Introduction Suicide attempts (SA) are the strongest predictor of future suicide. Hospitalization after SA is a critical clinical decision, yet longitudinal outcomes of this population remain poorly described in Portugal. Understanding socio-demographic and clinical risk factors is essential to guide prevention strategies. Objectives To longitudinally characterize psychiatric inpatients admitted after SA and evaluate recurrence, rehospitalizations, and mortality. Methods We conducted a retrospective observational study including all patients admitted to the Acute Psychiatric Inpatient Unit of Unidade Local de Saúde Lisboa Ocidental between January 2022 and August 2025, following a documented SA in the previous month. Socio-demographic, clinical, index-SA, and follow-up variables were collected from electronic health records. Descriptive statistics were applied. Results Ninety-three patients were included (53.8% female; mean age 44.9 ± 19.0 years). A majority were single (60%), 54.8% were unemployed or otherwise inactive, and only 43% had children (mean 0.7), indicating fragile social support. Nearly half (48.4%) had a prior SA, 71% presented suicidal ideation, and 46.2% reported substance misuse; 21.5% had a family history of suicidal behaviour. The index SA was impulsive in 52.7%. Methods included drug overdose (49.5%), self-cutting (11.8%), defenestration (6.5%), and hanging (5.4%). Median length of stay was 17.5 days (IQR 11–33). Discharge diagnoses were mainly depressive disorders (52.7%), bipolar disorder (11.8%), borderline personality disorder (7.5%), and schizophrenia/psychosis (8.6%). During follow-up, 15.1% presented new non-fatal SAs (15 events), 7.5% required rehospitalization (18 episodes), and 2 patients (2.2%) died by suicide. Three additional deaths of unknown cause were recorded. Conclusions This is the first longitudinal characterization of SA-related psychiatric hospitalizations in Lisbon. Patients showed a profile of psychiatric morbidity combined with marked social vulnerability (isolation, unemployment, absence of children) and frequent substance misuse. Recurrence of suicidal behaviour and fatal outcomes occurred despite hospitalization, underscoring that inpatient care alone is insufficient. These findings highlight the need for multimodal post-discharge strategies integrating early follow-up, psychosocial and occupational support, and targeted interventions for substance misuse. Addressing social determinants alongside psychiatric treatment should be a cornerstone of suicide prevention. Disclosure of Interest None Declared
Introduction Autism spectrum disorder (ASD) is a pervasive neurodevelopmental condition characterized by social, communication, and sensorimotor difficulties. Traditional diagnostic procedures such as ADOS-2 are effective but time-consuming, costly, and reliant on highly trained professionals. Machine learning (ML) offers opportunities for more efficient and objective approaches. Intellectual functioning has historically been linked to ASD, yet its diagnostic role remains unclear. Objectives This study aimed to evaluate whether cognitive profiles measured with the Stanford-Binet Intelligence Scales, Fifth Edition (SB-5), can predict ICD-10 ASD subtypes using ML models. Methods A total of 68 children with ASD (51 males, 17 females; age range 2–17 years, M = 8.9, SD = 3.8) completed the SB-5. ICD-10 classifications included F84.0 (n = 11), F84.1 (n = 21), and F84.5 (n = 36). A control group of non-ASD children referred for IQ testing was used for comparison. Eight ML algorithms were tested: k-nearest neighbors, support vector machines (linear and RBF), decision tree, random forest, logistic regression, neural network, and a majority voting ensemble. Hyperparameters were tuned via grid search with nested 5-fold cross-validation, and models were evaluated using 10-fold cross-validation. Accuracy, precision, recall, and F1 score were computed. Analyses were repeated controlling for age. Results ASD subgroups differed significantly in intellectual functioning. Children with F84.5 had higher mean IQ scores (M = 102.1, SD = 12.7) compared with F84.0 (M = 78.1, SD = 14.5) and F84.1 (M = 71.2, SD = 18.9; p < .001). Among ML models, logistic regression achieved the best performance (75% accuracy), followed closely by random forest and majority voting (74%). Controlling for age increased accuracy up to 79%. Predictions for F84.1 vs. F84.5 reached 88% accuracy, whereas F84.0 proved more difficult to classify, likely due to small sample size and heterogeneity. Conclusions Machine learning applied to standard cognitive test data can distinguish ICD-10 ASD subtypes with promising accuracy. Logistic regression and ensemble models were most effective, suggesting interpretable approaches may have clinical utility. The predictive role of intellectual ability supports its relevance in diagnostic frameworks, though further validation in larger, diverse samples is required. Integrating ML with cognitive assessments could contribute to faster, more cost-effective, and objective ASD diagnostics. Disclosure of Interest None Declared
Introduction ADHD is a neurodevelopmental disorder characterised by inattention, hyperactivity, and impulsivity. According to NICE, an estimated 3-4% of adults are diagnosed with ADHD in the UK. Gabalfa CMHT in Cardiff manages a large caseload, with ADHD making up a significant proportion. In 2025, an estimated 186 ADHD patients used the CMHT’s repeat prescription service. There is increasing demand for the provision of repeat prescription medication, as well as diagnosis, with almost 500 patients on the waiting list. The previous repeat prescription process for patients with ADHD was identified as inefficient. Requests were made either in person or via telephone, with patients verbally communicating their needs to administrative staff. The administrative staff (who also triage mental health queries and any emergencies) recorded the requests in a prescription book. Prescribers then accessed the book to review, print and sign prescriptions - a process that was often time-consuming. Objectives To digitise the repeat prescription process within the CMHT, with the aim of improving patient accessibility and modernising the service. Methods An NHS shared mailbox was introduced, which was accessible to the prescribers within the team, as well as administrative staff. The service was limited to core working hours to maintain clinical boundaries and minimise inappropriate or out-of-hours requests. Patients were informed of the change via SMS or email, including guidance on how to use the new system. Anonymous patient feedback was gathered via a survey form, sent by SMS. Results Within 6 months of the email inbox being piloted, many patients had started to use the email inbox service. Twelve patients responded to the initial survey. 83% of these patients were ‘very satisfied’ or ‘satisfied’ with the new email service. 64% found this service more convenient. 67% of patients would prefer to continue to use the email service, rather than the previous method and 25% had no preference. In addition, staff feedback noted a significant decrease in the number of calls received relating to ADHD medication, thus reducing workload. Conclusions The initial patient feedback demonstrates the positive impact of using digital tools in repeat prescription services. The key benefits are improved patient accessibility, reduced administrative burden on staff and increased efficiency. However, there is an ongoing digital divide, with some patients not using the service at all. The next steps would be to gain further feedback, to ensure a more representative patient opinion has been collected, over a longer period of time. With continued improvements to the new service, the CMHT could look at expanding to a wider patient group, not just those with ADHD. Disclosure of Interest None Declared
Introduction Encoding a key ‘hub’ scaffolding protein, the ‘Disrupted-In-Schizophrenia-1’ (DISC1) gene has been strongly implicated in brain development and functions. Genetic variance in this gene is associated with major neuropsychiatric disorders, including schizophrenia, bipolar disorder, and major depression. DISC1 is abundantly expressed in the brain of humans and various model organisms. In mice, mutation Q31L in the Disc1 gene reduces binding of DISC1 protein to GSK-3 (glycogen synthase kinase-3 alpha) and PDE4B (3’,5’-cyclic AMP phosphodiesterase 4B), accelerates degradation of BMAL1, which is involved in the regulation of glucocorticoid synthesis in the adrenal glands and the sensitivity of glucocorticoid receptor target genes. Objectives The aim of this study is to distribution of BMAL1 protein in brain regions involved in behavior regulation in mice with Q31L mutation in Disc1 gene and to investigate the behavioral response to different duration of stress effects. Methods The male mice C57BL/6 mice and those with a mutation in the Disc1 gene - Q31L were kept under standard conditions, with a 12:12 light regime, and their behavior was studied in the “open field”, “social preference” and “forced swimming” tests. Immunohistochemical analysis was performed on frozen brain sections using primary antibodies against BMAL1 and secondary ones - Alexa fluor 488. Results Mice carrying point mutation Q31L in second exon Disc1 exhibit depression-like phenotype: increased floating time and decreased social interaction. In CA1 and lateral habenula of Q31L mice was observed decreased BMAL1 protein expression. Both of this areas, especially lateral habenula involved in affective disorders. Conclusions By the fact that BMAL1, regulate oligodendrocite proliferation and migration it can be proposed that DISC1-GSK3-BMAL1 pathway is potential mechanism regulating myelination and involved in affective disorders pathogenesis, so this protein can act as a molecular target for the treatment of affective disorders. The study is supported the Russian Science Foundation No. 23-75-00023. Disclosure of Interest None Declared
Introduction Population-based screenings for substance use are highly recommended, but little is known about the screening status for substances in primary care settings (PCSs). Objectives This systematic review aims to understand what screening tools are used, barriers that impact screenings, and opportunities to improve screenings. Methods Articles focusing on the utilization of screening tools for substance use disorders (SUDs) within PCSs were considered for inclusion. Searches on PubMed and Embase databases were conducted from April 25, 2022 through July 22, 2022, which yielded 487 results; only 32 articles meeting the inclusion and exclusion criteria were included. Results Among all substances, alcohol is the most screened for at PCSs, with 20 studies included in this article. This was followed by opioids (n = 11), other illicit drugs (n = 8), cannabis (n = 6), and nicotine (n = 2). The Alcohol Use Disorders Identification Test was the screening tool most used. Most studies utilized different screening tools to screen for the same substances. Identified barriers include no referral to treatment upon positive screenings, lack of training in providing intervention, and no recognition of the effectiveness of intervention services. Conclusions Our search resulted in only 32 studies focusing on substance screening at PCSs, indicating that substance screening may not be conducted as recommended or formally. Integrating patient-centered SUDs screening in PCSs is challenging but could be feasible via implementation strategies to address barriers. Future studies focusing on the feasibility, acceptability, and effectiveness of screening tools in PCSs may increase the screening rates for early identification and intervention for SUDs. Disclosure of Interest None Declared
BACKGROUND:There has been growing interest in developing a clinical staging model for eating disorders (EDs). The aim of this scoping review is to investigate the extent, range, and nature of research that can inform clinical staging approaches for EDs. METHODS:This review was conducted following the Joanna Briggs Institute methodology for scoping reviews and the PRISMA extension for Scoping Reviews. A search of PubMed, PsycINFO, MEDLINE and Web of Science databases was conducted. Sixty studies were included for review, and data were extracted according to revised pre-published criteria. Studies were summarized using narrative synthesis. RESULTS:Thirty-six of the included manuscripts were quantitative, 18 were qualitative, and six were conceptual/review papers. The included studies primarily focused on anorexia nervosa in white, female samples and compared only two stages of illness, equating to "early" and "longstanding" or "severe and enduring" EDs. "Middle" stages or transitions between stages were not well-defined. Stages have mostly been conceptualized using duration of illness or other behavioral or clinical parameters, with few (neuro) biological studies identified. Factors that may modify illness stage have not yet been integrated into ED staging literature. CONCLUSIONS:There is growing consensus that early- and later-stage EDs have distinct characteristics and treatment requirements; however, these distinctions require clearer definition. Recommendations for future research are outlined, emphasizing integration of psychosocial and biological research, focused investigation of illness phases outside of early and longstanding phases, and evaluation of personalized, stage-based treatments. A tentative staging model for EDs is proposed.
BACKGROUND:Cannabinoids are increasingly discussed as adjuncts in addiction treatment, yet whether they improve clinically meaningful substance use disorder (SUD) outcomes beyond short-term symptom relief is unresolved. We determined whether cannabinoid exposure confers directional efficacy across opioid, alcohol, cocaine, tobacco, and methamphetamine use disorders, distinguishing symptomatic targets from sustained therapeutic outcomes. METHODS:PubMed and Embase (1975-2025) were searched for human studies evaluating cannabinoid exposure in relation to SUD outcomes. Two reviewers independently screened and extracted data. Risk of bias was assessed using RoB 2 for randomized controlled trials (RCTs), ROBINS-I for cohort studies, and JBI checklists for cross-sectional, case, and qualitative designs. Six prespecified endpoints (treatment retention, relapse, abstinence, craving, withdrawal severity, consumption) were mapped to each target SUD. Following Synthesis Without Meta-analysis (SWiM) guidance, structured narrative synthesis used a design-based weighting scheme (RCT 1.00 to qualitative 0.25). PROSPERO: CRD420251151193. RESULTS:Ninety-seven studies (41,954 participants) contributed 195 endpoint instances: 89 Beneficial (45.6%), 80 No Significant Effect (41.0%), 12 Mixed/Partial (6.2%), and 14 Harmful/Inferior (7.2%). Short-term symptom targets accounted for most Beneficial findings (76.4%). Sustained outcomes were predominantly No Significant Effect, most pronounced in opioid use disorder. Beneficial symptom findings derived overwhelmingly from weaker study designs (craving 81.5%; withdrawal severity 85.7%; consumption 80.0%). CONCLUSIONS:Cannabinoids confer short-horizon symptomatic benefits but do not demonstrate efficacy for sustained abstinence, relapse prevention, or retention, most clearly in opioid use disorder, where evidence is strongest. Findings for other disorders remain preliminary. Adequately powered adjunctive randomized trials with biochemically verified endpoints are needed.
BACKGROUND:Bipolar disorder (BD) is frequently misdiagnosed as major depressive disorder (MDD), with diagnostic delays averaging 6-10 years leading to suboptimal treatment and poorer outcomes. Early identification of patients at risk for diagnostic conversion remains a critical clinical priority. We aimed to identify clinical predictors of BD conversion in a large population-based cohort and develop a clinically applicable prediction model. METHODS:We conducted a retrospective cohort study using electronic health records from the Catalan PADRIS-PRESTO database (2010-2019). Patients aged ≥18 years with incident MDD diagnosed in specialized mental health services (2015-2019) were followed until BD diagnosis, death, or study end. Time-varying Cox regression addressed proportional hazards violations, with competing risks analysis accounting for mortality. RESULTS:Among 52,001 patients (median follow-up 4.0 years), 3,958 (7.61%) converted to BD. Psychotic features emerged as the strongest predictor (HR = 3.18, 95% CI: 2.92-3.45, E-value = 5.82), with consistent effects across follow-up. Anxiety disorders showed time-varying protection (baseline HR = 0.41, 95% CI: 0.35-0.48) that attenuated over time (interaction HR = 1.30 per period). The time-varying model achieved good discrimination (C-index = 0.745). Machine learning validation confirmed predictor importance rankings (Spearman r = 0.87). CONCLUSIONS:Psychotic features during depressive episodes robustly predict BD conversion, warranting enhanced clinical monitoring. Our time-varying prediction model using routinely collected clinical data enables early risk stratification to guide diagnostic vigilance and treatment decisions.
Large language models (LLMs) are poised to become a ubiquitous feature of our lives, mediating communication, decision-making and information curation across nearly every domain. Within psychiatry and psychology the focus to date has remained largely on bespoke therapeutic applications, sometimes narrowly focused and often diagnostically siloed, rather than on the broader and more pressing reality that individuals with mental illness will increasingly engage in agential interactions with AI systems as a routine part of daily existence. While their capacity to model therapeutic dialogue, provide 24/7 companionship and assist with cognitive support has sparked understandable enthusiasm, recent reports suggest that these same systems may contribute to the onset or exacerbation of psychotic symptoms: so-called ‘AI psychosis’ or ‘ChatGPT psychosis’. Emerging, and rapidly accumulating, evidence indicates that agential AI may mirror, validate or amplify delusional or grandiose content, particularly in users already vulnerable to psychosis, due in part to the models’ design to maximise engagement and affirmation, although notably it is not clear whether these interactions have resulted or can result in the emergence of de novo psychosis in the absence of pre-existing vulnerability. Even if some individuals may benefit from AI interactions, for example where the AI functions as a benign and predictable conversational anchor, there is a growing concern that these agents may also reinforce epistemic instability, blur reality boundaries and disrupt self-regulation. In this paper, we outline both the potential harms and therapeutic possibilities of agential AI for people with psychotic disorders. In this perspective piece, we propose a framework of AI-integrated care involving personalised instruction protocols, reflective check-ins, digital advance statements and escalation safeguards to support epistemic security in vulnerable users. These tools reframe the AI agent as an epistemic ally (as opposed to ‘only’ a therapist or a friend) which functions as a partner in relapse prevention and cognitive containment. Given the rapid adoption of LLMs across all domains of digital life, these protocols must be urgently trialled and co-designed with service users and clinicians.
Internet- bzw. onlinebasierte psychologische Interventionen haben sich bereits in mehreren Studien als wirksam erwiesen. Da sowohl die Therapieadh & auml;renz als auch die Abbruchraten immer noch eine gro ss e Herausforderung darstellen, hat sich der Autor Blake Dear im Rahmen einer Sekund & auml;ranalyse der Daten von 4 randomisierten und kontrollierten Studien nun eingehend mit m & ouml;glichen Einflussfaktoren befasst.
Cognitive Disengagement Syndrome (CDS) refers to symptoms of fogginess, daydreaming, blank staring, and slowed behaviour. Growing evidence suggests that CDS reflects a transdiagnostic construct. This study aimed to examine (1) associations between age, sex, and CDS severity; (2) differences in CDS across Attention-Deficit/Hyperactivity Disorder (ADHD), Autism Spectrum Disorder (ASD), and internalizing disorders; and (3) associations between CDS and quality of life, as assessed by the KIDSCREEN-27. CDS was assessed in 958 clinically referred children and adolescents (mean age = 10.7 +/- 2.8 years; 36.3% female) with a primary diagnosis of ADHD (n = 353), ASD (n = 494), or internalizing disorders (n = 111). Four Child Behaviour Checklist (CBCL) items (13, 17, 80, 102) were used to operationalize CDS. Analyses included ANOVAs, MANOVAs, and multiple regression analyses in the full sample and a subsample without comorbidities. CDS levels didn't differ by sex, whereas adolescents (12-17 years) showed higher CDS levels than children (6-11 years). Across diagnostic groups, ASD (M = 2.6, SD = 1.8) and internalizing disorders (M = 2.6, SD = 1.9) displayed higher CDS scores than ADHD (M = 1.8, SD = 1.6; p < .001). These findings remained significant in the subsamples without comorbidities (p < .001). Multiple regression analyses indicated that higher CDS scores were significantly associated with lower quality of life (B = -2.20, p < .001), independent of diagnostic group, sex, and age, without significant interaction between CDS and diagnosis. Correlational analyses indicated stronger associations between CDS with autistic (r = .25-.36; p < .05) and affective CBCL symptom dimensions (r = .19-.36; p < .05) dimensions than with ADHD symptoms. CDS appears to represent a clinically relevant transdiagnostic construct. Its strong expression in ASD suggests that it captures disengagement mechanisms beyond current nosological frameworks.
