This Week in Psychiatry — Oct 12, 2026

This AudioScholar briefing is generated by artificial intelligence for healthcare professionals and trainees. It is not medical advice.

Welcome to This Week in Psychiatry. This week we're covering 10 notable papers spanning suicide risk and psychological treatment in young people, the uneven road to precision psychiatry, and neurodevelopmental conditions across the lifespan. Let's dive in.

We start with suicidal thinking and self-harm in young people, where two papers look at who goes on to act and whether a phone app can help. In the British Journal of Psychiatry, Morse and colleagues followed roughly 3,300 students at five universities in the United Kingdom through autumn and spring terms between 2021 and 2024 [1]. Just under half of students reported lifetime suicidal ideation at baseline, and about one in ten reported a past attempt. Lifetime ideation was linked to minoritised sexual orientation or ethnicity, male or gender-diverse identity, prior mental health treatment, and adverse childhood experiences, so most of the risk was present before students arrived on campus. Ideation was fluid: about six in ten students with prior ideation no longer reported it by spring, while new-onset ideation appeared in a few percent of participants. The striking contrast was in attempts. Among students with ideation alone, about one in a hundred made an attempt during the year, compared with about one in ten among students with a previous attempt. Rising depressive symptoms and rumination tracked with new attempts. The authors conclude that ideation by itself is a weak marker of who will attempt, and that prior attempt history and changing symptoms carry more signal. This is observational survey data, but it fits a growing consensus that screening on ideation alone is poorly specific. From the Journal of Child Psychology and Psychiatry, Liu and colleagues randomised 122 Chinese adolescents with depressive episodes to a safety planning app called LIFES plus usual care, or usual care alone, with six months of follow-up [2]. The app group showed greater reductions in non-suicidal self-injury frequency, seriousness of suicide planning, suicide threats and gestures, and total suicidal behaviour, after adjusting for depression, anxiety and sleep, and users reported good satisfaction. The trial is small, single-country, and reports interaction effects rather than clear absolute differences, so it is early evidence that digital safety planning can work as an adjunct, not yet proof that it should be generalised.

Our second theme is what actually drives benefit in psychological interventions for children and families. In Psychotherapy and Psychosomatics, Li and colleagues pooled 48 randomised trials with about 6,000 children and adolescents exposed to complex trauma, using a component network meta-analysis to pull treatments apart into their ingredients [3]. Trauma-focused approaches came out on top: multimodal group interventions, trauma-focused cognitive behavioural therapy with or without parental involvement or a phase-based structure, and eye movement desensitisation and reprocessing all reduced post-traumatic stress symptoms, with large effects for several. When the components were separated, exposure was the only individual element that significantly reduced symptoms. Affective modulation added to psychoeducation was an efficient combination for stress symptoms, and exposure plus parental involvement looked efficient for depression. This bears directly on the persistent worry that children with complex trauma are too fragile for exposure; the pooled evidence supports exposure as the active ingredient, although component analyses depend on how trials were coded and the precision of some estimates is wide. A more sobering picture comes from a three-arm trial in the Journal of Child Psychology and Psychiatry, led by Gonzalez, which randomised about 500 caregivers of children aged two to six in Ontario to online Triple P, online Circle of Security Parenting, or usual services [4]. Triple P, a social learning program, produced short-term reductions in emotional symptoms and conduct problems and improvements in hostile parenting, but most of those differences faded after the program ended, and the effect on total child behaviour problems at twelve months did not survive correction for multiple testing. Only improved parental attributions persisted. The attachment-based Circle of Security showed no measurable effect against usual care. Notably, families receiving usual services also improved. Read alongside the trauma meta-analysis, the common thread is that structured, skills-based behavioural components seem to carry the benefit, while the durability of community-delivered online parenting programs remains an open question.

The third theme is precision psychiatry, where three papers show both the promise and the gap between a genetic signal and a clinical tool. In Molecular Psychiatry, Byrne and colleagues combined two Brisbane birth cohorts to ask whether polygenic risk for depression helps sort out which distressed teenagers go on to major depression [5]. Teenagers in the top tenth for internalising symptoms at age 14 had roughly two and a half times the odds of lifetime depression by young adulthood, and those in the top tenth of polygenic risk had about three and a half times the odds. Adolescents in the top tenth on both measures had roughly eight times the odds compared with peers high on neither. The authors argue polygenic scores may be most useful within already symptomatic adolescents. The cohort is modest, with about 300 cases, the estimate for the combined group is imprecise, and the sample is of European ancestry, so this is a proof of principle rather than a screening tool. In the British Journal of Psychiatry, Sharew and colleagues used the International Consortium on Lithium Genetics, about 2,400 patients with bipolar disorder treated for at least six months, to build a multi-trait polygenic score from 60 phenotypes [6]. The combined score explained roughly six to nine percent of the variation in lithium response, with accuracy of about 68 percent in separating good from poor responders, and psychiatric and behavioural traits contributed most. That is a substantial improvement over single-trait scores, but still well short of what would justify withholding or prescribing lithium on genetics alone. Both papers land neatly in the framework set out by Baune in a Personal View in The Lancet Psychiatry [7]. Baune proposes a Translation Evidence Map that stages precision claims from Level 0, a defined use case or discovery, through to Level 6, post-deployment monitoring, tying each level to measurement reliability, generalisability, decision value, safe delivery and proportionality to risk. The Map is meant to sit alongside, not replace, GRADE and regulatory frameworks, and to make explicit the highest claim a given body of evidence can defend. By that yardstick, the depression and lithium scores remain early-stage discovery and validation work.

Our final theme is neurodevelopmental conditions across the lifespan and how services are organised around them. The headline paper is in JAMA Psychiatry, where Bauer-Negrini and colleagues drew on the United States All of Us Research Program, about 187,000 adults aged 50 and older with both health records and genomic data, followed for an average of about ten years [8]. Among people with an ADHD diagnosis, about seven percent developed mild cognitive impairment or dementia, roughly double the rate among those without ADHD, and in adjusted models ADHD was linked to more than a fourfold higher hazard of cognitive impairment. Adjusting for comorbidities, education and health care use weakened but did not erase the link. Importantly, polygenic risk for ADHD showed the same direction, with the highest fifth carrying about a third higher hazard than the lowest, which helps address worries about diagnostic ascertainment. Among people who did develop cognitive impairment, those with ADHD had worse mental health, quality of life and social satisfaction, and used more health care. This is observational and cannot show that ADHD causes dementia, nor whether treatment changes the trajectory, and only a small minority of participants carried a diagnosis. Still, convergent clinical and genetic evidence strengthens the case that adult ADHD is a late-life cognitive risk marker. Nature Reviews Disease Primers offers a comprehensive update on Tourette syndrome from Eapen and colleagues [9], framing it as a disorder of cortical-striatal-thalamic-cortical circuitry with shared genetic risk explaining frequent ADHD and obsessive-compulsive comorbidity. The review describes education, psychosocial support and behavioural therapy as the foundation, with medication for severe or unresponsive tics and deep brain stimulation reserved for refractory adult cases, and flags the need to delineate functional tic disorders. Finally, in the British Journal of Psychiatry, Thapar, Cooper and Cortese argue that separating neurodevelopmental from psychiatric services is itself a clinical hazard [10]. Given high comorbidity, siloed services risk fragmented care, misdiagnosis, deskilled clinicians and longer waits, and the authors call for more person-centred systems. It is an opinion piece, but the ADHD and Tourette papers this week provide a concrete illustration of why the comorbidity argument matters.

If you only have time for one paper this week, make it the JAMA Psychiatry cohort on ADHD and late-life cognitive impairment [8]. It moves the question of whether adult ADHD is a genuine dementia risk marker beyond diagnostic artefact by pairing clinical diagnoses with genetic liability in the same people, while leaving open whether treatment alters that risk.

Here is what this week's evidence adds up to in Psychiatry. First, in university students, prior suicide attempts and worsening depression and rumination carried far more signal for future attempts than ideation alone, though this rests on observational survey data. Second, pooled trial evidence points to exposure as the active ingredient in treating complex trauma in young people, while a well-run community trial suggests online parenting programs deliver mostly short-lived gains. Third, polygenic scores for depression risk and lithium response are improving but remain at the discovery and validation stage, and a new evidence map offers a way to judge how far such claims can go. Fourth, adult ADHD is now linked, by both clinical and genetic evidence, to a substantially higher risk of late-life cognitive impairment and a heavier burden once impairment develops, though causality and the effect of treatment are unsettled. And finally, a single small trial supports digital safety planning as an adjunct for depressed adolescents, pending replication in broader populations.

That's your roundup for This Week in Psychiatry. The full transcript and references are available on the episode page. This is an AI-curated summary — for clinical decisions, always consult primary sources and current guidelines. See you next week.

If this weekly briefing is useful, follow the show in your podcast app so new episodes arrive automatically. And a quick rating on Apple Podcasts or Spotify helps other physicians discover the show.

This is an automated summary generated by artificial intelligence, which can make mistakes. Always review the original source materials.