Faculty Bibliography

Identifying updated, evidence-based risk and protective factors for psychotic disorders is essential to advancing scientific and clinical knowledge of their etiology. Since the most recent umbrella review published in this journal, the evidence base has expanded substantially. Following a pre-registered protocol, we searched the Web of Science for systematic reviews with meta-analyses of observational studies published between the search date of the last umbrella review (February 1, 2017) and March 31, 2025, examining associations of sociodemographic, parental, perinatal, and later factors or antecedents with ICD/DSM (any version) diagnoses of non-organic psychotic disorders. We graded associations between each putative factor and psychotic disorders using standardized international classification criteria: convincing (class I), highly suggestive (class II), suggestive (class III), and weak (class IV). We also conducted sensitivity analyses: a) lowering the classification criterion requiring more than 1,000 cases to 500 cases, b) restricted to prospective studies (class I-II factors), and c) testing more robust estimates of uncertainty (Hartung-Knapp-Sidik-Jonkman method). Then, we estimated the population attributable fraction (PAF) for potentially modifiable class I-III factors. Study quality was assessed with A MeaSurement Tool to Assess systematic Reviews (AMSTAR). A total of 69 systematic reviews and meta-analyses were included, reporting on 970 individual studies and 221 risk or protective factors for psychotic disorders. Six risk factors showed convincing (class I) evidence of association: Black-African ethnicity in England (odds ratio, OR=4.89, 95% CI: 4.04-5.92), South Asian ethnicity in England (OR=2.15, 95% CI: 1.69-2.73), paternal age <20 years (OR=1.34, 95% CI: 1.20-1.50), birthweight under 2,000 g (OR=1.77, 95% CI: 1.48-2.13), birthweight under 2,500 g (OR=1.50, 95% CI: 1.38-1.64), and maternal history of three or more previous pregnancies (OR=1.32, 95% CI: 1.20-1.45). Eighteen additional factors were highly suggestive (class II): maternal psychosis, Black-Caribbean ethnicity in England, any maternal mental health disorder, ethnic minority status in a low ethnic density area, first-generation immigrant status, cannabis use 5 to 7 days per week, Toxoplasma gondii IgG positivity, history of attention-deficit/hyperactivity disorder (ADHD), trait anhedonia, olfactory identification ability, premorbid IQ, and seven factors pertaining to minor physical anomalies. Two additional factors were upgraded from class IV to class II in our sensitivity analyses: clinical high-risk for psychosis (CHR-P) state, and soft neurological signs. Twenty-six factors (including four protective factors) were suggestive (class III), 75 were weak (class IV), and 96 were non-significant. Sensitivity analyses restricted to prospective studies downgraded some of the above factors, but only total minor physical anomalies became non-significant. The largest PAFs (>10%) were found for childhood adversities (40.69%), Toxoplasma gondii IgG positivity (20.47%), cannabis use 5-7 days per week (12.75%), and CHR-P state (12.68%). The sensitivity analyses employing robust uncertainty estimates confirmed that, among class I factors, Black-African ethnicity in England, birthweight under 2,000 g, and birthweight under 2,500 g retained their level of evidence. The mean AMSTAR score was 7.73 (SD=2.09). These findings provide an evidence-based, updated atlas of risk and protective factors for psychosis, advancing epidemiological knowledge, refining precision psychiatry, identifying targets for preventive intervention, and guiding global public health research.

BACKGROUND/UNASSIGNED:Radiation pneumonitis (RP) remains a significant treatment-related toxicity in patients with unresectable, locally advanced non-small cell lung cancer (NSCLC) undergoing chemoradiotherapy (CRT). Most existing predictive models rely on static baseline demographic or dosimetry variables and lack real-time clinical applicability. We developed a novel predictive framework that integrates longitudinal symptom data extracted from clinical notes using natural language processing (NLP) with clinical and dosimetry features to improve early RP prediction. METHODS/UNASSIGNED:We retrospectively identified 227 patients with locally advanced NSCLC treated with definitive CRT at a high-volume cancer center in the United States. We included all patients older than 18 years who were diagnosed between Jan 1, 2006, and Dec 31, 2022 with histologically or cytologically confirmed unresectable Stage 2 or 3 NSCLC and treated with conformal radiotherapy to a minimum dose of ≥45 Gy with or without chemotherapy. Of these, 31 RP events were identified through manual adjudication using radiologic criteria and chart review. NLP was used to extract the temporal relationship of 16 pre-specified symptoms with treatment from over 100,000 clinical notes spanning pre- and during-treatment intervals. We trained and validated machine learning models on combinations of baseline clinical data, radiation dosimetry, and NLP-derived symptom features. Model performance was evaluated using a nested cross-validation framework, with an outer cross-validation loop reserved for performance assessment and an inner cross-validation loop used for model training and integration, and summarized using area under the receiver operating characteristic curve (AUC) and partial AUC (pAUC) at high specificity thresholds. Clinical utility was evaluated using decision curve analysis (DCA). FINDINGS/UNASSIGNED:The best-performing model incorporated longitudinal NLP features and achieved a median AUC of 0.759 (90% confidence interval 0.753-0.766), significantly outperforming baseline models using only dosimetry (AUC 0.613) or clinical variables (AUC 0.635). NLP-based features such as cough trajectory, shortness of breath, and wheezing were among the most important predictors. Inclusion of NLP-derived symptom data improved early identification of high-risk patients, particularly in the clinically relevant high-specificity range (pAUC 0.021 vs. 0.010 for dosimetry alone). DCA showed that the calibrated MLP model provided greater net benefit than default strategies of treating all or no patients across clinically relevant threshold possibilities. INTERPRETATION/UNASSIGNED:In this early work, NLP-based extraction of longitudinal symptoms from routine clinical documentation meaningfully enhances RP prediction in patients undergoing CRT for NSCLC. This approach leverages existing electronic health record infrastructure to deliver real-time, scalable, and interpretable risk estimates, offering a pathway toward potential early intervention and personalized toxicity management. The model and DCA requires external and prospective validation before clinical deployment; as such, future work should focus on this validation and integration into clinical decision support systems. FUNDING/UNASSIGNED:AstraZeneca.

We present here the first bottom-up review of the lived experience of trans persons, co-written by experts by experience and academics. We sourced the medical and non-medical literature, focusing on first-person narratives, and critically extracted experiential themes, which were reviewed and appraised in collaborative workshops and eventually shared through a cloud-based system. Numerous experts by experience and academics from diverse gender, ethnic and cultural backgrounds were involved in the above steps and co-wrote the paper. The four overarching themes were: a) the lived experience of trans persons, b) the lived experience of gender affirmation, c) the experience of trans persons in the social and cultural context, and d) the lived experience of receiving care and support. The lived experience of trans persons is characterized by a rich diversity of gender identity experiences, which include identifying as male or female, or simultaneously as male and female and other genders, or as gender fluid. The first-person narratives indicate that common experiences may be of discovering an uncertain gender identity and feeling in the wrong body, together with some degree of mind-body disconnection. However, many trans individuals may feel satisfied with their bodies and identities. The lived experience of gender affirmation includes a process of searching for a bodily and mental redefinition of the self and - at times - fighting oneself to accept one's gender identity. Gender affirmation involves changes in the body to communicate the lived gender identity, as well as behavioral changes. The first-person narrative also highlights the emotional impact of gender affirmation, including the restoration of a sense of self, experiencing a strong emotional reaction to gender-affirming care, and, in a few cases, deciding to stop or reverse gender affirmation. The experience of trans persons in the social and cultural context can be characterized by an initial attempt to hide one's gender identity. This may be followed by coming out to others and wanting to be recognized by others in their lived gender. Exposure to high levels of transphobia, institutional discrimination and violence are commonly shared experiences. These include feeling wrong and dehumanized, pressured to "pass" as females or males, and suffering misgendering, deadnaming and verbal microaggressions. These experiences are frequently amplified by institutional discrimination and emotional, physical and sexual abuse. While some trans persons may struggle with family acceptance, others experience family and kinship structures as a safe base that provides support. Similarly, friends and romantic relationships may be experienced as sources of rejection or instead of happiness, resilience and restoration. The lived experience of seeking help and receiving general health care and support may be characterized by feeling rejected, disvalued and objectivized, but in other cases validated and cared for. Trans persons may feel lost in the maze of gender-affirming care and challenged when receiving a gender dysphoria diagnosis, although others experience a sense of relief. Mental health care may be feared, as individuals feel vulnerable and exposed, and regard it as a gatekeeping to receive gender-affirming care. In other cases, it is experienced as restorative and truly healing. The narratives provided in this paper have high educational and social value, informing medical and psychological practices and research as well as public health approaches, and promoting anti-discrimination policies and social change. This study gives voice to the lived experience of trans persons to overcome stigma and allow us to fully understand the varied nature of our gender expression and identity.

Schizophrenia and related psychoses occur in all human populations, with the highest rates of diagnosis among Black individuals and those of mainly African ancestry¹. Decades of research have established a highly heritable and polygenic basis for schizophrenia, which is mostly shared across populations^2-4. However, a recruitment bias towards European cohorts⁵ has led to discoveries that are poorly generalizable to African populations. This exclusion of the world's most genetically diverse populations narrows our understanding of disease biology and risks exacerbating health disparities. Here we show that electronic health records linked with genomic data from the Million Veteran Program (MVP)⁶-a national research programme that looks at the effects of genes, lifestyle, military experiences and exposures on the health and wellness of veterans-enable a comprehensive assessment of schizophrenia genetics in populations of African ancestry in the USA. We identify ancestry-independent associations in African populations and expand the catalogue of implicated regions by more than 100 loci. Through statistical fine-mapping and integrative transcriptomic analyses, we refine disease-associated signals to consensus genes with convergent neurobiological functions. These findings provide a much-needed view of schizophrenia's genetic architecture in populations of African ancestry, and offer biological insights that both extend previous work and broaden its global relevance.

Prenatal inflammation is a recognized risk factor for autism spectrum disorder (ASD), but the neural pathways linking in utero immune exposure to later outcomes remain unclear. We examined whether chronic placental inflammation (CPI), a sustained immune response within the placenta, is associated with fetal brain functional connectivity differences and later ASD traits. A cohort of 105 pregnant individuals (85.7% African American/Black, 1.0% Asian American, 4.8% Bi-racial, 4.8% White, 1.0% Other, 2.9% Missing) underwent fetal resting-state fMRI between 24 and 38 weeks' gestation. After delivery, placental tissues were evaluated by expert pathologists for CPI lesions. Fetuses exposed to CPI (n = 44) showed significant differences in resting-state functional connectivity (RSFC) compared to non-exposed fetuses (n = 61) across ten network pairs, with the greatest effects in prefrontal, cerebellar, visual, and motor regions. These connectivity differences were associated with behavioral outcomes: cerebellar-medial visual RSFC was the strongest prenatal predictor of ASD traits at age 3. An exploratory structural equation modeling indicated that CPI was associated with reduced cerebellum-medial visual RSFC, which was associated with higher ASD trait scores, though the indirect pathway did not reach statistical significance. This study provides the first human evidence that chronic placental inflammation is associated with large-scale fetal brain network differences that are also associated with autism traits, suggesting that prenatal immune activity may be relevant to neurodevelopmental trajectories.

BACKGROUND/UNASSIGNED:Adolescence is a period of heightened risk for psychopathology. Recent studies suggest shared neural correlates across psychiatric disorders in brain networks involved in emotional and cognitive processes. However, the visual network-crucial for extracting and processing emotional information-has received less attention. Here, using a transdiagnostic approach, we focus on the visual network to investigate how youth with psychopathology represent and differentiate between emotional face categories during a cognitively demanding task, for which emotional content is irrelevant. METHODS/UNASSIGNED:= 4952 after quality control). Symptoms of psychopathology were based on the Child Behavior Checklist. Participants performed an emotional n-back task in the scanner. Representational similarity analysis was performed to examine dissimilarities in neural representations between emotional categories in functional brain networks. RESULTS/UNASSIGNED:< .001), and predicted symptoms a year later. CONCLUSIONS/UNASSIGNED:Using a transdiagnostic approach, we showed that the distinctiveness in neural representations of task-irrelevant emotional information in the visual cortex is positively correlated with psychopathology. Together, our findings suggest the importance of the visual network and early visual areas in understanding neural mechanisms of emotional processing in youth with psychopathology.

Family-based interventions have been shown to improve behavioral and justice-related outcomes for youth. There has been less focus on whether family-based interventions are effective at reducing youth recidivism, especially for newer or lesser known interventions. To address this gap, we conducted a systematic review of studies utilizing family-based interventions to address recidivism for youth involved in the juvenile justice system. Of the identified 3,497 unique articles, 235 full-text articles were reviewed and 31 met criteria for inclusion in the review. A variety of family-based interventions were used, including both widely known (e.g., multisystemic therapy) and lesser known interventions. Most studies demonstrated significant reductions in recidivism, with some mixed findings depending on the type of charge, assessment timepoint, measure of recidivism, or moderating factors. The findings from this review corroborate prior research on the effectiveness of well known interventions for reducing recidivism and highlight the promise of newer interventions. The results also highlight the need for additional research in this area, as well as a greater focus on better understanding the moderating factors that may affect treatment outcomes.

Study DesignRetrospective cohort study.ObjectivesTo characterize longitudinal durability of cervical deformity (CD) correction, identify predictors of alignment failure, and define failure mechanism timing across the first postoperative year.Methodsof 155 operative CD patients in a prospectively maintained multicenter database, 89 had complete four-timepoint radiographic follow-up. Well-alignment was defined as cSVA <4 cm. Patients were grouped as A (maintained), B (corrected then deteriorated), or C (never aligned). Analyses included Cox regression (n=47, 17 events), Kaplan-Meier with restricted mean survival time (RMST), and mixed-effects logistic regression across four timepoints (n=356; ICC=0.36). Concurrent thoracolumbar deformity was excluded.ResultsWell-aligned proportion fell from 48.3% at 3 months to 33.7% at 1 year (A=34%, B=19%, C=47%). DJK was most strongly associated with alignment loss (HR 5.21, 95% CI 1.94-13.98; OR 0.09), with alignment maintained 4.1 versus 9.8 months (RMST difference 5.7, p=0.001). Lower postoperative cSVA paradoxically increased failure hazard (HR 0.58/10 mm, p=0.002). Older patients deteriorated faster (Time×Age OR 0.88, p=0.022). Failure mechanisms shifted from DJK at 3 months (50%) to PJK at 6 months (75%) to global compensation loss at 1 year (71%). No surgical parameters differed between groups.ConclusionsFewer than one-third of patients maintain alignment at 1 year. DJK, advanced age, and lower postoperative cSVA were most strongly associated with failure (17 events; limited precision). Mechanisms appeared to shift from junctional failure early to global decompensation late. Alignment loss was not accompanied by significant 1-year NDI or EQ-5D differences, framing radiographic deterioration as an early surveillance marker rather than clinical failure. Individualized risk stratification and surveillance are supported.

IMPORTANCE:Although fluoride exposure at high levels during the perinatal period has been associated with socioemotional problems among children, there is limited evidence regarding potential associations at lower levels of fluoride exposure, particularly in the US. OBJECTIVE:To examine the association between prenatal exposure to public drinking water fluoride levels and children's socioemotional problems in a pooled US pediatric consortium. DESIGN, SETTING, AND PARTICIPANTS:This cohort study used observational pregnancy cohort data from January 1, 2005, to December 31, 2019, on children aged 1.5 to 17 years from the Environmental Influences on Child Health Outcomes (ECHO) Cohort, which took place at 20 prenatal pediatric cohort sites across 34 US states. Statistical analysis was conducted from December 2024 to December 2025. EXPOSURE:Area-level, time-weighted, mean fluoride concentrations in public water were estimated based on residential addresses during pregnancy. MAIN OUTCOMES AND MEASURES:Children's socioemotional problems were assessed using caregiver reports of internalizing and externalizing problems via the Child Behavior Checklist. Generalized estimating equation models evaluated the adjusted mean difference in T scores (1) across the range of exposure values in cubic spline models, (2) per 500.0-µg/L higher fluoride level in linear models, (3) across quartiles of exposure, and (4) at the US Public Health Services recommended cutoff (≤700.0 µg/L; 3998 less than the cutoff). RESULTS:The cohort comprised 5520 children (mean [SD] age, 5.8 [3.0] years; 2849 boys [51.6%]). The mean (SD) prenatal time-weighted water fluoride level was 407.9 (361.0) µg/L (range, <1.4-3272.5 µg/L). The adjusted mean difference in emotional and behavioral problem T scores per 500.0-μg/L higher exposure to prenatal public water fluoride was -0.15 (95% CI, -0.85 to 0.56) for internalizing problems and 0.06 (95% CI, -0.44 to 0.55) for externalizing problems, consistently showing no associations at the continuous level. However, some positive associations were observed only when using change point models for externalizing problems (adjusted mean difference, 1.59 [95% CI, 0.72-2.47] per 500.0-µg/L increase above the inflection point), and for some subgroups. CONCLUSIONS AND RELEVANCE:In this cohort study, prenatal exposure to fluoride levels in regulated public water was not associated with internalizing and externalizing problems among offspring. In change point models, positive associations were found for externalizing problems. The results contribute to growing evidence that could inform future policy decisions regarding fluoride in the US public water systems. Additional studies are needed to further investigate this association, especially examining other sources of fluoride that may have an association with children's emotional and behavioral problems.