A recent Psychiatry digest,
as sent to subscribers

Study Type: Cross-sectional online survey (self-selected sample, Netherlands, 2023–2025)

Key Question: How do relatives' experiences, emotions, involvement, and views of the due care criteria differ between psychiatric and somatic medical assistance in dying (MAID)?

Key Findings:

• Of 616 relatives (110 psychiatric, 506 somatic), >80% judged the due care criteria positively in both groups, with no group difference. Over 25% felt insufficiently involved.
• After psychiatric MAID, the most strongly felt emotions were gratitude, resignation, relief and sadness (mean 3.5–4.2/5). Relatives reported greater powerlessness (p = 0.026) and guilt (p = 0.004) than after somatic MAID.
• 68.3% of psychiatric-MAID relatives found the trajectory too long, versus 21% for somatic MAID (p < 0.001).

Clinical Relevance: MAID is not legal in the UK, but this informs the ongoing debate on assisted dying legislation and may guide support for families of patients with severe, treatment-refractory mental illness who seek MAID abroad.

Limitations: Recruitment was self-selected (social media, patient federation), so results are vulnerable to selection bias and cannot be generalised, and the cross-sectional design precludes causal inference.

Study Type: Cross-sectional observational study (Australian Genetics of Depression Study; self-reported data plus polygenic risk scoring)

Key Question: Does a circadian subtype of depression, defined by ≥3 circadian-related features, show distinct clinical course, antidepressant response and genetic risk profile?

Key Findings:

• 23% of participants (n=2604) met circadian criteria. This group had earlier onset and more severe features (hypo)manic-like and psychotic-like experiences, suicidality, distress, somatic complaints and weight gain (all P<0.001).
• Self-reported response was poorer to SSRIs (OR 0.88, 95% CI 0.82–0.94) and SNRIs (OR 0.89, 0.83–0.97), with more side-effects.
• The subtype was associated with higher polygenic risk for ADHD, MDD, anxiety, bipolar disorder, BMI, IL-6, insomnia and later sleep midpoint (ORs ~1.08–1.15).

Clinical Relevance: Routine assessment of chronotype, social jetlag and seasonality may help identify a more severe, treatment-resistant depressive phenotype, though no alternative treatment is tested here.

Limitations: Cross-sectional, self-reported design (including medication response) precludes causal inference, effect sizes are small, and findings await independent replication.

Study Type: Nationwide register-based cohort study with interrupted time-series analysis (Denmark).

Key Question: Do the metabolic benefits of GLP-1 receptor agonists seen in schizophrenia RCTs translate to routine clinical practice?

Key Findings:

• Among 3,042 people with schizophrenia starting a GLP-1RA, those treated for diabetes showed post-initiation reductions of 11.4% in HbA1c, 6.2% in LDL and 14.8% in triglycerides (all p < 0.001).
• Those treated for overweight showed reductions of 3.4% (HbA1c), 4.9% (LDL) and 12.7% (triglycerides) (all p < 0.001).
• LDL and triglycerides were used as proxies for weight loss; weight itself was not measured.

Clinical Relevance: Supports real-world use of GLP-1RAs to address cardiometabolic risk in people with schizophrenia, relevant to NHS physical health monitoring and the persistent mortality gap in this population.

Limitations: No control group and no direct weight or BMI data, so changes cannot be causally attributed to GLP-1RA treatment and the proxy outcomes limit interpretation of weight effects.

Study Type: Nonrandomised clinical trial (per title). No abstract text was provided, so design details cannot be confirmed.

Key Question: Does psilocybin alter inflammatory and mitochondrial stress markers in people with Parkinson disease? (Inferred from the title only.)

Key Findings:

• Unable to summarise: the abstract body was empty, so no data on sample size, dosing, outcomes, effect sizes or safety are available.
• The title suggests biomarker endpoints (inflammatory and mitochondrial stress markers) rather than motor or psychiatric outcomes, but this cannot be verified.

Clinical Relevance: If confirmed, findings would inform early-phase work on psychedelic-assisted therapy in Parkinson disease, though it is not currently relevant to UK clinical practice, where psilocybin remains a Schedule 1 substance outside research settings.

Limitations: Nonrandomised design (per title) limits causal inference; further appraisal requires the full abstract or paper. *Please resend with the complete abstract for a full summary.*

Study Type: Naturalistic longitudinal prognostic cohort study (machine learning prediction model; Marburg-Münster Affective Disorders Cohort, Germany).

Key Question: Can integrating multimodal baseline data improve prediction of 2-year relapse in major depressive disorder?

Key Findings:

• Of 564 patients with current or lifetime MDD, 270 (48%) relapsed within 2 years.
• Balanced accuracy ranged from 47.2% (SD 5.9) to 64.4% (SD 6.6), i.e. close to chance at the lower end and only moderate at best.
• The best-performing model used only a few clinical risk factors. Adding neuroimaging, genetics, neurocognition, blood markers and gut microbiome did not improve prediction.

Clinical Relevance: Routine clinical risk factors remain the most informative predictors of relapse, so expensive multimodal biomarker work-ups are not justified for individual relapse prediction in NHS practice.

Limitations: Single time-point baseline measurement, with no external validation reported in the abstract and a mixed sample (current and lifetime MDD) from two German sites, limits generalisability to UK populations.

Study Type: Cannot be determined. The title alone does not indicate whether this is original research, a review, or an editorial or commentary.

Key Question: Presumably the evaluation of large language models in mental health care, but the specific question and scope are not stated.

Key Findings: None available. No data, effect sizes or themes were provided, and I have not inferred any from the title.

Clinical Relevance: Cannot be assessed without the content.

Limitations: Not applicable. No abstract text was supplied. If you paste the full abstract, I will complete the summary. Alternatively, if you can share the full text, I can summarise from that.

Study Type: Mixed-design study: two-sample Mendelian randomisation using GWAS data, plus a prospective single-cohort neuroimaging study (122 MDD, 66 healthy controls) with machine-learning subtyping and survival analysis.

Key Question: Does structural brain network variation define biologically distinct MDD subtypes with different immune-inflammatory profiles and treatment outcomes?

Key Findings:

• Genetic liability to MDD was causally associated with lower circulating IGF-1 (pFDR = 0.013). IL-5 and leptin were negatively associated with grey matter volume in the right cingulate gyrus and bilateral putamen respectively.
• Structural covariance analysis identified a hypoconnectivity subtype (n=49) with greater anxiety and anhedonia, and a hyperconnectivity subtype (n=73) with broadly elevated inflammatory markers.
• The hyperconnectivity subtype showed faster antidepressant response during follow-up (p = 0.032).

Clinical Relevance: Supports the development of imaging- and immune-based stratification for treatment prediction, though this remains a research-stage finding with no current NHS application.

Limitations: Modest sample size, single cohort without external validation, and unspecified antidepressant regimens and follow-up duration limit generalisability and the strength of the prognostic claim.

Study Type: Cross-sectional neuroimaging analysis (resting-state fMRI) using UK Biobank data.

Key Question: Do functional connectivity patterns differ between people with a history of suicide attempt and those with non-suicidal self-directed violence, and do they track self-harm frequency and recency?

Key Findings:

• Among 1,070 participants with self-directed violence (579 suicide attempt, 491 NSSDV), the attempt group showed reduced connectivity in bilateral amygdala and right putamen, and increased connectivity in right caudate.
• More frequent self-harm correlated with increased orbitofrontal–occipital connectivity.
• Self-harm within the past 12 months showed trends towards reduced orbitofrontal connectivity (rectus gyrus, temporal lobe). Effect sizes and confidence intervals were not reported in the abstract.

Clinical Relevance: Limbic and reward circuit differences may eventually support risk stratification for suicide attempt among UK patients presenting with self-harm, but the findings are not yet clinically actionable.

Limitations: Cross-sectional design with self-reported group classification precludes causal or predictive inference, and UK Biobank participants are healthier and less representative than NHS self-harm populations.