JAMA Pediatrics

Objective: To evaluate whether a locally hosted open-weight large language model (LLM) can extract documented psychosocial factors from pediatric psychiatric intake notes and apply validated extraction to a large emergency psychiatry cohort. Materials and Methods: We identified emergency department presentations at Cincinnati Children's Hospital Medical Center from January 1, 2016, through December 31, 2024, among patients younger than 18 years with psychiatric billing diagnoses. Using full-text intake notes, gpt-oss:120b classified peer conflict, sleep disruption, and school-related academic, attendance, and disciplinary issues as detected, negated, or indeterminate. Four human raters independently reviewed 50 notes. We compared Fleiss' kappa among humans alone versus humans plus the LLM, assessed repeated-query stability across 50 independent calls per note, and applied the workflow to all eligible notes. Results: Among 37,315 eligible admissions, 22,284 had eligible intake notes; 22,270 produced parseable JSON. In detected-versus-not-detected coding, human-plus-LLM reliability did not differ significantly from human-only reliability across measures (human {kappa} 0.71-0.94; human-plus-LLM {kappa} 0.70-0.93). Stability was associated with human agreement: mean LLM-human agreement increased from 42.6% for classifications with less than 80% stability to 82.7% for classifications with 100% stability (Pearson r = 0.36). Full-cohort extraction showed frequent and overlapping documented factors: sleep disruption was most frequently detected (57.7%), followed by peer conflict (47.2%), academic issues (43.4%), disciplinary issues (43.3%), and attendance issues (16.9%). Discussion: Agreement varied by construct and was strongest when repeated model outputs were stable. Conclusion: Locally hosted open-weight LLMs can support scalable structured extraction of documented psychosocial factors from pediatric psychiatric intake notes after local validation.

We assembled a multimodal clinical dataset describing demographics, placement history, prenatal substance exposure (PSE), birth characteristics, adverse childhood experiences (ACEs), International Classification of Diseases (ICD) diagnoses, and laboratory results for 3,685+ pediatric patients evaluated between 2014 and 2024 at the University of Minnesotas Adoption Medicine Clinic (AMC). Data were curated from electronic medical records through a combined manual and automated extraction protocol using a standardized operating procedure. The resulting dataset integrates structured EMR fields including neuropsychological, laboratory, and diagnostic information with manually pulled fields of ACE scores, PSE history, and placement history. We provide an overview of the population represented and describe the datasets structure, variable definitions, and validation procedures. This resource enables investigations into how early adversity impacts medical and developmental outcomes, and provides one of the largest standardized clinical placement history, PSE, and ACE datasets in an adoption and foster care pediatric population.

Introduction: While growing evidence suggests that music training supports child development, few long-term randomized controlled trials (RCTs) have rigorously tested these claims. Moreover, it remains unclear whether the benefits are confined to music-specific domains or extend to higher-order cognitive functions such as inhibitory control (IC), a core executive function associated with long-term outcomes in academic achievement, career success, socio-emotional health, and physical well-being. This paper presents the protocol for the Extracurricular Activity and Child Early Learning and Development (EXCEL) trial, an RCT designed to assess the feasibility of a long-term music training program focusing on the brain and behavioral correlates of IC. Methods: A total of 126 children, aged 6 to 8 years and residing in neighborhoods with limited resources in Los Angeles, were individually randomized to either a music (intervention) or theatre (active control) after-school program. Both programs were delivered over 24 months by established community arts organizations. Eligibility criteria included: average intellectual functioning, no major medical or psychiatric conditions, and MRI eligibility. Children with prior formal music training exceeding six months or severe hearing impairment were excluded. Before the intervention began, all participants completed baseline behavioral and neuroimaging assessments. The primary trial aim was to assess the effects of extended music training, relative to theatre training, on changes in measures of IC (i.e., Go/No-Go task and delayed gratification) and related neural functional activation. A secondary interim aim of the trial was to evaluate the feasibility of conducting a long-term RCT of music education in a first cohort, measured by participant retention, adherence to the program, willingness to continue at the 12-month mark, and fidelity. Progress: Recruitment, screening, baseline testing, randomization, and program enrollment began in August 2022, and after-school programming began in October 2022. The randomized interventions and all data for the first cohort (N = 42) have been collected. Intervention and active control programs for a second cohort are ongoing and will end in Fall 2026. Discussion: This paper reports the EXCEL trial protocol and provides feasibility estimates for implementing a long-term randomized controlled trial of music training in real-world, community-based settings with children. While similar neuroimaging RCTs are currently underway in Europe, the EXCEL trial is among the first in the United States to integrate longitudinal neuroimaging with arts intervention. Findings will inform the viability of scaling such programs and contribute to our understanding of how sustained music engagement may influence the development of inhibitory control circuitry in childhood.

Background Large-for-gestational-age (LGA) and macrosomic newborns are at increased risk of adverse perinatal outcomes, including death, yet the burden of neonatal mortality associated with these conditions in low- and middle-income countries (LMICs), where ongoing nutritional and epidemiological transitions suggest their prevalence will rise, remains poorly quantified. In this study, we quantify the neonatal mortality risk associated with LGA and macrosomia from 16 subnational birth cohorts in low- and middle-income countries between 2000 and 2017. Methods and findings This is an individual-participant meta-analysis to estimate neonatal mortality rates (NMRs) and relative risks among LGA infants (>90th and >97th percentile birth weight-for-gestational-age using INTERGROWTH-21st) versus appropriate-for-gestational-age (AGA, 10th-90th percentile) infants. Macrosomic ([≥]4000 g and [≥]4500 g) neonates were compared with those weighing 2500 g-3999g. Missing birth weights were imputed using recalibration and multiple imputation methods. We used random effects meta-analysis to pool relative risks. Median prevalences of LGA >90th and >97th percentile were 5.3% (interquartile range 3.6-8.2) and 2.6% (IQR 1.3-4.5), respectively; macrosomia ([≥]4000 g and [≥]4500 g) prevalences were 1.0% (IQR 0.3-3.1) and 0.06% (IQR 0.0, 0.30), respectively. Mortality was highest among preterm plus LGA infants (61.3 per 1000). LGA infants in the >90th percentile had over twofold increased mortality compared with appropriate-for-gestational-age infants (RR: 2.46; 95% CI: 1.86-3.25), while >97th percentile infants had a higher risk (RR: 3.77; 95% CI: 2.50-5.69). Term LGA >97th percentile infants also showed elevated mortality (RR: 3.14; 95% CI: 1.58-6.22). For LGA >97th percentile, the risk was higher in the early neonatal period (RR: 2.71; 95% CI: 1.92-3.82) than late (RR: 1.69; 95% CI: 1.22-2.34). There was no overall association between macrosomia ([≥]4000 g) and neonatal mortality. Population attributable fractions were 7.2% for LGA >90th percentile and 0.4% for macrosomia ([≥]4000 g). Conclusions Neonatal mortality risks were elevated among LGA infants in low- and middle-income countries, particularly at extreme values (>97th percentile) and during the early neonatal period. Macrosomia showed weaker, less robust associations. Although LGA prevalence is currently low ([~]5%) and contributes less to neonatal mortality than small newborns, ongoing nutritional and epidemiological transitions suggest increasing prevalence. This highlights the need for strengthened surveillance, monitoring, and improved delivery planning to ensure that no population is left behind.

AI-assisted clinical care may compound, rather than correct, existing health inequities. We applied Omar and colleagues' validated four-domain emergency-medicine benchmark to OpenEvidence (OE), a literature-grounded clinical LLM used by tens of thousands of US physicians daily, across 100 emergency-department cases and 20 sociodemographic labels. OE was consistent on the codified clinical decisions, triage, workup, and treatment, but diverged sharply on mental-health screening, where it flagged many historically marginalized groups between three and ten times more often than demographically unmarked cases. Cases labeled as unhoused received recommendations in 78 to 87 percent of responses (versus a 9 percent no-identifier-control rate); cases labeled as transgender in 22 to 24 percent; and Black transgender women specifically in 47 percent. A pre- registered audit of 193 free-text rationales localized the differential to the inner layer of the response, in the structure and tone of the rationale rather than the recommendation itself. Literature grounding may redistribute sociodemographic disparity in clinical AI rather than remove it. As clinical LLMs move toward agentic deployment, equity audits should examine how evidence is applied to each patient, not only whether citations are present.

Aim: To systematically evaluate evidence on the effects of post-discharge early developmental intervention programs (EI) on behavioral development, quality of life, participation, executive functioning, parent-child interaction, and use of medical services from infancy through adolescence in children born preterm. Method: Four bibliographic databases and one trial registry were systematically searched for randomized controlled trials up to April 23, 2024. Two reviewers independently screened studies and extracted data. In clinically and methodologically comparable studies, random-effects meta-analysis were performed. Risk of bias was assessed with the Cochrane RoB 2 tool, and certainty of evidence with the GRADE approach. Results: Twenty-six studies met inclusion criteria, eleven studies including 2,315 preterm born infants reported relevant outcomes, and seven contributed to meta-analyses. Most reported results showed some concerns or high risk of bias; certainty of evidence ranged from very low to moderate across outcomes. EI may offer small benefits for selective attention, behavioral problems and parent-child interaction. Little to no effect was found for special educational needs, language skills, executive functioning and the use of medical services. No included studies evaluated the effect of EI on ADHD, quality of life, or participation related to mobility or leisure activities. Interpretation: EI may improve problems typically seen in preterm children and should be offered especially to those with additional medical or social risk factors. High-quality, contemporary trials are needed to establish reliable clinical recommendations regarding EI strategies and complementary interventions throughout childhood.

Background: Adverse childhood experiences (ACEs) are traumatic or adverse events in early life that can have lasting effects on behavioral, emotional, and psychological functioning. Prior research suggests ACEs relate to later psychiatric outcomes through threshold, cumulative, and individual-specific risk patterns. Few studies, however, have operationalized all three models to test ACE-specific associations with diagnosed psychiatric disorders in individuals who are adopted or with foster care histories. Methods: We conducted a cross-sectional retrospective study using electronic health record data from foster care and adopted patients aged 0-21 years old seen at the University of Minnesota Adoption Medicine Clinic (UMN-AMC) between 2014-2024. Extracted measures included ACE history, demographics, and psychiatric diagnoses. We used latent class analysis and logistic regression to identify clusters of adversity and estimate associations with psychiatric diagnosis domains, adjusting for Sex and Age at Initial Visit. Results: ACEs showed a threshold pattern across psychiatric domains, with higher ACE counts associated with greater odds of psychiatric diagnoses. Individual risk modeling indicated that exposure to abuse or violence was associated with higher odds of psychiatric diagnoses. Across cumulative and individual risk approaches, Anxiety Disorders, Mood Disorders, and Behavioral or Emotional Disorders showed the greatest sensitivity to adversity. Conclusion: Current ACE models may not fully capture neurodevelopmental impacts reflected in diagnosed psychiatric disorders among adolescents, particularly in high-risk groups such as foster and adopted individuals. In a large clinic sample our findings support a nuanced association between ACEs and later psychiatric diagnoses and highlight the need for ACE-focused assessment, prevention, and treatment strategies tailored to foster care and adopted populations.

Objectives To explore caregiver and clinicians perspectives on implementing mental health conversations and supports for caregivers of children with chronic conditions in paediatric outpatient clinics. Specifically, views were sought on (a) screening approaches and measures (phase 1) and (b) how feedback and support could be provided to caregivers experiencing mental health difficulties (phase 2). Methods Caregivers and clinicians from two outpatient clinics (neuromuscular and diabetes) at a tertiary paediatric hospital in Melbourne, Australia participated in online focus groups in July and August 2024. Caregivers were recruited from outpatient clinics and clinicians were recruited via email. Both groups were combined for phase 1 before separating into breakout rooms for phase 2. Two authors conducted reflexive thematic analysis of transcripts using NVivo. Results Sixteen participants (caregivers n = 8; and clinicians n = 8) took part in in two semi-structured focus groups. Analysis generated two overarching domains, each comprising multiple themes. Domain 1, Addressing caregiver mental health, captured themes of overwhelm and invisibility, diverse caregiving roles, and the need for time and resources to support wellbeing conversations. Domain 2, Housing the mental health conversation, encompassed themes of screening preferences, caregiver agency in confidentiality, delivery of feedback, and access to tailored supports. Conclusions Caregivers and clinicians support routine caregiver mental health discussions in paediatric outpatient settings. Caregivers favour screening at diagnosis and key transitions, with clear, and actionable feedback delivered away from the child. Questions about record-keeping warrant further exploration, as do the perspectives of fathers.

Background Post-acute sequelae are well described following COVID-19 but may also occur after other respiratory infections and Aedes-borne infections. Evidence remains fragmented due to heterogeneity in study design, populations, and exposure, outcome, and follow-up definitions. Methods We synthesized and compared post-acute sequelae across influenza, RSV-ARI, dengue fever, chikungunya, Zika, and yellow fever. We searched five databases from inception to 25-08-2025 for articles quantifying risk, incidence, or rates of post-acute sequelae following these diseases. Eligible non-randomized observational studies assessed post-acute neurological, psychiatric, gastrointestinal, cardiovascular, respiratory, renal, musculoskeletal, autoimmune, or endocrine outcomes after confirmed infection. Risk of bias was assessed using ROBINS-E. Random-effects meta-analyses with restricted maximum likelihood estimation were conducted when comparable effect estimates were available (PROSPERO #CRD420251124994). Findings 51 studies were included, predominantly from high-income regions. Most were retrospective cohorts using ICD-coded diagnoses; prospective studies used laboratory-confirmed infections. Data sources, comparator groups, exposure definitions, outcome ascertainment, and follow-up periods varied substantially. Meta-analyses were feasible for RSV, influenza, and dengue fever. All RSV-ARI studies were pediatric and assessed infections during infancy, which were associated with higher pooled odds of physician-diagnosed asthma (OR:2.93 [95%CI: 2.12-4.06]). Influenza studies used COVID-19-positive comparators; pooled estimates showed lower risk for neurological (HR:0.82 [0.76-0.89]) and composite outcomes (RR:0.88 [0.82-0.95]), with other organ systems non-significant. Dengue fever studies spanned all ages and showed increased risks of anxiety (HR:1.34 [1.01-1.78]), dementia (HR:1.61 [1.10-2.35]), autoimmune (RR:1.39 [1.17-1.67]), cardiovascular (HR:1.51 [1.27-1.80]), psychiatric (HR:1.17 [1.07-1.28]), and any sequelae (HR:1.19 [1.13-1.25]) versus those without prior infection. Interpretations Post-acute sequelae contribute to overall disease burden following RSV-ARI and dengue fever. The evidence remains limited by heterogeneity in study design, exposure and outcome definitions, comparator selection, and follow-up duration. Greater standardization in study design and reporting is needed to improve comparability and strengthen causal inference.

Background Orofacial clefts (OFCs) and other palate abnormalities (PAs) are among the most common birth defects worldwide and are characterized by the abnormal formation of the lip and/or palate. Genetic studies have traditionally classified OFC cases as either syndromic, involving OFCs alongside other congenital anomalies, or nonsyndromic, which represent the majority of cases and occur in isolation. Emerging genomic evidence indicates that genes traditionally associated with syndromic forms of OFC can also harbor variants contributing to isolated cases, challenging the notion of a strict dichotomy between these categories and supporting their integration for gene discovery. Methods In this study, we applied multiple analytic approaches to characterize the genetic architecture of OFC and PAs by integrating genomic data from 2,497 trios with an OFC (n=2080) and PA (n=417) affected proband. We compared these findings across OFC subtypes and syndromic status with those from 5,515 control trios to identify enriched biological pathways and mechanisms and to prioritize candidate genes using variant burden testing. Results We observed a significant enrichment of de novo protein-truncating and damaging missense variants in cases compared to controls (OR = 2.17, p = 1.21x10-32), with particularly strong signals in biologically relevant gene sets involving OFC-associated, constrained, Mendelian disorder, and mouse candidate genes. Variant burden testing identified 39 OFC risk genes at FDR [≤] 0.05, which we then integrated with 593 established OFC genes to interrogate the functional underpinnings of OFC via network analysis. This analysis revealed 309 high-order interactor genes not previously associated with OFC. Notably, this OFC network clustered into ten distinct biological pathways, with nucleosome-associated genes showing significant enrichment among cases in our cohort (OR = 14.8, p = 8.1x10-4). In a final integrative step, we combined evidence across all analyses to nominate 231 candidate genes, 32 of which contained at least two deleterious de novo variants in our cohort. Conclusions These findings underscore the value of integrating diverse OFC and PA subtypes, syndromic status, and variant classes to refine the genetic architecture of these disorders, highlighting both phenotypic expansion of known disease genes and the emergence of novel gene-phenotype associations.

Objective Clinical narrative provides a unique window into provider reasoning and attribution, but use has been limited by resource requirements and extensive fine-tuning, and LLMs in particular have traditionally not performed well at medical coding. We optimize and evaluate a reproducible method for automated diagnosis assignment using LLMs in clinical notes and compare with EHR structured diagnoses. Methods We used GPT-OSS for prompt engineering and task segmentation to create a model that extracts ICD-10-CM diagnoses, with estimates of severity, currency, and importance, from progress notes. We assessed performance across multiple cohorts of patients aged 0-21 years. For each, 100 outpatient provider notes were selected across levels of severity, along with coded diagnoses from that visit (EHR); a subset of 130 notes were subjected to clinical expert review. Results Comparison showed 18.7% exact code and 33.3% ICD-10-CM category match between EHR and LLM, but semantic similarity of 0.93 at the category level. Compared to expert review, LLM precision was 0.84 and recall 0.49 for exact matches, and 0.92 and 0.62, respectively, for category-level matching. In contrast, EHR coded diagnoses showed slightly higher precision (0.94 for both cases) and substantially lower recall (0.27 and 0.43) versus expert review. Codes not identified by the LLM were more often rated by the reviewer as lower importance or certainty. Conclusion We demonstrate a reusable approach to optimizing a pretrained LLM for use in diagnosis extraction from clinical notes, facilitating large-scale diagnosis screening by LLMs without the need for expensive study-specific model refinement.

Introduction. Routine childhood immunisation is frequently disrupted in conflict-affected settings, leaving many children unvaccinated (zero-dose [ZD]). Their vaccination is now a global priority, but published evidence on restoring immunisation services in these settings is limited. We evaluated a nurse-led, community-based Expanded Programme on Immunisation adapted to a conflict-affected setting in Myanmar, focusing on factors associated with full immunisation (FI) among ZD children. Methods. This mixed-methods observational cohort study enrolled children from November 2023 to December 2025; analyses of FI outcomes were restricted to children enrolled >=18 months, with primary analyses focused on ZD children. Associations between programme delivery factors including vaccination opportunity (the ratio of vaccination sessions available to visits required for FI based on age and vaccination schedule [accelerated versus routine]) and FI were assessed using mixed-effects logistic regression with a random intercept for site. Programme cost and qualitative data from document review and questionnaires were also analysed. Results. Of 13,263 children enrolled, 6563 (49%) were in the analytic cohort; 2,684 (20%) were ZD. Among ZD, 452 (17%) were FI at 12 months and 1329 (50%) at 18 months. Accelerated schedule (OR 3.00, 95% CI 1.11-8.13) and greater vaccination opportunity (OR 2.1 per 0.5 unit increase in opportunity, 95% CI 1.8-2.4) were strongly associated with FI at 12 months, with smaller effects at 18 months. The cost per fully immunised ZD child was US$147, primarily reflecting substantial vaccine costs. Qualitative findings indicate that community engagement increased demand and access, but insecurity and logistical challenges limited service continuity and vaccination opportunities. Conclusion. FI improved over time but remained suboptimal through 18 months. Vaccination opportunity and schedule influenced the timing of FI, but sustained follow-up was critical for completion. Community-based delivery enabled restoration of immunisation services where formal systems had collapsed, demonstrating what is possible and what it demands in active conflict.

Ambient AI documentation systems rely on automatic speech recognition to transcribe patient-provider conversations before generating clinical notes. However, little empirical evidence exists on how these systems perform in mixed-language clinical encounters. We conducted a mixed-method heuristic evaluation of an ambient AI documentation tool using 24 reenacted primary care conversations involving Spanish-English and Mandarin-English code-switching. Quantitative analyses measured mixed error rate (MER) and code-switching detection. Overall MER was low, with a median of 4% and less variation in Spanish-English conversations, and 9% in Mandarin-English conversations, but with outliers reaching 67%. The system generally detected language switches reliably, although deletions occurred frequently in Mandarin-English transcripts at switch points. Qualitative analysis revealed transcription errors related to phonetic similarity, automatic language translation, clinical terminology recognition, and language-specific challenges. These findings highlight considerations for improving ambient AI clinical documentation systems to support multilingual providers in delivering care for linguistically diverse populations.

Importance. Large language models are increasingly explored as clinical decision support tools in orthodontics, yet existing evaluations have been confined to knowledge based question answering where reported accuracy ranges from 18% to 100%. No study has evaluated performance on the computational and classificatory tasks that define daily diagnostic work. Furthermore, 84.3% of published healthcare large language model studies fail to report the number of repeated queries performed, leaving output stochasticity unexamined. Objective. To compare the diagnostic accuracy and output consistency of three frontier reasoning-enhanced large language models, namely, ChatGPT 5.4 (Thinking), Gemini 3 (Thinking), and Claude Opus 4.6 (Extended Thinking), on Bolton analysis, Index of Orthodontic Treatment Need-Dental Health Component (IOTN DHC) classification, space analysis, and lateral cephalometric interpretation. Methods. In this comparative cross-sectional study with a repeated-measures design, each model, accessed through its respective consumer facing web interfaces under default provider settings rather than through application programming interfaces, processed 200 purpose-built items (50 per task) across four independent trials, yielding 2,400 observations. Responses were scored against a pre-established reference standard by two independent raters using strict binary exact match criteria. Accuracy was reported with exact binomial 95% confidence intervals. Inter-model comparisons used Cochran's Q test with post-hoc McNemar's tests and Bonferroni correction. A supplementary context-rich prompting evaluation was conducted on 40 items (480 observations). Results. Claude Opus 4.6 (Extended Thinking) achieved the highest accuracy (99.0%; 95% CI: 96.4 to 99.9%), followed by Gemini 3 (Thinking) (95.5%; 91.6 to 98.1%) and ChatGPT 5.4 (Thinking) (94.0%; 89.8 to 96.9%) (Cochran's Q=6.87, p=0.032). Each model exhibited distinct, non-overlapping error profiles concentrated at the normal-abnormal classification boundary. An accuracy-consistency paradox emerged: the most accurate model was the least consistent (93.0%), while the least accurate was the second-most consistent (98.0%). Context rich prompting eliminated all errors across all three models. Interpretation. Frontier reasoning large language models achieved high overall accuracy on orthodontic diagnostic tasks but retained concealed, task-specific vulnerabilities detectable only through repeated-trial evaluation. An accuracy-consistency paradox, in which the most accurate model was the least consistent, demonstrates that single-trial evaluations cannot characterise clinical risk. The reasoning modes were associated with high arithmetic accuracy but did not compensate for imprecise parametric knowledge on classification tasks; however, the absence of a non-thinking baseline means this association cannot be attributed to the thinking mode itself. Context-rich prompting eliminated all errors on synthetic data but should be regarded as a necessary yet insufficient prerequisite for clinical deployment pending prospective validation on real patient data.

Lower cognitive ability measured in childhood or late adolescence has been consistently associated with higher mortality risk across adulthood. However, this evidence largely relies on single assessments, leaving it unclear to what extent mortality risk reflects cognitive differences established early in life versus developmental divergence during adolescence - a period of substantial neurocognitive plasticity. Using two nationally representative Swedish cohorts comprising 9,412 males born in 1948 and 1953, we linked cognitive ability assessed in primary school at age 13 years and military conscription at age 18 years to all-cause and cause-specific mortality recorded in nationwide registers through 2025. We decomposed late-adolescent cognitive ability into childhood cognitive level and adolescent cognitive change and evaluated their independent associations with mortality. Childhood cognitive level (HR = 0.81; 95% CI, 0.78-0.85) and adolescent cognitive change (HR = 0.84; 95% CI, 0.79-0.89) independently predicted lower mortality risk, also after adjustment for parental education. Childhood cognitive level and adolescent cognitive change showed partially distinct cause-specific patterns. Childhood cognitive level was most strongly associated with mortality from intrinsic causes, whereas adolescent cognitive change showed relatively stronger associations with external causes, particularly accidental deaths. Although adolescent cognitive change was associated with psychosocial factors including education and psychiatric diagnosis at conscription, its association with mortality persisted after adjustment for these factors. These findings suggest that cognitive development during adolescence carries independent prognostic information regarding long-term survival beyond cognitive level established by late childhood, highlighting adolescence as a consequential period for lifelong health.

Background: Prenatal substance exposure (PSE) occurs when an individual is exposed to substances in utero. PSEs may have lasting effects on mental health. We tested whether PSEs show threshold, cumulative, or individual substance associations with childhood psychiatric diagnoses. Methods: Clinical variables (demographics, ICD-9/10 diagnoses, PSE history) were extracted from electronic health records from the University of Minnesota Adoption Medicine Clinic. PSEs were identified from caregiver and child-protective-services narratives and/or toxicology (cord tissue/blood, meconium). For each ICD-9/10 diagnostic category, we fit logistic regression models comparing (1) exposure thresholds (0, 1, 2, 3, 4+ exposures), (2) a cumulative exposure count, and (3) individual substances to estimate marginal odds ratios (ORs) with 95% Confidence Intervals (CIs). Results: Psychiatric diagnoses increased with the number of PSEs. Relative to no exposure, odds of an Anxiety Disorder rose from OR 1.47 (95% CI 1.16-1.87) with one exposure to OR 2.03 (1.64-2.52) with >=4 exposures. Higher cumulative exposure scores were associated with Anxiety Disorders (OR 1.28, 1.18-1.38), Behavioral and Emotional Disorders (OR 1.42, 1.31-1.54), Substance Use Disorders (OR 1.52, 1.29-1.79), and Mood Disorders (OR 1.16, 1.04-1.30). Alcohol, tobacco, and marijuana exposures were associated with increased odds of at least one psychiatric diagnosis, and each substance showed at least one significant diagnostic cluster when modeled independently. Conclusion: Increasing numbers of PSEs were associated with higher odds of psychiatric diagnoses, with patterns varying by substance and outcome. These findings motivate research on exposure timing and combinations to support earlier identification and intervention for at-risk children.

Objective: To understand if sociodemographic and neuropsychiatric characteristics of people diagnosed with autism in the United Kingdom (UK) and Sweden have changed since 2010. Design: Cross-context population-based cohort studies. Setting: UK primary care records from 2010-2023 and Swedish population-wide register linkages from 2010-2021 Participants: 24,537,039 individuals age 16 or over, registered with general practices in the UK, including 141,119 with an autism diagnosis. 9,096,874 people age 16 or over in the Swedish Total Population Register, including over 100,817 with an autism diagnosis. Main outcome measures: Annual age-standardised incidence and prevalence of adult autism diagnoses within different sociodemographic groups. Annual age-standardised proportion of adults with new autism diagnoses, lifetime autism diagnoses, and no autism diagnoses, with prior records of other neuropsychiatric conditions or medications. Results: Incident adult autism diagnoses were consistently higher in Sweden than the UK, however incidence increased rapidly in the UK after 2020. Incident diagnoses increased fastest for 16-25-year-olds and females in both nations, as well as people in White ethnic groups in the UK and people with Swedish-born parents in Sweden. For example, in the UK in 2023 the age-standardised incidence of autism diagnoses among 16-65 years olds was 11 diagnoses per 10,000 person-years (95%CI: 10.7, 11.3) in the White ethnic group and 2.2 diagnoses per 10,000 person-years (95%CI: 1.9, 2.5) in the South Asian ethnic group. Over time there has been a consistent decline in the proportion of autistic adults with a prior diagnosis of epilepsy, psychosis and intellectual disability and an increase in the proportion with a prior diagnosis of ADHD, anxiety, depression and several other mental illnesses. For example, in the UK between 2010 and 2023 the age-standardised proportions of newly diagnosed autistic adults with prior records of epilepsy decreased from 10% (95%CI: 7.6, 13) to 4% (95%CI: 3.6, 4.5), while the proportion with records of anxiety increased from 28.7% (95%CI: 24.4, 33.6) to 58.3% (95%CI: 56.6, 60.1). Mental health conditions were generally more common in females and the reduction over time in intellectual disability was greater in females than males. Conclusions: The socio-demographic and neuro-psychiatric characteristics of individuals diagnosed as autistic have changed dramatically since 2010, a phenomenon observed both in the UK and Sweden. The extent to which these changes indicate nuanced recognition of autism or broadening of diagnostic practice needs investigation.

Abstract Background Gene dosage disorders impact cognition and psychopathology, but outcomes vary widely amongst carriers of the same variant. Recent work has sought to better predict proband outcomes using measures of corresponding traits in family members. However, family-based models have not yet been prospectively quantified across several traits in different genetic disorders, nor evaluated for the precision they afford: both crucial issues for clinical implementation. Methods In a first test case for these questions, we apply regression analyses to quantify and compare family-based prediction of 12 traits (including IQ, autism- and ADHD-related traits) in 433 individuals from families including a proband with XXY or XYY syndrome (N=93 and 58, respectively). Results The 12 traits vary substantially in their proband-family associations (0.001<|r|<0.55) - with differences emerging between XXY and XYY syndrome. Only two traits also show significant and similar proband-family associations in both aneuploidies, with the greatest concordance found for IQ. A family-based model for IQ prediction in male sex chromosome trisomies significantly reduces error vs. a group mean IQ model (F = 7.4, p = 0.006), but only in 65% of probands, and with mean error reduction of ~2 IQ points. Conclusions Family-based prediction of neuropsychiatric traits in genetic syndromes likely requires trait- and syndrome- specific models. Family models can significantly improve outcome prediction for IQ, but to variable degrees across individuals and with a small mean improvement. By mapping and quantifying these limits, our work helps draft a roadmap for refinement of family-based prediction of proband outcomes in gene dosage disorders.

Background: The ECAIL trial, launched in 2017, targets hard-to-reach families and evaluates a multicomponent childhood obesity prevention intervention. At a maternity hospital in Lille, France, healthcare providers screened pregnant women experiencing social vulnerability, and dietitians delivered a home-based intervention until age 2. The COVID-19 pandemic led to a six-month suspension in 2020. This study compared eligibility and participation before the pandemic and after resumption, and examined how the pandemic and subsequent cost-of-living crisis shaped implementation and reach. Methods: We analyzed 5,744 eligibility questionnaires distributed at the maternity ward. Inclusion criteria included [&ge;]1 indicator of social vulnerability (e.g., socioeconomic disadvantage, precarious housing, or social isolation). To capture implementation experiences, a psychosocial researcher conducted a focus group with six dietitians delivering the intervention; it was recorded, transcribed, and analyzed thematically focusing on reach, acceptability, and adaptation. Results: Eligibility increased from 29.7% (n=955) prepandemic to 33.6% (n=849) after resumption, while the distribution of vulnerability criteriaremainedsimilar across periods:78.3% received social/medical benefits; employment was not the main source of household income for 58.7%; 24.4% experienced financial hardship; 14.7% reported social isolation; 6.0% lived in precarious housing; and 19.0% had three or more vulnerabilities. Participation among eligible women remained stable (24.6%; n=443). Qualitative findings indicated dietitians satisfaction and participants enthusiasm for the resumption of home visits, particularly in addressing social isolation. After resumption, the introduction of a pre-visit COVID-19 questionnaire reduced missed appointments. Converging qualitative and quantitative findings indicated sustained, and in some cases strengthened, provider engagement despite pandemic-related strain on hospital services. Conclusions: This study shows that a complex intervention can maintain reach and acceptability through adaptive implementation under major contextual disruptions.The rapid resumption of home-based services emerged as a robust strategy for engaging and retaining socially disadvantaged families, highlighting the importance of flexible, context-sensitive approaches during social and economic crises.

BackgroundSETBP1 haploinsufficiency disorder (SETBP1-HD) is a neurodevelopmental condition characterized by developmental delay, speech apraxia, motor deficits, and autism spectrum disorder (ASD), caused by deficiency of the transcription factor (TF) SETBP1. Since current management is limited to symptomatic relief, we defined a robust consensus molecular signature for SETBP1-HD and prioritized drugs that converge with SETBP1 regulatory targets. MethodsWe performed a meta-analysis of three independent transcriptomic datasets derived from in vitro models of SETBP1 deficiency. Using Robust Rank Aggregation (RRA), we overcame transcriptomic heterogeneity to establish a SETBP1-HD consensus signature. We integrated this signature with known and spatio-temporal-aware SETBP1 interactions to prioritize key regulatory and drug targets. Finally, we screened the LINCS L1000 drug perturbation database for compounds capable of reversing the consensus SETBP1-HD gene signature. ResultsOur regulatory analysis yielded potential, critical therapeutic windows, while our consensus analysis identified LIN28A (a key regulator of developmental timing) and SPON1 ( F-spondin) as novel, robustly upregulated targets across models. LIN28A provides a potential mechanism for the dysregulated neurogenesis and progenitor proliferation observed in patient-derived cells, while SPON1 links molecular dysfunction to motor and connectivity deficits seen in patients. Our in silico drug repurposing screens prioritized celecoxib and buspirone as top therapeutic candidates. Mechanistically, celecoxib targets the prioritized SETBP1 regulatory target COX-2 (PTGS2) to mitigate neuroinflammation, while the potent neuromodulator buspirone chemically stabilizes vulnerable circuits to correct the underlying neurotransmitter imbalance driven by SETBP1 loss. LimitationsThis study relies on in silico consensus modeling derived from cell culture and animal models. While robust statistical thresholds were applied, the therapeutic efficacy of the identified candidates on behavioral phenotypes--specifically motor deficits and speech apraxia--requires in vivo validation. Additionally, the specific contribution of LIN28A to the ASD phenotype warrants further functional characterization. ConclusionsWe defined a unified molecular mechanism for SETBP1-HD, in which dosage deficiency leads to a delayed neurogenic state driven by LIN28A, coupled with connectivity defects driven by SPON1, ultimately resulting in a destabilized excitation/inhibition balance. We identify celecoxib and buspirone as top therapeutic candidates, offering an immediate translational pathway to address the core neurodevelopmental, circuitry, and motor challenges associated with SETBP1-HD variants.