JAMIA Open最新文献

{"title":"A framework for assessing algorithmic discrimination risks in training data: a case of pediatric type 1 diabetes.","authors":"Ioannis Bilionis, Ricardo C Berrios, Antonio de Arriba Muñoz, Luis Fernandez-Luque, Carlos Castillo","doi":"10.1093/jamiaopen/ooag125","DOIUrl":"10.1093/jamiaopen/ooag125","url":null,"abstract":"<p><strong>Objectives: </strong>To develop a generalizable framework for identifying algorithmic discrimination risks arising from subgroup imbalances in machine learning training data, with relevance to medical informatics applications where heterogeneous real-world data can bias model behavior.</p><p><strong>Materials and methods: </strong>We introduce a discrimination risk assessment framework for training datasets, a 4-step methodology integrating: (1) controlled representation sampling, (2) ensemble-based model training, (3) multilevel subgroup disparity quantification, and (4) mitigation-oriented interpretation. The framework systematically perturbs training subgroup composition while holding evaluation sets fixed to isolate representation effects. Validation was performed across 7 publicly available pediatric type 1 diabetes datasets using sociodemographic variables and continuous glucose monitoring data to assess robustness under heterogeneous data sources.</p><p><strong>Results: </strong>Representation balance alone does not guarantee stable or equitable model outputs. Ensemble analyses revealed subgroup-dependent volatility, with some groups consistently contributing to model generalization, while others inducing instability despite increased representation. These findings demonstrate structural sensitivity to data composition that is not captured by standard performance-only evaluations.</p><p><strong>Discussion: </strong>The framework exposes disparities and instability patterns that remain hidden under single-model or balanced-data evaluations. By quantifying representation-driven behaviors and subgroup-specific training value, it offers a diagnostic tool for understanding data-induced risks prior to model deployment.</p><p><strong>Conclusion: </strong>This work provides a methodology for assessing discrimination risks in training datasets used in medical informatics. By integrating data composition effects, prediction stability, and subgroup-level disparity analysis, the framework supports more reliable and transparent development of machine learning systems across diverse clinical and nonclinical contexts.</p><p><strong>Clinical trial registration: </strong>This research did not involve any new clinical trial. All analyses were performed on data from previously registered clinical trials, as listed below: <b>PEDAP dataset</b>: The source of the data is the PEDAP Trial Study Group. (2024). PEDAP Public Dataset (Release 4). Retrieved from https://public.jaeb.org/dataset/599. The analyses content and conclusions presented herein are solely the responsibility of the authors and have not been reviewed or approved by PEDAP Trial Study Group. <i>ClinicalTrials.gov</i> Identifier: NCT04796779; registered March 15, 2021. <b>IOBP2 RCT dataset</b>: The source of the data is the Insulin Only Bionic Pancreas Pivotal Trial. Retrieved from https://public.jaeb.org/dataset/579. The analyses content and conclusions presented herein are solely the re","PeriodicalId":36278,"journal":{"name":"JAMIA Open","volume":"9 4","pages":"ooag125"},"PeriodicalIF":4.7,"publicationDate":"2026-07-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13368804/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148457020","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Develop and validate a fair machine learning model to identify patients with high data-continuity in electronic health records data.","authors":"Yao An Lee, Tiange Tang, Yu Huang, Jiang Bian, Lizheng Shi, Jingchuan Guo","doi":"10.1093/jamiaopen/ooag124","DOIUrl":"10.1093/jamiaopen/ooag124","url":null,"abstract":"<p><strong>Objectives: </strong>Electronic health record (EHR) data discontinuity, defined as receiving care outside of a particular EHR system, may cause misclassification of study variables. We aimed to: (1) quantify misclassification across levels of EHR data discontinuity and identify an optimal continuity threshold, (2) develop a machine learning (ML) model to predict EHR continuity and optimize fairness across racial and ethnic groups, and (3) externally validate the EHR continuity prediction model using an independent dataset.</p><p><strong>Materials and methods: </strong>We used linked OneFlorida+ EHR-Medicaid claims data for model development and Research Action for Health Network (REACHnet) EHR-Louisiana Blue Cross Blue Shield (LABlue) claims data for external validation. A Harmonized Encounter Proportion Score (HEPS), adapted from prior continuity metrics, was applied to quantify patient-level EHR data continuity and the impact on misclassification of 42 clinical variables. Machine learning models were trained using routinely available demographic, clinical, and healthcare utilization features derived from structured EHR data.</p><p><strong>Results: </strong>Higher EHR data continuity was associated with lower rates of misclassification. A HEPS threshold of approximately 30% effectively distinguished patients with sufficient data continuity. Machine learning models demonstrated strong performance in predicting high continuity (area under the receiver operating characteristic curve [AUROC] = 0.77). Fairness assessments showed bias against Hispanic group, which was substantially improved following bias mitigation procedures. Model performance remained robust and fair in the external validation.</p><p><strong>Discussion: </strong>Our study offers a practical metric for quantifying data continuity in EHR networks. The current ML model incorporating EHR-routinely collected information can accurately identify patients with high care continuity.</p><p><strong>Conclusions: </strong>We developed a generalizable data-continuity classification tool that can be easily applied across EHR systems, strengthening the rigor of EHR-based research.</p>","PeriodicalId":36278,"journal":{"name":"JAMIA Open","volume":"9 4","pages":"ooag124"},"PeriodicalIF":4.7,"publicationDate":"2026-07-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13361708/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148438282","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Using a translational data platform to create clinical-grade genome-informed risk assessments.","authors":"Jennifer L Morse, Megan He, Hana Bangash, Wendy K Chung, Joshua Cortopassi, Sophie Forman, Adam S Gordon, Marwan Hamed, Paul A Harris, Elizabeth W Karlson, Eimear E Kenny, Krzysztof Kiryluk, Iftikhar J Kullo, Adam Lewis, Nita Limdi, Cong Liu, Priya Marathe, Kyle McGuffin, Jacqueline Odgis, Emma Perez, Jacob Petrzelka, Cynthia A Prows, Luke V Rasmussen, Ellis Thomas, Georgia L Wiesner, Travis Wilson, Jodell E Linder, Josh F Peterson","doi":"10.1093/jamiaopen/ooag118","DOIUrl":"10.1093/jamiaopen/ooag118","url":null,"abstract":"<p><strong>Objective: </strong>To describe the development and implementation of an automated platform for genomic risk prediction that integrates multiple data types.</p><p><strong>Materials and methods: </strong>Using the REDCap infrastructure, we constructed the R⁴ (Recruitment, Results, and Risk Reduction) platform to intake data from clinical sites, partner laboratories, participant surveys, and electronic health record (EHR) data across 13 institutions.</p><p><strong>Results: </strong>The R⁴ Portal successfully integrated data to generate genome-informed risk assessments (GIRAs) across 11 conditions for a 23 840 person cohort. Testing phases and quality control led to network-wide protocols ensuring consistency and accuracy.</p><p><strong>Discussion: </strong>As the science of estimating disease risk evolves, standardized and high-throughput methods of collecting and manipulating complex data are required. Platforms should be open-source, modular, and reusable, ensuring flexibility, security, and integration across healthcare environments.</p><p><strong>Conclusion: </strong>The electronic MEdical Records and GEnomics (eMERGE) network successfully generated and returned comprehensive risk profiles using logic and data specific to 11 conditions in a secure and semi-automated fashion employing a customized REDCap database.</p>","PeriodicalId":36278,"journal":{"name":"JAMIA Open","volume":"9 4","pages":"ooag118"},"PeriodicalIF":4.7,"publicationDate":"2026-07-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13361706/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148438268","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Real-time prediction of atrial fibrillation in intensive care unit: a meta-learning approach.","authors":"Mehran Moazeni, Sebastian Mildiner Moraga, Maaike Wösten, Paul Elbers, Patrick Thoral, Linda Wilhelmina Van Laake, Folkert Wouter Asselbergs, Emmeke Aarts","doi":"10.1093/jamiaopen/ooag110","DOIUrl":"10.1093/jamiaopen/ooag110","url":null,"abstract":"<p><strong>Objective: </strong>Atrial fibrillation (AF) is a common arrhythmia affecting a large fraction of patients in intensive care units (ICUs). Predicting patients at risk of AF in the ICU is a challenging task and is not commonly practiced but could have clinical implications considering that AF is a proxy for poorer outcomes. Our goal is to develop a real-time AF prediction model using ICU numerical data to improve alarm quality, helping clinicians to identify at-risk patients and intervene before AF onset, thereby potentially improving clinical outcomes.</p><p><strong>Methods: </strong>We employed AmsterdamUMCdb, Europe's first openly accessible ICU dataset. The dataset includes static features, including demographics, as well as dynamic bedside monitoring data, like blood pressure and respiratory rate, along with detailed records of medication and fluid administration. To enhance generalizability across diverse ICU patients, we trained a long short-term memory (LSTM) model combined with a Model-Agnostic Meta-Learning (MAML) approach. Model performance was tested on an imbalanced test set, reflective of the real-world ratio of AF to non-AF patients. Direct external validation was performed using the MIMIC-IV database to test generalizability across different clinical settings.</p><p><strong>Results: </strong>The LSTM-MAML model achieved an AUC of 0.92, accuracy of 0.89, and precision of 0.29 on the internal validation set. In external validation with MIMIC-IV, it showed near-equivalent performance with an AUC of 0.89, accuracy of 0.85, and precision of 0.18.</p><p><strong>Conclusions: </strong>The real-time prediction model demonstrated predictive value for AF and has potential for future clinical benefits. However, for clinical implementation, further refinement is necessary to improve its performance in identifying patients at risk for AF.</p>","PeriodicalId":36278,"journal":{"name":"JAMIA Open","volume":"9 4","pages":"ooag110"},"PeriodicalIF":4.7,"publicationDate":"2026-07-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13355586/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148425062","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Build fair machine learning models to predict adverse outcomes for heart failure patients with preserved ejection fraction and with reduced ejection fraction.","authors":"YaYun Yeh, YaoAn Lee, Yu Huang, Wenxi Huang, Stephen E Kimmel, Carl Yang, Zhe Jiang, Tingsong Xiao, Yi Guo, Jiang Bian, Jingchuan Guo","doi":"10.1093/jamiaopen/ooag136","DOIUrl":"10.1093/jamiaopen/ooag136","url":null,"abstract":"<p><strong>Objectives: </strong>Develop and validate subtype-specific, fairness-aware machine learning (ML) models that integrate clinical and social determinants of health (SDoH) information to predict 6-month readmission or mortality after hospitalization among patients with HFpEF or HFrEF, and to evaluate the incremental contribution of SDoH, model explainability, and subgroup error disparities across demographic groups.</p><p><strong>Materials and methods: </strong>We used University of Florida Health electronic health record (EHR) data (2016-2022) to identify adult heart failure (HF) hospitalizations and followed patients for 6 months for a composite outcome of readmission or mortality. Features included clinical characteristics, contextual SDoH (eg, neighborhood deprivation), and individual SDoH extracted from clinical notes via natural language processing (NLP). Logistic regression and XGBoost models were trained with random oversampling. Performance metrics included the C statistic, F1-score, and recall. Fairness was evaluated using false negative rate (FNR) parity across sex, race/ethnicity, and age band, and mitigation methods were applied (eg, Disparate Impact Remover, Adversarial Debiasing, and Calibrated Equalized Odds).</p><p><strong>Results: </strong>Adding SDoH improved the C statistic for logistic regression in HFpEF (0.603 vs 0.586) and HFrEF (0.641 vs 0.637). SHapley Additive exPlanations (SHAP) highlighted sodium, financial constraint level, and emergency department visit count in HFpEF, and utilization measures and financial constraint level in HFrEF. FNR ratios indicated race/ethnicity disparities; HFpEF FNRBlack/FNRWhite was 0.7834 (0.8728 after Disparate Impact Remover), and HFrEF FNRHispanic/FNRWhite was 1.2217 (0.9880 after Adversarial Debiasing).</p><p><strong>Discussion: </strong>SDoH integration and mitigation can modestly improve performance while reducing subgroup error disparities.</p><p><strong>Conclusion: </strong>Subtype-specific, fairness-aware ML models for HFpEF and HFrEF provided interpretable 6-month risk stratification and enabled subgroup fairness assessment. Integrating clinical and SDoH information added modest discrimination gains while strengthening interpretation and fairness assessment. These findings support further validation of explainable, equity-aware HF prediction models.</p>","PeriodicalId":36278,"journal":{"name":"JAMIA Open","volume":"9 4","pages":"ooag136"},"PeriodicalIF":4.7,"publicationDate":"2026-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13353219/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148425009","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Mapping oral-systemic health relationships: a data-driven analysis of co-occurrence patterns in biomedical literature.","authors":"Bhumi Patel, Janusz Wojtusiak","doi":"10.1093/jamiaopen/ooag130","DOIUrl":"10.1093/jamiaopen/ooag130","url":null,"abstract":"<p><strong>Objective: </strong>To systematically map co-occurrence patterns between systemic medical conditions and oral diseases in the biomedical literature.</p><p><strong>Materials and methods: </strong>We conducted a large-scale bibliometric analysis of PubMed-indexed literature from 1995 to 2025. Using automated Application Programming Interface queries, we retrieved publications jointly referencing 256 systemic conditions (defined using the Agency for Healthcare Research and Quality's Clinical Classifications Software) and 5 oral health conditions (periodontal disease, dental caries, tooth loss, oral ulcers, and oral cancers). Search strategies combined Medical Subject Headings and title/abstract terms. Co-occurrence data were normalized using max scaling and analyzed using hierarchical clustering, K-means clustering, and network visualizations. This study reports associations between conditions and not causal relationships.</p><p><strong>Results: </strong>The search identified 79 817 co-mention instances across 39 782 unique publications. Of 256 medical conditions, 237 co-occurred with at least one oral condition. Periodontal diseases and oral cancers accounted for the largest share of co-mentions (42 554 and 38 735, respectively). Hierarchical clustering of systemic conditions revealed eight clusters, including malignancies strongly co-occurring with oral cancers and immunity disorders co-occurring with periodontal diseases. Dental conditions grouped into three clusters: (1) caries, ulcers, and tooth loss; (2) oral cancers; and (3) periodontal diseases. Publication-level clustering identified 6 thematic domains dominated by oncology and inflammation research.</p><p><strong>Conclusion: </strong>This study provides a data-driven overview of oral-systemic health research across three decades, highlighting well-studied co-occurrence patterns, such as periodontal-diabetes, oral cancer-malignancies and less frequently represented areas involving caries, ulcers, and tooth loss, underscoring opportunities for interdisciplinary research and integrated care models.</p>","PeriodicalId":36278,"journal":{"name":"JAMIA Open","volume":"9 4","pages":"ooag130"},"PeriodicalIF":4.7,"publicationDate":"2026-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13354603/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148425090","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Development of a benchmarking dataset for symptom detection using large language models.","authors":"Joshua Davis, Brigitte N Durieux, Chloe Van Dongen, Kate Sciacca, Juan Manuel Gutierrez, Charlotta Lindvall","doi":"10.1093/jamiaopen/ooag134","DOIUrl":"10.1093/jamiaopen/ooag134","url":null,"abstract":"<p><strong>Objectives: </strong>To develop a pipeline for evaluating large language models (LLMs) on the task of capturing symptoms from clinical encounters.</p><p><strong>Materials and methods: </strong>We created a gold standard dataset of symptom annotations from simulated doctor-patient encounter excerpts (264 encounters; 16 symptoms; double-coded and adjudicated). Nine different LLMs from 4 vendors (OpenAI, Meta, DeepSeek, Moonshot AI) were used as examples to test our evaluation pipeline; outputs were assessed for correct structure and symptom information.</p><p><strong>Results: </strong>Of 3085 excerpts, 2087 (68%) contained symptoms. Pain, cough, and shortness of breath were most common; LLMs achieved F1 scores ranging 0.66-0.88 for these symptoms with minimal prompt engineering. Of tested models, GPT-4.1 demonstrated the best overall performance.</p><p><strong>Discussion: </strong>Our evaluation pipeline and benchmarking dataset are publicly available and applicable to various LLMs, including open-source models.</p><p><strong>Conclusion: </strong>This work supports the development and optimization of models that seek to improve patient symptom understanding.</p>","PeriodicalId":36278,"journal":{"name":"JAMIA Open","volume":"9 4","pages":"ooag134"},"PeriodicalIF":4.7,"publicationDate":"2026-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13354605/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148425069","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Sociotechnical safety risks in oncology: a framework-guided qualitative study.","authors":"Chantelle Recsky, Michelle B Tam, Arlyn Heywood, Ruth A Dillon, Charlene E Ronquillo, Sandra B Lauck, Leah K Lambert","doi":"10.1093/jamiaopen/ooag103","DOIUrl":"10.1093/jamiaopen/ooag103","url":null,"abstract":"<p><strong>Objective: </strong>To analyze oncology nurses' accounts of technology-related safety events and the sociotechnical factors contributing to patient-, workforce-, and system-level harms.</p><p><strong>Materials and methods: </strong>A framework-guided qualitative descriptive study was conducted with 28 oncology nurses from outpatient, chemotherapy, inpatient, and radiation settings at a large Canadian cancer center. Semistructured interviews and small focus groups elicited 98 technology-related safety events. Data were analyzed using a framework-guided qualitative descriptive approach informed by the WHO International Classification for Patient Safety, a sociotechnical model of health information technology, and a health data-related harm framework.</p><p><strong>Results: </strong>Reported adverse events most often involved clinical administration, documentation, and organizational processes. Contributing sociotechnical factors included workflow complexity, communication breakdowns, people-related challenges, and interface usability issues. Harms extended beyond patients and included system inefficiencies, delays in care, and adverse effects on workforce well-being such as frustration and moral distress.</p><p><strong>Discussion: </strong>Nurses described technology-related safety issues as recurring sociotechnical patterns arising from workflow, documentation, and organizational misalignments, rather than isolated technical failures. These patterns had interconnected impacts across patients, care delivery, and the healthcare workforce. Structured framework analysis of clinicians' narratives helped illuminate these relationships.</p><p><strong>Conclusion: </strong>Technology-related safety events in oncology reflect recurring sociotechnical misalignments, rather than isolated technical failures, and have consequences that affect patients, care processes, and the healthcare workforce.</p>","PeriodicalId":36278,"journal":{"name":"JAMIA Open","volume":"9 4","pages":"ooag103"},"PeriodicalIF":4.7,"publicationDate":"2026-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13353214/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148425095","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"From study design to executable code: automating target trial emulation with large language models.","authors":"Hanjae Kim, Minseong Kim, Seonji Kim, Seng Chan You","doi":"10.1093/jamiaopen/ooag131","DOIUrl":"10.1093/jamiaopen/ooag131","url":null,"abstract":"<p><strong>Objective: </strong>Implementing target trial emulation (TTE) studies as standardized, reproducible analytic workflows is technically demanding. We developed Text-guided Health-study Estimation and Specification Engine Using Strategus (THESEUS), which uses large language models (LLMs) to translate free-text study descriptions into structured analytic specifications and Strategus R scripts within the Observational Health Data Sciences and Informatics (OHDSI) ecosystem.</p><p><strong>Materials and methods: </strong>THESEUS executes 2 steps: an LLM maps study descriptions to a JavaScript Object Notation (JSON) schema, and validated specifications are converted into Strategus R scripts through rule-based logic. For standardization evaluation, we compared specifications generated by 8 LLMs using 15 OHDSI-based TTE studies and 15 non-OHDSI studies under primary-analysis and full-analyses settings.</p><p><strong>Results: </strong>Under the primary-analysis setting, overall standardization accuracy ranged from 0.93 to 0.97 across models in OHDSI studies and from 0.82 to 0.95 in non-OHDSI studies. Gemini-3.1-Pro achieved the highest overall accuracy in OHDSI studies, while Gemini-3.1-Pro and Gpt-5.5 jointly achieved the highest overall accuracy in non-OHDSI studies. Under the full-analyses setting, field-level sensitivity ranged from 0.83 to 0.97 in OHDSI studies, with 0.07-0.80 false positives (FPs) per study, and from 0.77 to 0.89 in non-OHDSI studies, with 0.53-1.20 FPs per study. Gpt-5.5 performed best at the field level. THESEUS was implemented as a web application and coding-agent tools.</p><p><strong>Discussion: </strong>Pairing a standardized data model with a structured analysis framework enables reliable LLM-assisted interpretation of study descriptions and deterministic workflow construction in observational research.</p><p><strong>Conclusion: </strong>THESEUS supports translation of natural language study descriptions into executable, shareable code in standardized observational research settings.</p>","PeriodicalId":36278,"journal":{"name":"JAMIA Open","volume":"9 4","pages":"ooag131"},"PeriodicalIF":4.7,"publicationDate":"2026-07-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13348706/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148425043","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Computational phenotyping of sexually transmitted infections with the All of Us Research Program from 2010 to 2023.","authors":"Fanghui Shi, Huiyi Xia, Sharon Weissman, Xiaoming Li, Xueying Yang","doi":"10.1093/jamiaopen/ooag127","DOIUrl":"10.1093/jamiaopen/ooag127","url":null,"abstract":"<p><strong>Objectives: </strong>This study aims to develop comprehensive computable phenotyping algorithms that integrate multiple domains of electronic health record (EHR) data within the All of Us (AoU) Researcher Workbench to identify sexually transmitted infection (STI) cases and characterize STI patterns.</p><p><strong>Materials and methods: </strong>We used AoU controlled tier data, version 8, which included participants enrolled from May 6, 2018, to October 1, 2023. Using data across multiple domains of EHRs in AoU, such as diagnostic codes, laboratory results, and medication records, we developed computational phenotyping algorithms to identify 3 leading STIs: chlamydia, gonorrhea, and syphilis.</p><p><strong>Results: </strong>Among 393 596 participants with EHR data in AoU, 2603 had confirmed chlamydia, 1520 had gonorrhea, and 2762 had syphilis. By utilizing information from diagnostic codes and medication records, we identified an additional 4843 individuals with possible/presumed chlamydia and 18 855 individuals with possible/presumed gonorrhea. Among 5784 participants with at least one confirmed STI, 115 (2.0%) had all 3 infections. A notable shift in STI trends occurred in 2019, with the occurrence increasing from 2010 to 2019 before declining through 2023 across different STIs.</p><p><strong>Discussion: </strong>The comprehensive computable phenotype algorithms developed in this study provide an enhanced approach to identifying STI cases that may be missed using traditional lab-based methods.</p><p><strong>Conclusion: </strong>The computable phenotypes developed in this study provide a practical framework for more nuanced STI classification and downstream analysis in All of Us. Our approach may also support clinical and public health efforts by improving case identification beyond laboratory-only methods and informing interventions addressing clinical and social needs.</p>","PeriodicalId":36278,"journal":{"name":"JAMIA Open","volume":"9 4","pages":"ooag127"},"PeriodicalIF":4.7,"publicationDate":"2026-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13344838/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148413326","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}