JMIR Mental Health

Modeling of Therapist-Client Dynamics in Psychotherapy Using Large Language Model–Based Assessments: Retrospective Observational Study of 1573 Psychotherapy Sessions

2026-11-05T16:00:22-05:00

Background: Psychotherapy is a primary treatment for many mental health conditions, yet the dynamic interplay among therapist behaviors, client responses, and the therapeutic relationship remains difficult to disentangle at scale. Traditional process research relies on labor-intensive human coding, limiting the ability to model moment-to-moment therapeutic dynamics across large datasets. Objective: This study aimed to develop and validate a computational framework for modeling therapist-client interaction processes and to examine how therapist behaviors and relational qualities relate to client self-disclosure and emotional expression in psychotherapy sessions. Methods: We developed automated assessment pipelines using large language models (LLMs) to measure therapist behaviors (empathy or exploration), relational quality (rapport), and client outcomes (self-disclosure, self-directed negative emotion, or outward-directed negative emotion). Model-generated scores were validated against human annotations. We then applied these measures to approximately 2000 hours of psychotherapy transcripts from the Alexander Street corpus. This yielded an estimation sample of 119,233 client utterances from 1573 sessions and 37 clients. Because utterances are nested within sessions and sessions within clients, we estimated a saturated multivariate structural equation model with SEs clustered at the session level. Results were confirmed in a multilevel regression model with random intercepts for session and client. Results: Following comparative validation of 6 candidate language models, automated measures showed good to excellent agreement with human annotations (intraclass correlation coefficient [ICC]=0.60-0.81) for rapport (ICC=0.81), self-disclosure (ICC=0.78), empathy components (ICC=0.60-0.75), and emotion measurements (ICC=0.66-0.80). Structural equation modeling analyses revealed that therapist empathy and exploration were each associated with increased client disclosure (empathy: β=.086, 95% CI 0.078-0.095; exploration: β=.086, 95% CI 0.078-0.094; both P<.001) and increased expression of self-directed negative emotions, comprising sadness, anxiety, fear, and depression (empathy: β=.118, 95% CI 0.110-0.127; exploration: β=.078, 95% CI 0.070-0.087; both P<.001). Both behaviors were also associated with reduced expressions of enjoyment (empathy: β=–.037; exploration: β=–.041; both P<.001). Surprisingly, rapport in the prior session showed no association with client self-disclosure or negative emotion expression. Conclusions: This study demonstrates that LLM-based measurement combined with structural modeling can capture core therapeutic processes at scale. Empathy and exploration showed small but consistent turn-level associations with client disclosure and emotional expression, whereas prior-session rapport did not show a robust association. These findings provide a computational foundation for precision modeling of psychotherapy processes and may offer actionable insights for scalable therapist training and AI-supported clinical education systems.

The Effect of Bipolar Disorder Illness Course on Physiological Digital Signals: Observational Study

2026-11-04T17:45:20-05:00

Background: Although bipolar disorder (BD) lacks objective biomarkers of illness progression, wearable devices enable continuous, real-world collection of physiological digital biosignals, offering a promising approach to address this gap. Measures such as electrodermal activity (EDA) and heart rate variability (HRV) fluctuate with mood state and symptom severity. Previous studies have reported reduced EDA during depressive episodes and increased HRV following symptom resolution after acute mood episodes. Furthermore, greater illness burden in BD has been associated with lower HRV, suggesting that digital biosignals may capture not only state-related changes but also aspects of illness course. Objective: This study investigated whether naturalistic EDA and heart rate (HR) profiles differ according to illness course variables in BD—including diagnosis, predominant polarity, years since onset, and lifetime number of episodes—in euthymic patients, to evaluate their potential as illness course biomarkers. Methods: This cross-sectional post hoc analysis used data from 80 euthymic patients on the BD spectrum from the INTREPIBD-TIMEBASE cohort. Participants were recruited at Hospital Clínic de Barcelona (Spain) between 2021 and 2024 and wore an Empatica E4 wristband for 48 hours in naturalistic settings. Inclusion criteria comprised a current or past (, Fifth Edition) diagnosis of BD confirmed by structured interview and sustained euthymia. We examined the effects of 4 illness course variables (diagnosis, predominant polarity, years since onset, and number of lifetime episodes) on 3 biosignals (mean EDA, EDA peaks per minute, and HR) using generalized additive models. Interactions between illness course variables and circadian patterns were also assessed. All models were adjusted for age, sex, and current antipsychotic use and included a participant-specific random effect. Results: Of the 80 participants, 49 (61.3%) were female; participants’ mean age was 46.9 (SD 14.1) years. After adjustment for covariates and multiple comparisons, no significant associations were observed between illness course variables and EDA or HR. In sensitivity analyses excluding patients with schizoaffective disorder (n=75), years since illness onset and predominant polarity were associated with EDA peaks per minute after multiple comparison correction. Specifically, patients with more than 22 years since illness onset had fewer EDA peaks compared with other groups (β=−1.06, 95% CI −1.71 to −0.41; =.001). No significant interactions were found between illness course variables and circadian patterns of EDA or HR. Conclusions: To our knowledge, this is the first study to examine the association between illness course variables and EDA and HR recorded in naturalistic settings in BD. In contrast to prior findings linking illness burden with HRV, our results suggest that years since illness onset may show modest associations with phasic EDA activity. These findings indicate that digital biosignals may be more sensitive to state-related changes than to illness course characteristics, limiting their use as longitudinal biomarkers of disease progression.

Social Media Use Motives and Mental Health Among Adolescents: Cross-Sectional Study

2026-10-28T16:45:16-04:00

Background: Social media use is ubiquitous among young people, yet its relationship with mental health may depend on why it is used. Understanding the motives underlying social media engagement may provide greater insights into its potential risks and benefits than considering overall use alone. Objective: This study aimed to identify latent dimensions of motives for social media use and examine their associations with anxiety, depression, well-being, and resilience in a sample of older adolescents. Methods: Cross-sectional data were drawn from the BrainWaves study (N=6820; mean age 16.72, SD 0.74 years). Social media use motives were assessed using 10 items and exploratory factor analysis identified 2 latent dimensions: avoidant-escapist use (boredom, addiction, distraction, and procrastination) and social-recreational use (socializing, communication, entertainment, and observation). Confirmatory factor analysis was used to estimate factor scores, which were examined using linear regression. Structural equation modeling estimated associations with anxiety, depression, well-being, and resilience, with secondary analyses accounting for gender and deprivation. Results: Linear regression analyses showed that avoidant-escapist motives were associated with higher anxiety and depressive symptoms, whereas social-recreational motives were associated with lower symptoms. No interactions with gender or deprivation were observed. The structural equation model showed acceptable model fit (comparative fit index=0.945; root mean squared error of approximation=0.057; standardized root mean squared residual=0.054). Avoidant-escapist motives were associated with higher anxiety (β=0.543) and depressive symptoms (β=0.629) and lower well-being (β=−0.610) and resilience (β=−0.469). Conversely, social-recreational motives were associated with lower anxiety (β=−0.228) and depressive symptoms (β=−0.342) and higher well-being (β=0.362) and resilience (β=0.242; <.001 in all cases). These patterns remained consistent after accounting for gender and deprivation. Conclusions: Social media use shows different associations with older adolescent mental health depending on underlying motives. Focusing on why young people engage with social media rather than on overall use alone may provide a more informative framework for understanding potential risks and benefits. Given the cross-sectional design, causal inferences cannot be made.

Meta Mesh Ontology for Cross-Domain Mental Health Knowledge Integration: Development and Evaluation Study

2026-10-23T16:15:19-04:00

Background: Psychiatric diagnostic practice faces persistent challenges in adequately capturing the complexity, heterogeneity, and dynamic nature of mental disorders. Conventional categorical classification systems such as the and the rely on rigid symptom-based criteria that often fail to represent multifactorial etiologies, dimensional symptom profiles, and high rates of comorbidity. Although dimensional models, domain-specific ontologies, and alternative taxonomies have been developed to address these limitations, they are typically applied in isolation, resulting in fragmented knowledge representations and limited interoperability. This lack of semantic integration hampers data harmonization and cross-domain analysis. Objective: The objective of this work is to develop and present the Meta Mesh Ontology (MMO) and its Core Domain Framework (CDF) as an integrative approach for mapping heterogeneous psychiatric knowledge bases and evaluating ontology structure, logical consistency, and specified retrieval and inference behavior. Methods: We designed a modular ontology architecture grounded in the Basic Formal Ontology to ensure structural and semantic consistency across biological, psychological, and social domains. Existing psychiatric frameworks and ontologies—including dimensional models, diagnostic taxonomies, and socioenvironmental knowledge systems—were systematically analyzed and represented through domain modules using a documented alignment and mapping procedure. The MMO was implemented using formal semantic web standards, including the Web Ontology Language and the Resource Description Framework (RDF), and aligned with established biomedical ontologies and informatics standards to support explicit mappings and modular extensions across domains. Evaluation used 2 MMO and CDF metric sets and 10 module metric sets to quantify ontology size, semantic alignments, external integration, and dataset readiness. We also assessed Web Ontology Language 2 Description Logic (OWL 2 DL) conformance, logical consistency, and 11 competency questions (including 1 entailment test) and 4 negative controls, in an outpatient scenario with 50 variables. Results: The MMO provides a shared semantic framework for representing links among heterogeneous psychiatric knowledge sources. Its modular design allows domain-specific subontologies to be incorporated under a common architecture. MMO and CDF comprised 2238 triples, 208 classes, 28 properties, 3 individuals, and 103 alignments (100 external) across 5 frameworks. The 10 modules added 967 triples and 171 alignments (116 external), with 21 integrations and complete label coverage (54/54 resources). The ontology resources conformed to OWL 2 DL; HermiT (Information Systems Group) detected no logical inconsistency, and all 11 competency questions and 4 negative controls produced the expected results. All 50 variables were linked to at least 1 CDF anchor, and 47 had an explicit Simple Knowledge Organization System (SKOS) mapping. Conclusions: The MMO provides a modular foundation for representing links among diverse psychiatric knowledge systems and data modalities. The current evaluation characterizes its structure, mappings, and behavior in the specified tests. Further research should assess mapping validity, governed clinical data, operational performance, and clinical utility.

Incorporating Data From Text Messages and Email Into Psychotherapy: Randomized Controlled Trial

2026-10-19T16:15:20-04:00

Background: As digital communication becomes central to daily life, psychotherapy increasingly has access to patients’ electronic media data. Although digital phenotyping has been widely studied, less is known about whether incorporating personal communication data, such as text messages, improves clinical outcomes in psychotherapy. Objective: This study aimed to determine whether integrating personalized text message data into psychotherapy improves depression, anxiety, health-related quality of life, and therapeutic alliance and to examine whether increased access to collateral information influences clinical decision-making. Methods: The Electronic Media and its Impact on Psychotherapy (EMAP) study was a randomized controlled trial conducted in adult and geriatric outpatient and partial hospitalization settings at an academic psychiatric hospital. Participants receiving psychotherapy for primary depression or anxiety (n=101) were randomized to electronic media–enhanced therapy (EME; n=52) or treatment as usual (TAU; n=49). In EME, research staff reviewed participants’ text messages prior to sessions using personalized mood-related key terms and presented aggregated findings on a Health Insurance Portability and Accountability Act (HIPAA)–compliant clinician dashboard. TAU participants received standard care without message review. Outcomes included change scores for the Patient Health Questionnaire-9 (PHQ-9), the Generalized Anxiety Disorder-7 (GAD-7), the 36-item Short Form Health Survey (SF-36), and the Working Alliance Inventory–Short Revised (WAI-SR). Linear regression and nonparametric tests compared groups. Moderation analyses assessed whether the frequency of information access influenced treatment changes. Results: No statistically significant between-group differences were observed in anxiety or working alliance. Among participants with SF-36 follow-up data (n=65), EME participants demonstrated greater improvement in the pain subscale. Greater frequency of information access was associated with increased treatment-related changes in the EME group. Access to additional collateral information was associated with more clinical actions but not with improved distal outcomes. Conclusions: Integrating personalized text message data into short-term psychotherapy did not significantly improve depression or anxiety, and it neither improved nor worsened therapeutic alliance. However, access to greater amounts of collateral information, including electronic media, influenced clinical decision-making, suggesting nuanced and context-dependent effects. Trial Registration: ClinicalTrials.gov NCT03712267; https://clinicaltrials.gov/study/NCT03712267

“There is a Place for it”—Practitioners’ Perspectives on Situated Ethical Acceptability of AI Technologies for Mental Health Practice: Qualitative Study

2026-10-19T16:00:03-04:00

Background: Design research on AI technologies for mental health has expanded rapidly, yet existing work has largely focused on practitioners’ perspectives within isolated therapeutic stages or specific services. As a result, there is limited understanding of how practitioners view the role of AI across the full therapeutic journey, including assessment, diagnosis, and treatment, or how these perspectives can be translated into actionable and ethically grounded design guidance. This lack of stage-specific guidance introduces several risks: (1) the development of AI tools that practitioners may not trust or adopt, (2) heightened ethical concerns, particularly in diagnosis where bias and erosion of human judgment are most likely, and (3) missed opportunities to leverage AI to support ongoing assessment and treatment adaptation. Objective: This study aimed to explore mental health practitioners’ perspectives on how AI may support or hinder therapeutic practice across all 3 therapeutic stages, with particular attention to their ethical considerations. Methods: We conducted semistructured interviews with 18 mental health practitioners to explore how AI may support or hinder therapeutic practice across assessment, diagnosis, and treatment. Data were analyzed thematically, and ethical challenges were organized using 4 biomedical principles. This study received approval from the Lancaster University Institutional Ethics Committee (ID 3530). Results: Findings suggest a stage-specific pattern: practitioners identified valuable roles for AI in supporting clinical assessment and treatment, while expressing strong resistance to autonomous AI diagnosis, which was regarded as the most ethically sensitive stage. Across contexts, AI was seen as supporting and validating, rather than replacing, clinical judgment. Ethical concerns were organized using these principles, with key issues including data validity, inclusivity, privacy, transparency, and explainability. Grounded in these findings, four design directions are proposed: (1) AI-enabled clinical assessment hubs for analyzing multimodal patient data to support transparent decision-making; (2) natural language processing-based tools to enable richer understanding of patients’ lived experiences; (3) new models of trust to support ethically sensitive AI-assisted diagnostic decision-making; and (4) large language model–powered personalized and adaptive mental health support systems that maintain human-AI partnership and preserve practitioners’ central role in care. Conclusions: This study provides a stage-specific understanding of practitioners’ perspectives on AI use across assessment, diagnosis, and treatment and systematically frames ethical challenges using biomedical principles. The findings translate practitioner insights into actionable design guidance for ethically informed AI-supported mental health practice.

Incorporating Objective Behavioral and Biological Measures in a Large-Scale Decentralized Mobile Health Trial for Depression: Randomized Controlled Trial

2026-10-09T17:30:26-04:00

Background: Decentralized (ie, fully remote) clinical trials (DCTs) offer a scalable and ecologically valid approach to evaluating digital mental health interventions (DMHIs). Yet, incorporating objective measures of well-being within DCTs poses novel methodological and feasibility challenges that can limit the rigor and precision of data collection in the field. Objective: We recently conducted the Behavior, Biology, and Well-being (BeWell) Study, a large-scale, 3-arm randomized DCT evaluating the Healthy Minds Program app (Humin)—a self-guided DMHI that includes didactic well-being information and meditation components. The current paper’s objective is to describe the methodological approach to collecting and processing a multimodal battery of psychological (questionnaires and structured clinical interview), biological (blood and stool assays), and behavioral (stimulus-elicited affect [SEA] task, video log, and Mnemonic Similarity Task [MST]) measures, longitudinally across baseline, the 4-week intervention period, and 17-week follow-up. Secondarily, we evaluate its feasibility. Methods: The BeWell Study methodology is described. Descriptive statistics characterized recruitment and retention, as well as data completion and data quality. Binary logistic regressions evaluated whether recruitment sources, demographics, and depression severity predicted measure completion and data quality. Results: From June 1, 2022, through July 15, 2024, 52,052 individuals started the screening questionnaire. Of these, 2.22% (n=1156 participants; mean age 38.69, SD 12.14 years; women=75.50%; non-Hispanic White=68.88%) completed the screener, met eligibility criteria (including elevated depressive symptoms [Patient Health Questionnaire-9 [PHQ-9] score≥5]), completed baseline measures, were enrolled, and were randomized. Participants were recruited from Craigslist (694/1154, 60.14%), a US-based advertisement website, and university email messages (107/1154, 9.27%); participants were enrolled from every US state. Across the study, 84% (971/1156) of participants were retained (completed ≥1 measure at 17-week follow-up). Across time points, ≥81.40% (941/1156) of participants completed questionnaires and ≥75.17% (869/1156) completed biological samples. Of those completing questionnaires, ≥84.14% (796/946) completed SEA and video log tasks and ≥91.75% (868/946) completed the MST. Of completed assessments, high-quality or analyzable data rates across all time points were ≥97.89% (926/946) for questionnaires, ≥90.20% (773/857) for SEA, ≥84.60% (725/857) for video logs, ≥94.20% (1056/1121) for the MST, ≥98.84% (849/859) for dried blood spots, and ≥99.43% (865/870) for microbiome samples. Greater depressive symptoms at screening, but not age, predicted completion of baseline measures (odds ratios 0.96‐0.99; <.001). For enrolled participants, recruitment sources and baseline depressive symptoms were not predictive of data completion or quality. Older age and having a college degree predicted significantly higher rates of data completion across all data types (questionnaire, behavioral, and biological) at week 4 or week 17 follow-ups. Conclusions: This methodological approach offers one model of incorporating objective measures of well-being into DCTs testing DMHIs. Implications and limitations of our approach are discussed. Trial Registration: ClinicalTrials.gov NCT05183867; https://clinicaltrials.gov/study/NCT05183867

Ensemble of Domain-Specific Natural Language Processing and Large Language Models for Detecting Suicidal Ideation and Mental Health Conditions in Social Media Text: Development and Evaluation Study

2026-10-08T17:30:58-04:00

Background: Mental illness contributes substantially to global disability, and public adoption of AI for mental health support is accelerating without commensurate safety evaluation. General-purpose large language models can fail to recognize naturalistic text expressing suicidal ideation, with immediate clinical consequences. Domain-specific natural language processing methods offer a contrasting approach, but ensembles integrating the two have not been formally characterized. Objective: We aimed to (1) benchmark 5 modeling approaches for classifying mental health–related social media text, (2) develop and optimize ensembles integrating a domain-specific natural language processing classifier with a fine-tuned large language model, and (3) derive a closed-form framework constraining ensemble weighting in safety-critical classification. Methods: We analyzed 60,889 publicly available Reddit and Twitter texts spanning 9 categories (anxiety, bipolar disorder, depression, a normal baseline, personality disorder, stress, suicidal ideation, attention-deficit/hyperactivity disorder, and autism spectrum disorder). Labels were derived from the originating subreddit or a self-stated condition and are not clinical diagnoses. Five architectures were compared: prompt-engineered base GPT-4o-mini; fine-tuned GPT-4o-mini; a domain-specific classifier (support vector machine over symptom-informed lexical features); and 2 ensembles of these models, hybrid probability–indicator fusion and soft probability fusion. Data were split into 70%, 10%, and 20% at the post level (n=42,622, n=6089, and n=12,178); weights were selected on the validation split and all reported performance comes from the held-out test split. Closed-form bounds on permissible language model weights were derived for both strategies. CIs are paired bootstrap intervals (n=10,000 replicates); accuracy differences were tested with exact McNemar tests. Results: The domain-specific classifier achieved 91.9% accuracy (95% CI 91.4%-92.4%), exceeding the base (58.7%) and fine-tuned (88.6%) large language models. Both ensembles outperformed either constituent model, reaching 93.9% (soft fusion; 95% CI 93.5%-94.3%) and 93.8% (hybrid fusion; 95% CI 93.4%-94.2%) at a weighting of 60% classifier to 40% fine-tuned model, selected on validation (P<.001 for both against the classifier alone; the 2 ensembles did not differ from each other [P=.16]). The hybrid ensemble reduced the miss rate for the suicidal ideation category from 32.5% (650/2000) to 1.9% (37/2000), a 17.6-fold reduction relative to the base model, although its advantage over the classifier alone within that category was not significant (P=.68). Accuracy fell discontinuously above a 50% language model weight, where the hybrid ensemble reduced exactly to the fine-tuned model; the selected weight satisfies the derived instance-level bound of 44.0%. Conclusions: Domain-specific clinical grounding remained necessary for stable ensemble performance in this corpus, and probabilistic fusion outperformed both constituent models when language model weight was bounded by a derivable, model-agnostic constraint. Because labels were community-inferred and only one corpus was analyzed, these results are hypotheses requiring external validation on clinically characterized data before use in decision support.

Effectiveness of Chatbots in Mental Health Screening and Assessment: Systematic Review

2026-10-06T15:00:18-04:00

Background: Mental health disorders affect 970 million people globally; yet, over 50% do not access timely evaluation due to structural barriers and professional shortages. Chatbots and AI-based conversational agents have emerged as promising tools for mental health screening and assessment. Objective: This study systematically evaluated the effectiveness, accuracy, reliability, and acceptability of chatbots and AI-based conversational agents for mental health screening and assessment in adults. Methods: Systematic search conducted in May 2025 across PubMed/MEDLINE, PsycINFO, Scopus, and Web of Science (2019‐2025), following PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 guidelines. Eligible studies evaluated chatbots or AI for mental health screening/assessment in adults (≥18 y). Risk of bias was assessed using appropriate tools (Risk of Bias 2, Quality Assessment of Diagnostic Accuracy Studies-2, Joanna Briggs Institute checklists, and Mixed Methods Appraisal Tool). This systematic review was registered with PROSPERO (International Prospective Register of Systematic Reviews; CRD420251072392). Results: Eighteen studies (2021‐2025) were included, with samples ranging from 20 to 3902 participants. Rule-based chatbots demonstrated high reliability (Cronbach α >0.85) and good acceptability (Acceptability of Intervention Measure >19/25). Generative models (large language models) achieved sensitivities of 0.84 to 0.93 and specificities of 0.80 to 0.96 for depression and anxiety, with correlations up to =0.96 with expert clinicians in suicide risk assessment. Hybrid approaches combining large language models with machine learning achieved exceptional performance for cognitive impairment (-score of 92.1%, specificity of 99.6%). Most studies reported high user satisfaction (≥70%), although barriers existed among older populations. Methodological quality was heterogeneous with a moderate risk of bias in critical dimensions. Conclusions: Chatbots and AI conversational agents demonstrate clinically relevant performance in mental health screening and assessment. However, safe implementation requires clear clinical protocols, professional supervision, integration with electronic health records, and active mitigation of algorithmic bias. These technologies should complement rather than replace clinical judgment. Trial Registration: PROSPERO CRD420251072392; https://www.crd.york.ac.uk/PROSPERO/view/CRD420251072392

Electronic Health Record–Based Phenotyping for Obsessive-Compulsive Disorder: Algorithm Development and Multicenter Validation Study

2026-10-02T15:15:11-04:00

Background: Obsessive-compulsive disorder (OCD) is a common psychiatric disorder, with two-thirds of affected individuals reporting severe impairment. Despite its substantial burden and moderate heritability, the etiology of OCD remains poorly understood, and treatments are often suboptimal. Although recent genome-wide association studies (GWAS) have identified some risk loci, much of the genetic architecture of OCD remains undiscovered, underscoring the need for scalable approaches to identify large, well-defined patient cohorts. Objective: This study aimed to develop and validate a scalable electronic health record (EHR)–based phenotyping algorithm for identifying OCD cases to support large-scale genetic and translational research. Methods: We leveraged EHR-linked biobank data from 2 large hospital systems, namely Vanderbilt University Medical Center (VUMC) and Mass General Brigham (MGB), to develop a high-throughput phenotyping algorithm integrating diagnostic codes, medication records, and natural language processing (NLP) of clinical notes. Algorithm performance was evaluated through expert chart review, and genetic analyses were performed in individuals of European genetic ancestry using the polygenic scores (PGS) of OCD, major depressive disorder (MDD), and height derived from the most recent GWAS. Results: Expert chart reviews demonstrated our algorithm combining both () codes and NLP achieved the highest positive predictive values (PPV) for OCD case identification (0.84 at VUMC; 0.91 at MGB) compared to using either codes or NLP alone, albeit with reduced case yield. At both sites, algorithm-defined OCD cases of European genetic ancestry showed significantly higher OCD PGS than controls. In sensitivity analyses adjusting for MDD status, OCD PGS associations were more robust than MDD PGS associations, while height PGS showed no association, supporting the genetic plausibility and relative specificity of the phenotype. Conclusions: This study presents a scalable and cost-efficient EHR-based approach for identifying OCD cases across health systems. The algorithm achieves high PPV, and among individuals of European genetic ancestry, algorithm-defined cases show significant OCD PGS enrichment, supporting its utility for large-scale genetic studies and advancing understanding of the disorder’s complex etiology.