Highlights of this issue

In man's life, his time is a mere instant, his existence a flux¹

Self-harm and suicide are a major theme in this month's journal. Cassidy et al (pp. 683–691) try to estimate the prevalence of autism, both diagnosed and undiagnosed, in those who died by suicide in England between 2014 and 2017. Individuals with autism have significantly greater rates of suicidal ideation than the general population. It has been postulated that risk might be increased via greater vulnerability to adversity, feelings of not fitting in society and cognitive inflexibility impairing the ability to see a way out of significant challenges. Although research suggests that a little over 1% of UK adults meet criteria, many go through life without receiving a diagnosis. To undertake such complex and nuanced work, the authors adopted a less common method of evaluating the records from coroners’ inquests (N = 372), following up with 29 interviews with next of kin where relevant clinical markers had been uncovered. From the initial trawl of the inquests of deaths by suicide, evidence of autism was about ten times that of the general population. Over 40% of the follow-up sample were considered to have elevated autistic traits. The authors propose that we should be better at providing quick autism screening in those presenting with suicidal thoughts, and that there should be tailored suicide prevention programmes.

Effective treatment of a mental illness is clearly a key way to help mitigate self-harm and suicide. Fitzgerald et al (pp. 692–700) test whether there are relative differences between medications in the treatment of bipolar affective disorder, evaluating a Danish population cohort from 1995 to 2016, over 30 000 individuals with the condition. This helped to circumvent the issue of randomised controlled trials frequently excluding individuals with suicidal ideation. They found that the use of any medication was associated with fewer psychiatric admissions, but there were clear and significant between-drug differences. Any guesses? Yes, lithium continues to outperform all others. Its use was associated with lower rates of self-harm and hospital admission, notably also in within-individual analyses, and it had the lowest rate of suicide. I've written before about the baffling decline of lithium when almost all data strongly support its effectiveness across a range of parameters. Too ‘old fashioned’, a lack of promotion (opening the hornets’ nest on the effectiveness of pharmacological marketing to doctors) or something else?

Although suicide was decriminalised over half a century ago, the persisting (though declining) phrase ‘commit suicide’ continues to infer criminal behaviour. Alex Thomson, always an insightful commentator on contemporary frontline emergency department psychiatry, writes with colleagues (pp. 653–654) to ask why criminal sanctions persist for suicidality in 21st century Britain. Data on the topic are relatively sparse, often pulled from media reporting or individuals’ accounts. However, they have included charges for breach of the peace, criminal damage and wasting police time, with outcomes including restrictive bail conditions and antisocial behaviour orders, which have included conditions forbidding a person from disclosing that they are suicidal. The interfaces between mental health services, the police and the criminal justice services are often complex but here to stay: a wider examination of this is called for.

All things of the body stream away like a river, all things of the mind are dreams and delusion¹

The second major theme of this month's BJPsych is post-traumatic stress disorder (PTSD). Forehand et al (pp. 676–682) bridge these topics with a paper on PTSD severity and death by suicide in a cohort of over three-quarters of a million US military veterans. Their retrospective study is the first to add symptom severity to the more binary aspect of having or not having such a diagnosis. Any symptoms breaching remission criteria were associated with a doubling of suicide mortality 1 month following assessment, though this abated with time. This finding held even after adjusting for all health comorbidities. Those with moderate-to-high symptoms did indeed have greater longer-term risks of such deaths, and, crucially, clinical improvement to the point of remission lowered this. Military veteran populations with PTSD have been shown to have notable sociodemographic and clinical characteristics that delineate them from civilian cohorts; the generalisability to the latter is not yet clear.

Saito et al (pp. 668–675) take us to the contemporaneously relevant issue of PTSD developing after assisting others during traumatic events. Here, they explored the trajectories of first-responders to the 2011 Great East Japan earthquake, the largest in that country's history, which led to both a tsunami and a nuclear disaster. Over 55 000 Ground Self-Defence Force personnel were followed up over 7 years, with nine potential risk factors and any PTSD symptom severity mapped. Most first responders did not develop any (or at least very few) PTSD symptoms. The authors identified five trajectories: resilient, recovery, incomplete recovery, late-onset and chronic. About 10% suffered the most severe forms, the main risks for these so-called ‘non-resilient’ outcomes being: older age, personal disaster experiences and working conditions. Regarding working conditions, there were some that were relatively specific to this type of disaster – namely body recovery and radiation exposure risk – but the others – longer working times and a lack of leave – may resonate with you. The authors note that their findings could inform policies for prevention and early detection. I agree, though the language of having staff defined as ‘resilient’ or ‘non-resilient’ continues to place the problem in the person rather than the environment.

There is a general lack of exploration of harm from psychological interventions. Indeed, in contrast to pharmacological trials, this is not a mandated research outcome point, though work in this journal has previously shown that, more broadly, it is not a rare outcome.² It is pleasing therefore to see Hoppen et al (pp. 658–667) from the University of Münster (my favourite German city) review the safety of psychological interventions for adults with PTSD. Across 56 randomised controlled trials encompassing over 4000 patients, a headline finding was that just over one-third reported on harms at all, with data on putative causes even rarer. Nevertheless, where this was recorded, such treatments showed generally low incidence of harms, notably being less than those of passive and active control groups. Among interventions, trauma-focused cognitive–behavioural therapy had the fewest serious adverse events. I was glad to see both continued growth in the underexplored area of potential harms from therapies and the reassuring results of such low risks here. It's a topic taken up by this month's Kaleidoscope (pp. 709–710) as it looks at treatments for hypochondriasis, as well as asking what makes human brains unique compared with our hominin sister species of Neanderthal.