BACKGROUND:People with schizophrenia-spectrum or bipolar disorder frequently have physical comorbidities, but the epidemiology of developing multiple physical conditions is unclear. We hypothesised that this group experiences an earlier accumulation of physical multimorbidity affecting multiple systems. Our aim was to estimate the relative and absolute incidence of physical multimorbidity in people with and without severe mental illness. METHODS:We did a cohort study using statewide hospital registries in Queensland, Australia, between Jan 1, 2000, and Dec 31, 2023. Individuals with a recorded diagnosis of schizophrenia-spectrum or bipolar disorder were age-matched and sex-matched (1:4) to individuals without these disorders. The study outcome was systemic physical multimorbidity, defined as the cumulative number of systems affected by chronic physical disease. We measured the relative incidence of developing systemic physical multimorbidity in at least one to at least five systems using Fine-Gray adjusted subdistribution hazard ratios (HRs); absolute differences were measured using differences in the cumulative incidence function at 20 years. Additional age-stratified and sex-stratified analyses were undertaken. This study featured no lived experience involvement. FINDINGS:A total of 178 665 individuals were included in the analysed sample: 30 189 people with schizophrenia-spectrum or bipolar disorder (11 940 females and 18 249 males; median age 32 years [IQR 24-42]) were compared with 148 476 hospital comparators in the control group (61 365 females and 87 111 males; median age 33 years [IQR 24-44]). Ethnicity data were not available. Relative incidence was higher in the disorder group with adjusted subdistribution HRs of 1·85 (95% CI 1·81-1·88) for at least one affected system, 2·25 (2·19-2·32) for at least two affected systems, 2·65 (2·56-2·74) for at least three affected systems, 2·93 (2·81-3·05) for at least four affected systems, and 3·16 (3·01-3·30) for at least five affected systems. At 20 years, absolute cumulative incidence differences were 9·79 additional cases per 100 persons (95% CI 9·06-10·53) for at least one affected system, 18·43 additional cases (17·58-19·28) for at least two affected systems, 18·57 additional cases (17·73-19·42) for at least three affected systems, 15·71 additional cases (14·92-16·51) for at least four affected systems, and 11·66 additional cases (10·95-12·37) for at least five affected systems. Elevated adjusted subdistribution HRs were consistently observed for younger strata, with the highest observed in females younger than 25 years with at least five affected systems (adjusted subdistribution HR 9·24 [95% CI 7·30-11·71]). INTERPRETATION:People with schizophrenia-spectrum or bipolar disorder experience increased relative and absolute incidence of systemic physical multimorbidity spanning multiple systems, compared with individuals without these disorders. The largest differences were observed among younger cohorts, emphasising the need for early prevention. FUNDING:None.
BACKGROUND:People with severe mental illness (SMI) had a higher risk of death following SARS-CoV-2 infection compared with people without SMI during the COVID-19 pandemic peak. It is unclear whether this risk reduced in later stages after vaccination roll-out. We aimed to assess all-cause and cause-specific mortality following SARS-CoV-2 infection in people with or without SMI across pandemic stages, using whole-country data, while accounting for sociodemographic and clinical characteristics. METHODS:In this retrospective, whole-country cohort study, we analysed linked electronic health records for all people in England registered to primary care who were alive in November, 2019 (>99% of the population). The study population included people who had COVID-19 between Jan 31, 2020, and June 30, 2023, with information on deaths and vaccination. Data extracted from the NHS England Secure Data Environment provided through the British Heart Foundation Data Science Centre CVD-COVID-UK-COVID-IMPACT Consortium linked datasets were curated to form a dataset in July, 2024. Cox regression estimated differences in all-cause and COVID-19 mortality. Analyses were stratified by pandemic stages, defined by WHO policy and UK Government vaccination roll-out: Jan 31-Sept 30, 2020 (stage 1); Oct 1, 2020-July 18, 2021 (stage 2, vaccination roll-out); and July 19, 2021-June 30, 2023 (stage 3, post-vaccination roll-out). SMI comprised schizophrenia, schizoaffective disorder, bipolar disorder, or other affective and non-affective disorder with psychosis. A patient and public involvement panel was consulted during design phases of the study. FINDINGS:13 463 945 people with COVID-19 were identified, of whom 160 190 (1·19%) had SMI. People with SMI were older [mean age [SD], SMI 49·7 years [17·9], non-SMI 45·0 years [17·6]), had a similar sex split (females SMI vs non-SMI: 56·3% vs 56·6%); and were mostly White British (77·3% or 76·3%). People with SMI had higher all-cause mortality than those without (adjusted hazard ratio [aHR] 1·56, 95% CI 1·53-1·59). All-cause mortality risk in people with SMI was elevated in stage 1 (aHR 1·27, 95% CI 1·22-1·32), but larger in the stage 2 (aHR 1·56, 1·51-1·61) and stage 3 (aHR 1·55, 1·50-1·59). Vaccination prevalence was initially higher for people with SMI (in stage 2: 112 795 [73·2%, 95% CI 72·9-73·4], compared with people without SMI, 8 852 970 [67·4%, 67·3-67·4]). By study end, fewer people with SMI were fully vaccinated (127 145 [79·4%, 79·2-79·6]), compared with people without SMI (11 605 915 [87·2%, 87·2-87·3]. Higher COVID-19 mortality risk in people with SMI persisted but was partially attenuated by vaccination (aHR 1·25, 1·17-1·34). INTERPRETATION:Excess mortality in people with SMI persisted throughout the COVID-19 pandemic. Preventive public health and clinical strategies are needed to address excess mortality and lower vaccination. FUNDING:Health Foundation-Academy of Medical Sciences.
BACKGROUND:Many patients with N-methyl-D-aspartate receptor (NMDAR) encephalitis show persistent neurological or psychiatric symptoms, despite immunotherapy. The neuroanatomical basis of these symptoms remains uncertain. Morphological complexity analysis with fractal dimensionality has emerged as a sensitive neuroimaging method in related conditions, but remains unexplored in autoimmune encephalitis. We aimed to characterise morphological brain complexity, symptom trajectories from peak illness to the post-acute stage, and the relationship between structural brain changes and post-acute outcomes in patients with NMDAR encephalitis. METHODS:In this cross-sectional study, we included patients with post-acute NMDAR encephalitis referred to Charité-Universitätsmedizin Berlin, Germany. Patients were matched (1:1) to healthy control participants for sex and age by use of logistic regression propensity scores and nearest-neighbour matching. Data were collected between July 18, 2011, and March 17, 2021. Patient outcomes were assessed with the modified Rankin Scale (mRS), Clinical Assessment Scale for Autoimmune Encephalitis (CASE), psychiatric phenotyping, and neuropsychological testing. Morphological brain complexity was quantified from T1-weighted structural MRI. Normative modelling was used to evaluate the relationship between structural brain changes and post-acute outcomes. Self-reports on post-acute symptoms by individuals with lived experience informed the study design. FINDINGS:We included 70 patients (60 [86%] female and ten [14%] male; mean age 28·1 years [SD 10·2; range 15·0-67·0]) and 70 age-matched and sex-matched healthy control participants (60 [86%] female and ten [14%] male; 29·0 [9·3; 16·0-65·2]; ethnicity data unavailable). Study visits occurred at a median of 22·2 months from disease onset (IQR 8·5-42·5). Patients were severely affected at peak illness, but showed significant clinical improvement in both mRS (median 5 vs 1, Z -7·2, p<0·0001) and CASE (median 11 vs 1, Z -7·2, p<0·0001) in the post-acute stage. Nevertheless, 47 (67%) patients continued to show residual CASE symptoms, most commonly memory dysfunction (43 [61%]) and psychiatric symptoms (25 [36%]). In patients with persistent symptoms, psychiatric manifestations shifted from a psychosis-dominant phenotype at peak illness to an affective-dominant phenotype in the post-acute stage. Brain imaging showed a characteristic pattern of reduced morphological complexity, including the hippocampus bilaterally and a fronto-cingulo-temporal cluster in grey and white matter. Normative modelling revealed that patients with persistent symptoms showed stronger alterations of brain morphology than those without (psychiatric symptoms: t -2·65, p=0·0100; memory dysfunction: t -3·98, p=0·0002; both vs no symptoms t -5·46, p<0·0001). Similarly, stronger morphological alterations were associated with lower visuospatial and verbal memory performance (all pFDR<0·015). INTERPRETATION:Post-acute NMDAR encephalitis is characterised by systematic alterations of brain morphology. These changes are associated with persistent psychiatric symptoms and memory dysfunction after immunotherapy, highlighting fractal dimensionality as a promising new imaging marker of long-term outcomes. The phenotypic shift in psychiatric symptoms and the persistence of memory dysfunction highlight the importance of structured post-acute care to address long-term affective and cognitive morbidity in patients with NMDAR encephalitis. FUNDING:German Research Foundation and German Ministry of Education and Research.
Ambient artificial intelligence (AI) documentation tools, known as AI scribes, have entered psychiatric clinical practice, bundling at least three operations into a single workflow, without sufficient attention to the differences among them. In this Personal View, we argue that AI-generated clinical narrative (history of present illness), AI-generated clinical observation (mental status examination), and AI-generated diagnostic reasoning (assessment and plan) are functionally distinct operations that stand in qualitatively different relationships to their shared input, with distinct failure modes and oversight demands. We describe why psychiatric practice exposes these distinctions with particular clarity and examine their implications for liability, informed consent, and regulation. Psychiatry is uniquely positioned to surface distinctions among these operations because of the therapeutic and legal weight of word choice, the unnarrated structure of the mental status exam, and the formulation-driven nature of diagnostic reasoning. We offer provisional practice considerations for clinicians currently using these AI scribes and call for the field to build frameworks that reflect the different claims these AI scribes' outputs make.
BACKGROUND:Although psychosocial interventions delivered by non-specialist providers effectively treat common mental disorders such as depression and anxiety, the treatment-specific elements driving their efficacy remain largely unknown. Our aim was to compare and rank the efficacy of the individual active components of task-shared psychosocial interventions and to predict efficacy using individual participant characteristics. METHODS:We performed a systematic review and Bayesian individual participant data component network meta-analysis of RCTs comparing different task-shared psychosocial interventions with control conditions for the treatment of adults with common mental disorders. We searched MEDLINE, Embase, PsycINFO, and CENTRAL from database inception to March 15, 2023. We included additional datasets published up to May 29, 2026 and unpublished RCTs. We sought individual participant data from the trial authors. A purpose-built taxonomy of treatment-specific elements was used to dismantle interventions into their constituent components. The primary outcome was efficacy in reducing symptoms of common mental disorders, as measured at study endpoint. We used incremental mean difference to indicate the added benefit (or detrimental effect) of adding a component to a treatment. We assessed the risk of bias of the included studies with the revised Cochrane Risk of Bias tool. The protocol was published in a peer-reviewed journal. FINDINGS:We included 34 RCTs, from which 30 trials (89%) contributed individual participant data from 10 612 participants. Of these participants, 7662 were women (72·2%) and 2950 were men (27·8%). The mean age of participants was 36·7 years (SD 5·5). Findings identified strengthening social support (incremental mean difference -9·48; 95% credible interval [CrI] -13·29 to -6·60), behavioural activation (-4·15; -7·48 to -0·05), and problem management (-4·08; -5·37 to -2·83) as the most beneficial components. Relaxation (7·97; 3·35 to 12·11) and, less clearly, cognitive reframing (5·17; 0·78 to 9·15) were identified as detrimental components. Interaction analyses revealed that component effects vary systematically by baseline severity and other sociodemographic characteristics. Ethnicity data were not available. Individuals with lived experience contributed to the interpretation of results. Personalised effect estimates can be computed via a freely accessible web application. INTERPRETATION:Strengthening social support, problem management, and behavioural activation were associated with the greatest incremental benefit within task-shared psychosocial interventions for adults with depression or anxiety. FUNDING:European Commission.
