BACKGROUND Large language models are increasingly deployed in mental health applications, yet growing evidence suggests they encode algorithmic biases that influence clinical outputs. Because these models now mediate patient-facing decisions, such biases carry the potential for direct harm. Whether they systematically affect psychiatric diagnosis across demographic groups remains underexplored. OBJECTIVE To examine whether large language models (LLMs) exhibit implicit demographic biases when generating psychiatric diagnoses. METHODS We developed 1,152 synthetic clinical vignettes using a matched-pair design that manipulated gender, race/ethnicity, age, socioeconomic status, English proficiency, and urbanicity while holding clinical content constant. Vignettes were divided into control (unambiguous anorexia nervosa) and ambiguous conditions designed to permit differential diagnosis. Ten LLM configurations across five model families were tested. RESULTS Control vignettes produced near-unanimous anorexia nervosa diagnoses (M = 100.0%), while ambiguous vignettes elicited greater variability (M = 23.6%). Inter-model agreement was moderate for ambiguous vignettes (Fleiss' κ = 0.410, 95% CI: 0.397–0.422). Mixed-effects logistic regression with LLM as a random intercept revealed significant demographic biases: Black patients were over six times more likely to receive a major depressive disorder diagnosis than White patients with identical presentations (OR = 6.09, 95% CI: 5.13–7.24), Latine patients were over nine times more likely (OR = 9.57, 95% CI: 8.00–11.45), and Asian patients were nearly three times more likely to receive an anorexia nervosa diagnosis (OR = 2.88, 95% CI: 2.44–3.42). Female patients were less likely than males to be diagnosed with anorexia nervosa (OR = 0.43, 95% CI: 0.37–0.49). CONCLUSIONS These findings demonstrate that LLMs exhibit systematic demographic biases in psychiatric diagnosis even when clinical content is held constant, revealing measurable patterns that can inform improvements to training data, model architecture, and clinical deployment frameworks.
This study uses sentiment analysis to examine changes in authorial stance over time in the flagship journal of rehabilitation counseling. The corpus linguistic analysis found overall positive sentiment, with higher happiness scores in 2021-2022 compared to 2019-2020. Other emotions remained constant. Implications and recommendations are discussed.
Background/Objectives: This study examines the academic discourse surrounding Critical Incident Stress Debriefing (CISD) and Critical Incident Stress Management (CISM) for first responders using Latent Dirichlet Allocation (LDA) topic modeling. It aims to uncover latent topical structures in the literature and critically evaluate assumptions to identify gaps and limitations. Methods: A corpus of 214 research article abstracts related to CISD/M was gathered from the Web of Science Core Collection. After preprocessing, we used Orange Data Mining software’s LDA tool to analyze the corpus. We tested models ranging from 2 to 10 topics. To guide interpretation and labeling, we evaluated them using log perplexity, topic coherence, and LDAvis visualizations. A four-topic model offered the best balance of detail and interpretability. Results: Four topics emerged: (1) Critical Incident Stress Management in medical and emergency settings, (2) psychological and group-based interventions for PTSD and trauma, (3) peer support and experiences of emergency and military personnel, and (4) mental health interventions for first responders. Key gaps included limited focus on cumulative trauma, insufficient longitudinal research, and variability in procedural adherence affecting outcomes. Conclusions: The findings highlight the need for CISD/M protocols to move beyond event-specific interventions and address cumulative stressors. Recommendations include incorporating holistic, proactive mental health strategies and conducting longitudinal studies to evaluate long-term effectiveness. These insights can help refine CISD/M approaches and enhance their impact on first responders working in high-stress environments.
As Taiwan’s population ages, traditional filial piety expectations face modernization challenges, yet few studies examine how emergent adults linguistically negotiate these cultural tensions digitally. This study addresses this gap by analyzing how emerging adults in Taiwan express and reinterpret filial obligations toward aging parents through online discourse. Emerging adults represent a particularly meaningful demographic because they straddle traditional filial norms and modern independence, making their language use a valuable indicator of cultural transition. We analyzed 1976 Dcard posts from 30 discussion threads (2017–2023) using computational linguistics. LIWC-22 assessed emotional expression patterns, while Sketch Engine conducted keyness analysis and collocation mapping around filial care keywords. Posts were compared against Chinese web corpus norms. Quantitative emotion analysis revealed dominant positive emotions (M = 3.93) alongside significant negative emotions (M = 3.30), with anger and sadness exceeding broader Chinese online communication norms. Keyness analysis identified economic concerns as central themes. Collocation analysis around “filial piety” showed associations with “limits”, “willingness”, and “define”, indicating conditional rather than absolute conceptualization. Findings indicate that emerging adults in Taiwan reinterpret filial piety through reciprocal emotional bonds rather than strict hierarchical duty, negotiating traditional expectations with contemporary economic realities and personal autonomy. The implications of these findings highlight how cultural values adapt in response to modernization and digital communication, offering insight into evolving intergenerational relationships and informing future cross-cultural aging and caregiving research.
The treatment landscape of attention-deficit/hyperactivity disorder (ADHD) has expanded to include digitally mediated interventions such as telehealth, mobile applications, neurofeedback, and serious games, yet whether digital mental health research is structurally integrated within the broader ADHD treatment ecosystem or developing in thematic silos remains an open question. This study applied BERTopic, an embedding-based topic modeling approach, to 2,551 peer-reviewed ADHD treatment abstracts published between 2014 and 2025 to map dominant research domains, examine community structure through topic network analysis, and evaluate longitudinal shifts in thematic prominence. Results reveal a landscape organized around high-centrality hubs (lifespan symptoms, cognitive training, and gender differences) that bridge biological, psychosocial, and clinical domains. Digital intervention topics show modest growth over time and remain only weakly connected within this network. They had little semantic overlap with high-risk comorbidity clusters including sleep disturbances, trauma, and substance use. Interpreted through a biopsychosociotechnical (BPST) framework, these findings characterize ADHD treatment research as occupying a transitional sociotechnical phase in which digital tools are expanding but not yet embedded within established care pathways. This structural map identifies candidate integration gaps for future research, including digital comorbidity overlap, implementation equity, and compulsive use risk. It also offers implications for the design, evaluation, and ethical deployment of digital ADHD interventions.
This study investigates the criterion concordance of the Bolante Threat Assessment Guide (B-TAG), a structured professional judgment (SPJ) tool designed to assess the risk of targeted violence. We compared ratings from seven trained human raters against ratings from three large language model (LLM)-based raters, each simulating a distinct professional perspective relevant to threat assessment. Both human and AI raters evaluated 12 anonymized case summaries against a criterion standard established by a majority-rule panel of three expert reviewers with more than 20 years of threat assessment experience each. Concordance with the expert panel was assessed at the case level using Gwet's AC1, selected over Cohen's kappa and Krippendorff's alpha for its robustness to prevalencerelated paradoxes under skewed base rates. Across the 12 cases, mean concordance with the expert panel was similar for human raters (M = 0.647) and AI raters (M = 0.604), though human raters agreed with the panel slightly more often. AI raters identified the risk indicators the panel judged present about as often as human raters did, but flagged more indicators the panel judged absent, producing more false positives. These findings suggest that AI may serve as a supplementary reviewer but is not a substitute for trained threat assessors, and offer preliminary evidence regarding the potential for AI-based tools to approximate expert human judgment in structured threat assessment coding, while underscoring specific limitations in the current evidence base.