Suicide prevention: Time to change the paradigm

There is perhaps no more emotive topic than suicide. It is among the most important determinants of mortality in medicine, particularly since it is now the most common cause of death from youth until middle age (Sinyor & Schaffer, this issue). And yet, the branch of medicine charged with trying to reduce suicide rates is Psychiatry. Much has been done to tackle suicide, but as Jorm (this issue) points out there has been a notable lack of impact of efforts to prevent suicide in Australia. He attempts to explain this by pointing out three factors. The first is that suicide is influenced by social factors that are largely outside the domain of mental health services. While this perspective has been accepted since Durkheim, it appears to make little difference to expectations imposed on mental health services. The second, and we need to face this fact, is that psychiatric treatments may have limited impact on suicide deaths. There is consistent moderate strength evidence for the benefit of lithium (which, ironically, is now used about half as often as it was 20 years ago; Rhee et al., 2020) and a little evidence for the benefits of brief intervention and contact but not much else. The third factor is that mental health professionals are often not present when suicide risk is highest. Suicidal actions are sometimes impulsive, particularly in males. These factors again emphasise the need for coherent national policies which recognise the role of broader social factors in suicide and do not expect mental health services to be the primary solution.

Sinyor and Schaffer (this issue) approach the lack of success in reducing suicide in a different way. They ask what other areas of medicine would do to approach this major challenge. They point out that our concern with emphasis on prediction of suicide diverts already scant resources away from a more important potential area, namely, intervention studies. They note how strange it is to exclude suicidal patients from research trials due to concerns that some of them may die while in the trial and point out that other branches of medicine, such as oncology, do not exclude patients at increased risk of death from their clinical trials; hence, they advocate that psychiatry reverses this practice. They go on further to point out that our preoccupation in developing models and algorithms in the hope of predicting risk are likely to fail since medicine has no history of reliably predicting sentinel events. We do not expect our colleagues to predict exactly when a coronary artery occlusion or a stroke will occur but seem to expect that we can accurately predict suicide which is generally the result of a more complex process. Since so-called ‘low risk patients’ account for the majority of suicides at a population level, they advocate offering good care to all but also to redirect and enhance treatments and resources for patients at high-risk times.

Turner et al. (this issue) expand on these issues and look closely at health care settings. They also note the need for a shift away from the current paradigm with its focus on assessment and categorical risk prediction, resulting in a lack of focus on meaningful interventions and a culture of blame. Looking through the lens of the clinician, they point out ‘the inconvenient truth’ that there is a second victim of patient suicide. This victim is the health care provider involved in this adverse event and subsequently often emotionally traumatised. Mental health staff experience a range of adverse outcomes and also fear legal retribution, prejudice and loss of reputation. As has been repeatedly pointed out over the past decade, risk stratification cannot predict individual suicides. Despite this overwhelming evidence, documentation used in clinical practice continues to support a categorical risk prediction model. We need to acknowledge that we are powerless to usefully classify individuals or groups of patients according to future suicide risk. This second ‘inconvenient truth’ lies at the heart of clinicians being judged in incident reviews after the loss of a patient to suicide.

The authors further note the problem of hindsight bias. Reviewers who are aware of an outcome will overestimate the likelihood of that outcome and of the ability of the involved clinicians throughout to predict it. They also point out that current retrospective reviews do not take into account the complexity of how clinical incidents occur and what perpetuates them. They point out that clinicians work in complex systems which are unpredictable and do not conform to the linear expectations and mechanistic thinking that often drives service improvement efforts. Then just as one is feeling increasing despair the authors focus on the implications of these arguments for incident review processes and root cause analyses. They offer a model they call a ‘Restorative Just Culture’ which can counteract the risk of the blame culture and system anxiety following a critical incident. They suggest it provides an ideal framework which can build trust among staff, mandate the involvement of all parties so that the complexities of work and suicide prevention can be better understood, and help to overcome pessimism and nihilism with respect to our ability to learn from and prevent suicides. It is being implemented at the Gold Coast University Hospital and Health Services and the lessons learnt from this are discussed. They conclude that a zero suicide aspiration demands a restorative justice culture as a necessary accompaniment.

Jackson et al. (this issue) contrast older aged and middle-aged adults that are admitted to hospital from self-poisoning. They point out the groups are similar although the older aged group had a greater proportion with cognitive impairment, higher medical morbidity and use benzodiazepines more often in the deliberate self-poisoning event. They suggest benzodiazepine prescription might be a focus of clinical attention during the inpatient stay and periods of follow-up.

Not everything in this issue of the Journal focuses on suicide and self-harm. Two studies examine techniques used to ‘look’ at the brain – that is, imaging. Zhang et al. (this issue) conducted a meta-analysis of 395 functional magnetic resonant imaging studies to identify depressive disorder–associated brain regions as regions of interest for non-invasive brain stimulation. They report that a number of brain regions present potential non-invasive brain stimulation locations for depressive disorders and may serve as a basis for further clinical investigations. Colloby et al. (this issue) investigate cortical thinning in different types of dementia. They report the pattern of structural change in dementia with Lewy bodies is similar to that in Alzheimer’s disease while Parkinson’s disease dementia appears to demonstrate an ‘intermediate’ degree of regional and global cortical thinning.

Two further articles look at innovative clinical practice. Mitchell et al. (this issue) evaluate imbedded psychiatric assessment and planning units in emergency departments (EDs). The model of care appears well accepted by consumers although feedback from ED staff was mixed. This may be related to a lack of understanding of the model. The units also had a mixed effect on ED flow. Fehily et al. (this issue) studied the effectiveness of imbedding a specialist preventive care clinician, in this case an occupational therapist with 12 years’ experience, in increasing preventive care provision. It is a large randomised controlled trial (RCT) and the first one to rigorously and comprehensively examine the effects of this intervention. The specialist preventive care had a significant effect on the provision of the majority of recommended elements of preventive care. The study was limited by low uptake and difficulties in follow-up leading the authors to suggest identifying strategies to try and increase retention rates.

Returning to broader issues, Burns et al. (this issue) report on a longitudinal study of Australia’s mental health over the past 17 years. They note very small changes in mental health which are most evident for the youngest and oldest individuals. And interestingly, there appears to be consistent evidence for better mental health with increasing age until you are very old when there are substantial declines particularly in males.

Overall, we once again have furnished you with a varied and interesting collection of papers, which provide a narrative for what we do. It is pleasing to see that some consensus around the prevention of suicide is emerging, but moving away from risk stratification to better clinical care and actually studying patients who are suicidal in clinical trials may be a way forward. Restorative Just Culture might encourage clinicians to work in this difficult and emotive area of practice and alongside testing clinical innovations in mental health services may help improve the care of all our patients. After all, improving general clinical care for all our patients must be our overarching ambition, not just tackling those with so-called high risk, and perhaps this will be the best contribution that mental health services can make to lower the suicide rate.