Said Shahtahmasebi, Ph.D.

You know, when I was at primary school and subsequently high school my teachers used to tell me and other vertically challenged students to play basketball so we would become tall. Well, because I was young and because society at that time had faith in authority, in particular, teachers and doctors, I believed them. I tried to get on basketball teams but no one would have me because I was too short!  This is known as ‘selection bias’ where it is not the game that causes children to grow tall, rather, it is the nature of the game that seeks out the selection of tall people.

Studies of human behaviour inevitably carry bias due to its dynamics, e.g. due to omitted information, feedback effect and temporal dependencies, which become too complex with the addition of selection bias and measurement error. Frequently, studies of human behaviour fail to account for bias.

Suicide research is even more complex because the main informant (suicide case) can no longer provide important insight into their decision to end their life. Therefore, in most cases, suicide research is designed within a mental illness framework with depression as its main parameter and has attempted to use various techniques to gather information about the mental status of the deceased.

Therefore, it is not surprising that research does not provide objective evidence, instead, practice and policy making is led by mindsets and beliefs.

There is a belief amongst the medical profession and the public that depression causes suicide. When this belief is challenged the medical profession quote that 80-90% of suicides had depression. Where did this estimate come from and what is it based on? If it is true, why then have we not observed a drop in the suicide rate given that prescriptions for anti-depressants have increased sharply over the last 12-15 years? For example the New Zealand government’s documents report that prescriptions for antidepressants had doubled by 2006, and a more recent document reports that it has doubled again since 2006, i.e. the rate has quadrupled over the last 12 years. Yet, suicide rates have maintained an increasing trend over the same period (1).

In New Zealand it is estimated that roughly 1 in 6 people will suffer from serious depression at some time in their life (http://www.depression.org.nz/depression). In other words, at any one point in time in New Zealand we can expect over half a million people to be suffering from depression. If the 80-90% probability is applied to the population of those with depression in New Zealand then we should expect thousands and thousands of suicides every year. Even if we apply the inverse of 90% (i.e. 0.01) to the population of those suffering from depression we could expect over 5500 cases per year. Yet, on average the number of suicides in New Zealand’s is 540 per year which is 540 too many. However, this suggests a crude suicide risk of 0.00014 (or roughly 14 per 100,000 based on a total population of 4 million). The main depression website (http://www.depression.org.nz/depression) which is also part of the New Zealand Government’s suicide prevention strategy states that depression increases the risk of suicide 20 fold. Again a crude calculation suggests (0.00014*20*500,000=) 1400 expected suicides per year. Tackling depression has been central to the New Zealand Government’s suicide prevention strategy with the launch of www.depression.org.nz about a decade ago and the quadrupling of antidepressant prescriptions. Over the same period suicide rates have maintained an upward trend.

Similarly, in Australia national surveys show that 20% of the general population experience significant mental problems each year and the overall suicide rate is about 10 per 100,000. Thus the likelihood of a person with a mental illness taking their own life is low. It is higher for  some mental disorders such as bipolar disorder and schizophrenia than others but it also high for those with alcohol and drug problems and addiction, chronic pain, debilitating conditions and so on (2).

So where does the figure of 80-90% of suicide cases having depression come from?

Some argue that this figure is based on reviewing suicide cases using medical records. If 90% of suicide cases had a recorded diagnosis of depression then they would have been receiving treatment for it. Surely, if depression is the root cause of suicide at least a small proportion of the 90% should have responded to treatment. 

On the contrary, not only between two-thirds and three-quarters of all suicides do not come into contact with psychiatric services (3, 4), but also of those who do have a psychiatric record not all have depression or a mental disorder as a diagnosis. For example, in the UK, following a confidential inquiry into homicides and suicides by mentally ill people – data from medical and hospital records were collected on all suicide cases who had been through a community and mental health Trust (4).The findings from the confidential inquiry revealed that 33% had no diagnosis, 17% had depression either as a diagnosis or mentioned in their hospital notes, followed by 12.5% with schizophrenia, 8% and 6% alcoholism and personality disorder respectively. Furthermore, the cases from the hospital formed only about one-third of all completed suicides, i.e. two-thirds of all suicides had had no contact with psychiatric services and were successful in their first attempt. Interestingly, for 46% of cases the reason for coming into contact with psychiatric services was due to previous attempts – yet the individuals still completed suicide.

So we are still none the wiser as to where the figure that 80-90% of all suicides had depression has come from.

But, the idea that depression is the cause of suicide is well established in the public mindset. In other words, depression and a mental disorder is readily assumed after the event of suicide. For example, a GP while giving evidence at a coroner’s inquest in New Zealand in 2005 stated: “I am desperately sad we had no insight into his mental health problem and so were not able to prevent this tragedy.” Which is surprising given that the young person had been described as a happy and popular person with no sign of health problems and no evidence of mental ill-health. In the GP’s mind, selection bias dictated that the young individual must have had mental problems and depression to commit suicide, i.e. we force suicide to fit into the model.

Another example is the case of an Australian celebrity (http://www.stuff.co.nz/4262547a1860.html) who following treatment for depression (was prescribed antidepressants) and was making future plans committed suicide. After the event (suicide) occurred, the psychiatrist’s explanation was that cases with deep depression are good at hiding their feelings and intentions. Once again, in the psychiatrist’s mind, selection bias dictated that nothing other than depression, in this case deep depression (because the case had earlier been treated for depression), could have caused her suicide.

Given this mindset, it is surprising that governments are happy to fund researchers to seek information about the mental status of suicide cases from third parties (i.e. family and friends), e.g. the Canterbury Suicide Project (5). These types of suicide research (psychological autopsies) are flawed theoretically, methodologically, and analytically leading to erroneous results and mis-conclusions, e.g. depression and mental illness are the main causes of suicide, and, 80-90% of suicide cases had depression, e.g. see (6). There is no way to ascertain these claims because the main informant (suicide case) is no longer able to give their reasons for wanting to die.

Clearly, not much reliance can be placed on the results from psychological autopsies, yet, as mentioned earlier, tackling depression is central to the Government’s suicide prevention strategy.

For this strategy to work there has to be a real link between depression and suicide. For the reasons explained above there is no statistical evidence to support the conclusion that depression leads to suicide. If there was a link between depression and suicide a big drop in the suicide rate would have followed the NZ Government’s suicide prevention strategy. But, over the years, cyclic patterns in suicide rates have been mistakenly credited to a working strategy when the cycle is on a down-turn, and when the cycle is on its way up complexities in suicide has been blamed.

Poor government policies based on poor research and inappropriate information has created a vicious circle. In other words, with a government and health profession focus on a mental illness/depression model which will continue to provide more of the same prevention/intervention each year at a much higher cost in terms of lives lost and resources – all because it cannot be anything else but depression. As a result we do not know anything about suicide and hence we cannot prevent it. In the meantime, at least a proportion of suicide cases will die needlessly because of our obsession with mental illness and our refusal to address and understand suicide.

References

1.    Shahtahmasebi S. De-politicizing youth suicide prevention. Front Pediatr 2013;1(8).
2.    Webster I. Appreciation of your recent article. Personal communication 2013.
3.    Hamdi E, Price S, Qassem T, Amin Y, Jones D. Suicides not in contact with mental health services: Risk indicators and determinants of referral. J Ment Health 2008;17(4):398-409.
4.    Shahtahmasebi S. Suicides by mentally ill people. ScientificWorldJournal 2003;3:684-93.
5.    Beautrais AL, Joyce PR, Mulder RT. The canterbury suicide project: Aims, overview and progress. Community Mental Health in New Zealand 1994;8(2):32-9.
6.    Cash SJ, Bridge JA. Epidemiology of youth suicide and suicidal behavior. Curr Opin Pediatr 2009;21(5):613-19.