Mental disorder and suicide: A faulty connection,or a faulty argument?

In this edition, Professor Pridmore continues his important series addressing the complexity of suicidal behaviour (Pridmore, 2015). He argues that the link between mental disorders and suicide has been systematically overstated. As a result of this ‘faulty connection’, he argues, suicide prevention efforts have focused too narrowly on mental disorders at the expense of broader strategies. Pridmore provides a persuasive argument about the importance of social factors in suicide. However, social and disease models of suicide are not incompatible. It is not necessary to prove that a connection between mental disorders and suicide is ‘faulty’ to accept the critical role of other factors.

Pridmore first takes aim at claims that suicide is nearly always due to mental disorder. He cites the National Institute of Mental Health (NIMH) chief Thomas Insel’s 2012 TEDx talk on early intervention, which (in passing) links 90% of suicides to mental disorders. Insel appears to draw this figure from the Institute of Medicine’s (IoM) 2002 suicide prevention report (Goldsmith et al., 2002), which offered a more balanced view than might be guessed. It cited findings that 90% of suicides in the USA were associated with mental health and substance use problems. However, it also argued that these are ‘not always the greatest risk factors for suicide’ (p.69), that cultural factors are also important, and that psychological autopsy studies from China and India report lower rates of mental disorder. It is likely that Insel would acknowledge the points that Pridmore is making.

Pridmore then argues against suicide prevention policies focused exclusively on mental disorders. However, no such policies are cited in the editorial. The IoM suicide prevention report (Goldsmith et al., 2002) considered psychological, biological, social, subcultural and cultural risk factors for suicide, and made recommendations much broader than merely diagnosis and treatment of mental disorders. Australia’s Living Is For Everyone (LIFE) suicide prevention framework considers both clinical and non-clinical dimensions. As well as early diagnosis and treatment of mental disorders, especially in high-risk groups, its strategies include improving the understanding of community risk and protective factors, improving the resilience of individuals and communities, and detecting and responding to suicidal thoughts and behaviour. A recent overview of suicide prevention strategies (Christensen and Petrie, 2013) considers mental disorders as only one factor alongside social and cultural causes of suicide. Christensen argues that the provision of mental health care is one of three evidence-based public health interventions for suicide prevention, but one with significant limitations.

Finally, Pridmore positions himself against what he terms a ‘medical model’, which sees psychiatric and substance use disorders as a sufficient cause of suicide. He offers four arguments against this model. First, between-country differences in the rate of suicide are greater than between-country differences in the prevalence of mental disorders. Second, suicide rates are higher in men, but affective disorders are more common in women. Third, social and cultural factors influence suicide rates. Fourth, researchers from other disciplines are ‘scathing’ about the medical model. These seem to be arguments against a slightly caricatured view of causation. Even when considering a single group of disorders, such as psychoses, it seems clear that there is no single cause, but many interacting ‘component causes’. Rates of psychosis differ by age and sex, and are influenced by cultural and environmental factors including migration, urban location, substance misuse and trauma. Some oppose a medical view of psychosis. However, none of these are arguments that psychoses cannot validly be seen as mental disorders.

Pridmore cites the high suicide rate in Lithuania, writing: ‘For mental disorder to be the paramount factor in suicide, the people of Lithuania would need to have three times the psychopathology of the people of Australia … This is not the case, and the initial premise is incorrect’. This argument assumes minimal between-country variation in the prevalence of mental disorders. However, there is considerable between-country variation in rates of many mental disorders, including a three- to fivefold variation in rates of psychosis. The rate of binge drinking in Lithuania is more than three times that of Australia (World Health Organization, 2014). Eastern and central European countries have rates of depressive disorder up to twice those observed in Australasia (Ferrari et al., 2013). Therefore, it is possible that between-country differences in the prevalence of mental and substance use disorders contribute to some between-country variation in suicide rates. More importantly, the argument assumes that if mental disorders were ‘paramount’ in causing suicide then there would be a simple linear relationship between mental disorder prevalence and suicide rates. Why should this be so? Even where there are clear causal links between risk factors and mental health outcomes, these associations are rarely linear. For example, psychotic disorders are clearly a ‘paramount’ cause of the burden of psychosis. However, considering between-country differences, there is only a modest association between the prevalence of psychosis and the burden of psychosis (Ayuso-Mateos, 2002). Factors such as imprecision in estimates, differences in population age structures, differences in social and community supports, and access to effective health services all mediate this relationship. Similar factors are likely to mediate any ecological association between the prevalence of mental disorder and suicide rates. Therefore, even if mental disorders were an essential condition for suicide, the relationship between disease prevalence and suicide rates would be likely to be non-linear.

This commentary does not argue that mental and substance use disorders are the only cause of suicide. Indeed, it is not entirely clear who does argue for such a position. Many would agree whole-heartedly with Professor Pridmore’s conclusions that social and environmental factors play a critical role in suicide. All government policies which influence social capital and well-being have the potential to also influence suicide: economic, industry, environmental and welfare policies are all also ‘suicide policies’. Government decisions on issues ranging from welfare thresholds and parental supports to the detention of refugees all have the potential to influence suicide in individuals, or suicide rates in communities. They are likely to act through many pathways, including by influencing the prevalence or impact of mental disorders. While the connection between mental disorders and suicide may be incomplete and complex, it is probably not ‘faulty’.

See Viewpoint by Pridmore, 2015, 49(1): 18–20.