural factors. This was demonstrated well in the landmark study of Statham et al. (1998), the results of which are worthy of wide dissemination.

They examined 5995 respondents from an Australian twin registry and took into account twin pair zygosity and their history of suicidality. They were able to allow for a range of socio-cultural variables including job loss, loss of property or home, child sex abuse, rape or physical assault, divorce, alcohol dependence, religious affiliation and the trait of neuroticism, and found that: ‘Overall, genetic factors accounted for approximately 45% of the variance in suicidal thoughts and behaviour’. A reasonable assumption from these results would be that 45% of the variance would be mediated by genetically determined biochemical changes related to psychiatric illness, and there is a developing literature in this area.

Useful reviews of the biology of suicidal behaviour have been provided by Currier and Mann (2008), whose comprehensive work addressed ‘Stress, genes and the biology of suicidal behavior’, and more recently by Turecki (2013), who focused on the role of polyamine dysfunction. These reviews illustrate the potential links between socio-cultural stressors and suicidal behaviours, and, although the findings are yet to be translated into clinical practice, they do serve to raise awareness of possible inherited factors in some families. They also serve as a powerful reminder of the bio-medical contribution to suicide.

A further approach to this issue is to reflect on the adequacy of care received by many who have died by suicide. There is an extensive literature commenting on this (Goldney, 2013), although we cannot assume that adequate treatment would necessarily prevent suicide. However, it is sobering to reflect that two comprehensive reviews of the management of those who died by suicide independently concluded that about 20% of those deaths could have been prevented, but for a number of factors (Burgess et al., 2000; National Confidential Inquiry, 2001). These included inadequate assessment and management of depression and other disorders, poor staff–patient relationships and lack of continuity of care. The clinical implications are self-evident, and to assert that there is a ‘faulty connection’ between mental disorders and suicide runs the risk of exacerbating such findings.

Pridmore makes a plea for ‘non-clinical people’ to be involved, and, as could be seen from the reviews of two national programs of suicide prevention, this has already occurred, arguably to the exclusion of the more clinical approach. It is also pertinent to note that the International Association for Suicide Prevention, which has been in existence for over 50 years, is comprised of many different professions, volunteers and survivors (Goldney et al., 2013), as is the local organization, Suicide Prevention Australia.

Debate in this area is welcome, but it has developed beyond simplistic observations. Contemporary research has confirmed and placed in perspective historical views of the importance of both illness and socio-cultural factors. There should be no dichotomy between the contribution of mental illness as opposed to that of socio-cultural factors. Ignore one or other at your peril, and do not denigrate those with expertise honed through specialized training. Indeed, the psychiatric profession is in a unique situation to apply hard won research findings and meet its responsibility, not only in acknowledging our shortcomings, but more specifically by delineating and treating those with mental disorders which may contribute to suicide.