Journal Of The American Medical Informatics Association[JOURNAL]
Sun
200 papers
Artificial intelligence and computer-aided diagnosis in diagnostic decisions: 5 questions for medical informatics and human-computer interface research.
Brunyé TT, Mitroff SR, Elmore JG
J Am Med Inform Assoc · 2026 Feb · PMID 41101774 · Full text
OBJECTIVES: Artificial intelligence (AI) has the potential to transform medical informatics by supporting clinical decision-making, reducing diagnostic errors, and improving workflows and efficiency. However, successful...
OBJECTIVES: Artificial intelligence (AI) has the potential to transform medical informatics by supporting clinical decision-making, reducing diagnostic errors, and improving workflows and efficiency. However, successful integration of AI-based decision support systems depends on careful consideration of human-AI collaboration, trust, skill maintenance, and automation bias. This work proposes five central questions to guide future research in medical informatics and human-computer interface (HCI). MATERIALS AND METHODS: We focus on AI-based clinical decision support systems, including computer vision algorithms for medical imaging (radiology, pathology), natural language processing for structured and unstructured electronic health record (EHR) data, and rule-based systems. Relevant data modalities include clinician-acquired images, EHR text, and increasingly, patient-generated content in telehealth contexts. We review existing evidence regarding diagnostic errors across specialties, the effectiveness and risks of AI tools in reducing perceptual and interpretive errors, and the human factors influencing diagnostic decision-making in AI-enabled contexts. We synthesize insights from medicine, cognitive science, and HCI to identify gaps in knowledge and propose five key questions for continued research. RESULTS: Diagnostic errors remain common across medicine, with AI offering potential to reduce both perceptual and interpretive errors. However, the impact of AI depends critically on how and when information is presented. Studies indicate that delayed or toggleable cues may outperform immediate ones, but attentional capture, overreliance, and bias remain significant risks. Explainable AI provides transparency but can also bias decisions. Long-term reliance on AI may erode clinician skills, particularly for trainees and in low-prevalence contexts. Historical failures of computer-aided diagnosis in mammography highlight these challenges. DISCUSSION AND CONCLUSION: Effective AI integration requires human-centered and adaptive design. Five central research questions address: (1) what type and format of information AI should provide; (2) when information should be presented; (3) how explainable AI affects diagnostic decisions; (4) how AI influences automation bias and complacency; and (5) the risks of skill decay due to reliance on AI. Each question underscores the importance of balancing efficiency, accuracy, and clinician expertise while mitigating bias and skill degradation. AI holds promise for improving diagnostic accuracy and efficiency, but realizing its potential requires post-deployment evaluation, equitable access, clinician oversight, and targeted training. AI must complement, rather than replace, human expertise, ensuring safe, effective, and sustainable integration into diagnostic decision-making. Addressing these challenges proactively can maximize AI's potential across healthcare and other high-stakes domains.
Evaluating ambient artificial intelligence documentation: effects on work efficiency, documentation burden, and patient-centered care.
Guo Y, Wang J, Hu D … +6 more , Tam S, Gilman C, Chow E, Perret D, Pandita D, Zheng K
J Am Med Inform Assoc · 2026 Feb · PMID 41100159 · Full text
BACKGROUND AND SIGNIFICANCE: Ambient listening tools powered by generative artificial intelligence (GenAI) offer real-time, scribe-like support that reduce documentation burden and may help alleviate burnout. This study...
BACKGROUND AND SIGNIFICANCE: Ambient listening tools powered by generative artificial intelligence (GenAI) offer real-time, scribe-like support that reduce documentation burden and may help alleviate burnout. This study assesses physician-perceived benefits and challenges of ambient AI implementation through surveys and evaluates its effectiveness in clinical workflows using automatically recorded electronic health record (EHR) time-efficiency metrics. METHOD AND MATERIALS: A quality improvement pilot has been underway at UCI Health since December 2023. Epic EHR Signal metrics were analyzed to assess changes in note length, documentation time, and same-day encounter closure rates. Matched pre- and post-implementation surveys evaluated physician-perceived changes in documentation burden, clinical efficiency, and care quality. We also examined open-ended survey responses using thematic analysis to supplement quantitative findings. RESULTS: Analysis on EHR usage data from 167 physicians showed significant reductions in note-writing time, despite an increase in note length. Survey responses (n = 65) also indicated statistically significant improvements across multiple domains. Physicians reported reduced cognitive demand (P = .031) and documentation effort (P = .014), alongside perceptions of enhanced clinical efficiency, patient-centered care, and EHR system usability. Thematic analysis confirmed these quantitative findings and identified opportunities for improvement, including specialty-specific customization and expanded AI functionality. DISCUSSION: Ambient AI tools demonstrated improved documentation efficiency, perceived care quality, and reduced cognitive workload. These benefits suggest potential to alleviate key burdens in clinical documentation. CONCLUSION: Future development should prioritize customization for specialty-specific and individual physician needs, ensure the reliability and accuracy of AI-generated content, and integrate ethical and legal considerations to facilitate safe and scalable implementation in patient-centered care contexts.
The real-world impact of artificial intelligence ethics frameworks across a decade in healthcare: a scoping review.
Chan A, Rahimi-Ardabilli H, Rogers WA … +1 more , Coiera E
J Am Med Inform Assoc · 2025 Nov · PMID 41093301 · Full text
OBJECTIVES: The number of ethical frameworks designed to guide artificial intelligence (AI) use has grown substantially over the past decade, yet their real-world effect remains unclear. We aimed to synthesize existing e...
OBJECTIVES: The number of ethical frameworks designed to guide artificial intelligence (AI) use has grown substantially over the past decade, yet their real-world effect remains unclear. We aimed to synthesize existing evidence to analyze the practical impact of AI ethics frameworks (AIEFs) operationalized in healthcare. MATERIALS AND METHODS: We conducted a scoping review across 4 academic databases (Ovid MEDLINE, Ovid Embase, Scopus, and Web of Science), Google, and Google Scholar from January 2014 to January 2025. Eligible studies reported primary research on the qualitative or quantitative impacts of AIEFs implemented in healthcare. Data synthesis was conducted via narrative review. RESULTS: Of 1807 records identified, 16 studies met inclusion criteria. These comprised 5 preliminary initiatives testing guidelines in practice, 5 case studies, 5 implementation studies, and a comparative case study. AIEFs were implemented: (1) to develop new AI governance structures and guidelines, (2) as ethical review assessment systems for adopting clinical AI technologies, and (3) as ethical "audit" tools for identifying ethical risks. Impact was reported through qualitative improvements to process measures such as improved trust in AI. No studies demonstrated a direct link between AIEFs and health-related outcome measures such as patient safety. DISCUSSION: AIEFs led to changes in organizational or clinical processes, including increased compliance with ethical standards. When embedded in governance, AIEFs improved oversight and evaluation, but audits were constrained by their reliance on organizational cooperation. CONCLUSION: Despite the proliferation of AIEFs over the past decade, their implementation in healthcare remains limited and impact on health outcomes unmeasured or underreported.
Derivation and validation of an algorithm for maternal-child linkage in electronic health records.
Rogerson CM, Bartlett CW, Price J … +3 more , Li L, Mendonca EA, Grannis S
J Am Med Inform Assoc · 2026 Feb · PMID 41093300 · Full text
INTRODUCTION: We created a probabilistic maternal-child electronic health record (EHR) linkage algorithm to promote clinical research in maternal-child health. METHODS: We used EHR data from 1994 to 2024 to create an XGB...
INTRODUCTION: We created a probabilistic maternal-child electronic health record (EHR) linkage algorithm to promote clinical research in maternal-child health. METHODS: We used EHR data from 1994 to 2024 to create an XGBoost model to predict maternal-child linkages. The model used standard EHR elements as predictor variables, including first name, last name, birthdate, address, phone number, email, and an EHR-embedded maternal-child indicator as the deterministic outcome. RESULTS: From 82 million unique records, 6.2 billion potential pairs met blocking criteria. Of the potential pairs, 33 364 674 contained the deterministic indicator and were used as cases, and an equal number of controls were randomly sampled. The final model obtained an accuracy of 92%, a precision of 98%, a recall of 87%, and an F1-score of 92%. CONCLUSION: We derived and validated a probabilistic maternal-child linkage algorithm using routinely collected EHR data elements that could benefit future observational research in maternal-child health.
Using natural language processing to extract information from clinical text in electronic medical records for populating clinical registries: a systematic review.
Liu L, Blake V, Barman M … +6 more , Gallego B, Churches T, Kennedy G, Ooi SY, Delaney GP, Jorm L
J Am Med Inform Assoc · 2026 Feb · PMID 41093296 · Full text
OBJECTIVE: Clinical registries advance healthcare by tracking patient outcomes and intervention safety. Manually extracting information from clinical text for registries is labor- and resource-intensive and often inaccur...
OBJECTIVE: Clinical registries advance healthcare by tracking patient outcomes and intervention safety. Manually extracting information from clinical text for registries is labor- and resource-intensive and often inaccurate. Therefore, this systematic review aims to evaluate the use and effectiveness of natural language processing (NLP) methods in extracting information from clinical text for populating clinical registries. MATERIALS AND METHODS: PubMed, Embase, Scopus, Web of Science, and ACM Digital Library were systematically searched. Studies were included if they used NLP techniques to populate clinical registries. The extracted data included details of the registry, the clinical text, the registry data elements extracted, the NLP methods used, and how their performance was evaluated. RESULTS: Fifteen articles were included in the review. Since 2020, the use of NLP methods for extracting information to populate clinical registries has been increasing steadily. Initially, rule-based NLP methods dominated the field, but machine learning-based approaches have gradually gained popularity. However, only one of the included studies employed generative large language models (LLMs). The diversity of clinical text and extracted data elements posed challenges to the generalizability of the NLP methods. CONCLUSION: To date, the application of NLP methods to clinical text for populating clinical registries has been limited in both the number of published studies and the scope of implementation. The NLP methods used thus far face significant challenges in effectively managing the complexity and diversity of clinical text and data elements. Moreover, the performance of the NLP methods varied significantly. This review underscores the need for a robust and adaptable NLP framework. Generative LLMs may provide direction for future research, but their use must account for challenges such as accuracy, cost, privacy, and limited supporting evidence.
FHIR-Former: enhancing clinical predictions through Fast Healthcare Interoperability Resources and large language models.
Engelke M, Baldini G, Kleesiek J … +2 more , Nensa F, Dada A
J Am Med Inform Assoc · 2025 Dec · PMID 41082356 · Full text
OBJECTIVE: To address the challenges of data heterogeneity and manual feature engineering in clinical predictive modeling, we introduce FHIR-Former, an open-source framework integrating Fast Healthcare Interoperability R...
OBJECTIVE: To address the challenges of data heterogeneity and manual feature engineering in clinical predictive modeling, we introduce FHIR-Former, an open-source framework integrating Fast Healthcare Interoperability Resources (FHIR) with large language models (LLMs) to automate and standardize clinical prediction tasks. MATERIALS AND METHODS: FHIR-Former dynamically processes structured (eg, lab results, medications) and unstructured (eg, clinical notes) data from FHIR resources. The pipeline supports multiple classification tasks, including 30-day readmission, imaging study prediction, and ICD code classification. Leveraging open-source LLMs (GeBERTa), we trained models on 1.1 million data points across ten FHIR resources using retrospective inpatient data (2018-2024). Hyperparameters were optimized via Bayesian methods, and outputs were mapped to FHIR RiskAssessment resources for interoperability. RESULTS: FHIR-Former achieved an F1-score of 70.7% and accuracy of 72.9% for 30-day readmission, 51.8% F1-score (88.1% accuracy) for mortality prediction, and 61% macro F1-score for imaging study classification. The ICD code prediction model attained 94% accuracy. Performance demonstrated promising performance for readmission and showed scalability across tasks without manual feature engineering. DISCUSSION: FHIR-Former eliminates institution-specific preprocessing by adapting to diverse FHIR implementations, enabling seamless integration of multimodal data. Its configurable architecture outperformed prior frameworks reliant on static inputs or limited to unstructured text. Real-time risk scores embedded in FHIR servers enhance clinical workflows without disrupting existing practices. CONCLUSION: By harmonizing FHIR standardization with LLM flexibility, FHIR-Former advances scalable, interoperable predictive modeling in healthcare. The open-source framework facilitates automation, improves resource allocation, and supports personalized decision-making, bridging gaps between AI innovation and clinical practice.
Efficient extraction of medication information from clinical notes: an evaluation in 2 languages.
Fabacher T, Sauleau EA, Arcay E … +6 more , Faye B, Alter M, Chahard A, Miraillet N, Coulet A, Névéol A
J Am Med Inform Assoc · 2025 Dec · PMID 41071911 · Full text
OBJECTIVE: To evaluate the accuracy, computational cost, and portability of a new natural language processing (NLP) method for extracting medication information from clinical narratives. MATERIALS AND METHODS: We propose...
OBJECTIVE: To evaluate the accuracy, computational cost, and portability of a new natural language processing (NLP) method for extracting medication information from clinical narratives. MATERIALS AND METHODS: We propose an original transformer-based architecture for the extraction of entities and their relations pertaining to patients' medication regimen. First, we used this approach to train and evaluate a model on French clinical notes, using a newly annotated corpus from Hôpitaux Universitaires de Strasbourg. Second, the portability of the approach was assessed by conducting an evaluation on clinical documents in English from the 2018 n2c2 shared task. Information extraction accuracy and computational cost were assessed by comparison with an available method using transformers. RESULTS: The proposed architecture achieves on the task of relation extraction itself performance that are competitive with the state-of-the-art on both French and English (F-measures 0.82 and 0.96 vs 0.81 and 0.95), but reduces the computational cost by 10. End-to-end (Named Entity recognition and Relation Extraction) F1 performance is 0.69 and 0.82 for French and English corpus. DISCUSSION: While an existing system developed for English notes was deployed in a French hospital setting with reasonable effort, we found that an alternative architecture offered end-to-end drug information extraction with comparable extraction performance and lower computational impact for both French and English clinical text processing, respectively. CONCLUSION: The proposed architecture can be used to extract medication information from clinical text with high performance and low computational cost and consequently suits with usually limited hospital IT resources.
Should we synthesize more than we need: impact of synthetic data generation for high-dimensional cross-sectional medical data.
Pilgram L, El Kababji S, Liu D … +1 more , El Emam K
J Am Med Inform Assoc · 2025 Dec · PMID 41069048 · Full text
OBJECTIVE: In medical research and education, generative artificial intelligence/machine learning (AI/ML) models to synthesize artificial medical data can enable the sharing of high-quality data while preserving the priv...
OBJECTIVE: In medical research and education, generative artificial intelligence/machine learning (AI/ML) models to synthesize artificial medical data can enable the sharing of high-quality data while preserving the privacy of patients. Given that such data is often high-dimensional, a relevant consideration is whether to synthesize the entire dataset when only a task-relevant subset is needed. This study evaluates how the number of variables in training impacts fidelity, utility, and privacy of the synthetic data (SD). MATERIAL AND METHODS: We used 12 cross-sectional medical datasets, defined a downstream task with corresponding core variables, and derived 6354 variants by adding adjunct variables to the core. SD was generated using 7 different generative models and evaluated for fidelity, downstream utility, and privacy. Mixed-effect models were used to assess the effect of adjunct variables on the respective evaluation metric, accounting for the medical dataset as a random component. RESULTS: Fidelity was unaffected by the number of adjunct variables in 5/7 SDG models. Similarly, downstream utility remained stable in 6/7 (predictive task) and 5/7 (inferential task) SDG models. Where significant effects were observed, they were minimal, resulting, for example, in a 0.05 decrease in Area under the Receiver Operating Characteristic curve (AUROC) when adding 120 variables. Privacy was not impacted by the number of adjunct variables. DISCUSSION: Our findings show that fidelity, utility, and privacy are preserved when generating a more comprehensive medical dataset than the task-relevant subset. CONCLUSION: Our findings support a cost-effective, utility, and privacy-preserving way of implementing SDG into medical research and education.
A SNAPpy use of large language models: using large language models to classify treatment plans in pediatric acute otitis media.
Pourian JJ, Michaels B, Vo A … +3 more , Holmgren AJ, Garcia-Agundez A, Flaherman V
J Am Med Inform Assoc · 2025 Dec · PMID 41069025 · Full text
BACKGROUND AND SIGNIFICANCE: Acute otitis media (AOM) is a leading cause of pediatric antibiotic overuse. Safety Net Antibiotic Prescriptions (SNAPs) are recommended for antibiotic stewardship but are difficult to identi...
BACKGROUND AND SIGNIFICANCE: Acute otitis media (AOM) is a leading cause of pediatric antibiotic overuse. Safety Net Antibiotic Prescriptions (SNAPs) are recommended for antibiotic stewardship but are difficult to identify due to lack of structured documentation. OBJECTIVE: This study validates the accuracy of Versa, a GPT-4o based HIPAA-compliant large language model (LLM), to classify AOM treatment plans from physician notes. METHODS: A retrospective cross-sectional study analyzed pediatric AOM encounters. Multiple prompting strategies were used to classify treatment plans and validated against a representative sample of manual reviews by 2 pediatricians. A locally fine-tuned model, Clinical-Longformer was also trained and tested against Versa and human review. RESULTS: In total, 5707 encounters were included; 374 reviewed manually. Zero-shot accuracy was 97.8%; few-shot accuracy was 85%. Clinical-Longformer achieved 93.3% accuracy. CONCLUSION: Versa effectively identifies AOM treatment plans, providing a cost-efficient quality improvement tracking tool for prescription practice patterns in pediatric antibiotic stewardship efforts.
Zero-shot medical event prediction using a generative pretrained transformer on electronic health records.
Redekop E, Wang Z, Kulkarni R … +7 more , Pleasure M, Chin A, Hassanzadeh HR, Hill BL, Emami M, Speier WF, Arnold CW
J Am Med Inform Assoc · 2025 Dec · PMID 41060255 · Full text
OBJECTIVES: Longitudinal data in electronic health records (EHRs) represent an individual's clinical history through a sequence of codified concepts, including diagnoses, procedures, medications, and laboratory tests. Ge...
OBJECTIVES: Longitudinal data in electronic health records (EHRs) represent an individual's clinical history through a sequence of codified concepts, including diagnoses, procedures, medications, and laboratory tests. Generative pretrained transformers (GPT) can leverage this data to predict future events. While fine-tuning of these models can enhance task-specific performance, it becomes costly when applied to many clinical prediction tasks. In contrast, a pretrained foundation model can be used in zero-shot forecasting setting, offering a scalable alternative to fine-tuning separate models for each outcome. MATERIALS AND METHODS: This study presents the first comprehensive analysis of zero-shot forecasting with GPT-based foundational models in EHRs, introducing a novel pipeline that formulates medical concept prediction as a generative modeling task. Unlike supervised approaches requiring extensive labeled data, our method enables the model to forecast the next medical event purely from a pretraining knowledge. We evaluate performance across multiple time horizons and clinical categories, demonstrating model's ability to capture latent temporal dependencies and complex patient trajectories without task supervision. RESULTS: The model's performance in predicting the next medical concept was evaluated using precision and recall metrics, achieving an average top-1 precision of 0.614 and recall of 0.524. For 12 major diagnostic conditions, the model demonstrated strong zero-shot performance, achieving high true positive rates while maintaining low false positives. DISCUSSION: We demonstrate the power of a foundational EHR GPT model in capturing diverse phenotypes and enabling robust, zero-shot forecasting of clinical outcomes. This capability highlights both its versatility across conditions like liver cancer and SLE, and its limitations in more ambiguous settings such as depression, while also revealing meaningful latent clinical structure. CONCLUSION: This capability enhances the versatility of predictive healthcare models and reduces the need for task-specific training, enabling more scalable applications in clinical settings.
What patients want from healthcare chatbots: insights from a mixed-methods study.
Dellavalle NS, Ellis JR, Moore AA … +4 more , Akerson M, Andazola M, Campbell EG, DeCamp M
J Am Med Inform Assoc · 2025 Nov · PMID 41051963 · Full text
OBJECTIVES: To understand whether patients prefer chatbots for certain tasks in healthcare, and their motivations for doing so, recognizing that chatbots are already assisting patients with various healthcare tasks. MATE...
OBJECTIVES: To understand whether patients prefer chatbots for certain tasks in healthcare, and their motivations for doing so, recognizing that chatbots are already assisting patients with various healthcare tasks. MATERIALS AND METHODS: We conducted a mixed-methods study with patient-users of a healthcare system multi-task chatbot integrated in an electronic health record. We purposively oversampled by race or ethnicity to survey 617/3089 (response rate, 20.0%) chatbot users using de novo and validated survey items. We conducted semi-structured interviews with 46 patient-users and 2 chatbot developers between November 2022 and May 2024. We used modified grounded theory to analyze interviews, descriptive statistics and Chi-square tests to compare survey results, and mixed-methods techniques to integrate findings. RESULTS: Patient-users preferred chatbots for administrative tasks to save providers' time, because of the chatbot availability, and to avoid unpleasant interactions. Some preferred to discuss sensitive tasks (such as mental health or gender-affirming care) with chatbots due to more privacy or anonymity and less embarrassment or judgment. Developer interviews corroborated this finding. Avoiding bias and using a preferred means of communication applied to all tasks. In surveys, patient-users were less likely to worry about being judged based on chatbot interactions (153/608, 25.2%) compared to interactions with a doctor (219/606, 36.1%) (P < .001). Patient-users preferred human clinicians for diagnostic tasks. DISCUSSION: Patient-users appear to simultaneously prefer chatbots for simple tasks or sensitive ones, with diverse motivations. Whether chatbots best meet patient needs while balancing ethical tensions regarding access, privacy, judgment, and bias is unclear. CONCLUSION: Future chatbot design must accommodate different and diverse patient preferences.
A standards-based approach to digital health research: implementing the people heart study.
Sayeed R, Kreda D, Mandel JC … +4 more , Larson B, Gordon W, Mandl KD, Kohane I
J Am Med Inform Assoc · 2025 Dec · PMID 41032388 · Full text
OBJECTIVE: To assess whether HL7 Fast Healthcare Interoperability Resources (FHIR) can underpin a fully standards-based, end-to-end digital research architecture, demonstrate it in a live study, and quantify its benefits...
OBJECTIVE: To assess whether HL7 Fast Healthcare Interoperability Resources (FHIR) can underpin a fully standards-based, end-to-end digital research architecture, demonstrate it in a live study, and quantify its benefits for interoperability and development efficiency. MATERIALS AND METHODS: We designed a generalizable standards-based architecture to accelerate digital health research relying on FHIR as the sole transactional model throughout a participant research lifecycle starting from API-based study discovery to results. It was instantiated for People Heart Study, a real-world digital health cardiovascular-risk assessment study with its protocol transformed into FHIR resources (eligibility, consent, tasks, and results). Evaluation examined workflow coverage, validator conformance across independent servers, and points requiring custom extensions or app logic. RESULTS: The architecture was implemented using cloud managed FHIR stores including an illustrative public research discovery API for first-/third-party apps. A participant-facing iOS app was published on the App Store. Our evaluation reveals that 6 of 10 research app workflows could be executed entirely from FHIR artifacts; 2 were partially standards-driven and 2 remained limited requiring custom development. All FHIR resources passed structural, semantic validation with minimal custom extension usage and terminology integrity issues. DISCUSSION: Our approach addresses persistent challenges in digital health research by enhancing data interoperability, minimizing redundant development, and supporting the full research lifecycle. The architecture aligns with national priorities and complements healthcare standardization efforts. CONCLUSION: By leveraging FHIR, our architecture enables generalizability, interoperability, and reuse across diverse digital health research contexts, transforming study design into data modeling rather than software development, and fostering a more inclusive and agile digital health ecosystem.
Supporting public transit research in healthcare settings: testing a free, fast, and secure method for routing public transit from patient address to the point of care.
Aktay SL, Aktay OA, Menon S … +2 more , Huang SJ, McCoy RG
J Am Med Inform Assoc · 2025 Dec · PMID 41032045 · Full text
OBJECTIVES: Gaps in transportation, particularly public transit, are a significant barrier to accessible, high-quality healthcare. Health systems, payors, and regulatory bodies recognize the need to identify and address...
OBJECTIVES: Gaps in transportation, particularly public transit, are a significant barrier to accessible, high-quality healthcare. Health systems, payors, and regulatory bodies recognize the need to identify and address these gaps. However, clinical research examining public transportation accessibility and its impacts on healthcare utilization, outcomes, and costs remains limited. Existing tools used for studying public transit are generally non-HIPAA compliant, expensive, proprietary, and/or difficult to use. A tool addressing these concerns is needed to enable the incorporation of transportation variables into research and clinical care settings. MATERIALS AND METHODS: We developed and implemented a novel framework for building a public transit routing system that is comprised of free, publicly available data and offline software to maintain HIPAA compliance. The system consists of a transit router and a geocoder for converting addresses into coordinates. RESULTS: A total of 463 879 out of 505 379 (∼91.8%) of Baltimore, Maryland, addresses were successfully routed to University of Maryland Medical Center in 24 hours of compute time. A significant portion of journeys consisted of walking (36% of median trip time) or using a transit vehicle (57.2%). Testing the router with varying random-access memory levels showed a plateau in routing speed between 12 and 20 GB. The geocoding approach is >90% consistent with a widely used but non-HIPAA compliant geocoder. DISCUSSION: The methodology and step-by-step guidance shared in this study can allow researchers, public health professionals, non-for-profit agencies, and other stakeholders to efficiently, effectively, and safely incorporate public transportation information into their work. CONCLUSION: Public transportation routing using freely available data and software is possible in a HIPAA-compliant manner.
Machine learning based prediction of medication adherence in heart failure using large electronic health record cohort with linkages to pharmacy-fill and neighborhood-level data.
Adhikari S, Stokes T, Li X … +12 more , Zhao Y, Fitchett C, Ladino N, Lawrence S, Qian M, Cho YS, Hamo C, Dodson JA, Chunara R, Kronish IM, Mukhopadhyay A, Blecker SB
J Am Med Inform Assoc · 2025 Dec · PMID 41032036 · Full text
OBJECTIVE: While timely interventions can improve medication adherence, it is challenging to identify which patients are at risk of nonadherence at point-of-care. We aim to develop and validate flexible machine learning...
OBJECTIVE: While timely interventions can improve medication adherence, it is challenging to identify which patients are at risk of nonadherence at point-of-care. We aim to develop and validate flexible machine learning (ML) models to predict a continuous measure of adherence to guideline-directed medication therapies (GDMTs) for heart failure (HF). MATERIALS AND METHODS: We utilized a large electronic health record (EHR) cohort of 34,697 HF patients seen at NYU Langone Health with an active prescription for ≥1 GDMT between April 01, 2021 and October 31, 2022. The outcome was adherence to GDMT measured as proportion of days covered (PDC) at 6 months following a clinical encounter. Over 120 predictors included patient-, therapy-, healthcare-, and neighborhood-level factors guided by the World Health Organization's model of barriers to adherence. We compared performance of several ML models and their ensemble (superlearner) for predicting PDC with traditional regression model (OLS) using mean absolute error (MAE) averaged across 10-fold cross-validation, % increase in MAE relative to superlearner, and predictive-difference across deciles of predicted PDC. RESULTS: Superlearner, a flexible nonparametric prediction approach, demonstrated superior prediction performance. Superlearner and quantile random forest had the lowest MAE (mean [95% CI] = 18.9% [18.7%-19.1%] for both), followed by MAEs for quantile neural network (19.5% [19.3%-19.7%]) and kernel support vector regression (19.8% [19.6%-20.0%]). Gradient boosted trees and OLS were the 2 worst performing models with 17% and 14% higher MAEs, respectively, relative to superlearner. Superlearner demonstrated improved predictive difference. CONCLUSION: This development phase study suggests potential of linked EHR-pharmacy data and ML to identify HF patients who will benefit from medication adherence interventions. DISCUSSION: Fairness evaluation and external validation are needed prior to clinical integration.
Enhancing Structured Team Communication in Acute Care Settings with Ambient AI Scribes.
Jalilian L, Lukac P, Lane-Fall M
J Am Med Inform Assoc · 2025 Nov · PMID 41003645 · Full text
OBJECTIVE: This perspective explores how ambient artificial intelligence (AI) scribes could support documentation and quality improvement (QI) of structured, team-based provider-to-provider communication in acute care se...
OBJECTIVE: This perspective explores how ambient artificial intelligence (AI) scribes could support documentation and quality improvement (QI) of structured, team-based provider-to-provider communication in acute care settings. BACKGROUND: In acute care settings, team-based discussions such as multidisciplinary rounds and handoffs are essential to the delivery of safe care. These discussions rely on standardized frameworks (eg, IPASS, checklists) to ensure consistent information transfer and shared understanding. Despite their importance, these verbal discussions are often incompletely documented or left undocumented in the electronic health record, leading to gaps in clinical narrative, difficulty in QI evaluation, and lost opportunities for organizational learning. APPROACH: We outline how ambient AI scribes could enhance documentation of team-based communication in daily rounding and handoff discussions. We examine key sociotechnical challenges, including workflow integration, multiprovider consent, surveillance concerns, and vendor collaboration. We describe our experience with proof-of-concept demonstrations as an early feasibility signal. RESULTS: Ambient AI scribes are a promising tool for capturing structured team communication. Their use should be explored for its potential to improve documentation, support clinician well-being, and enable data-driven approaches to QI and communication fidelity assessments. Effective implementation requires workflow adaptations incorporating scribe output verification, transparent governance, and trust-building efforts to ensure clinician acceptance. DISCUSSION: Ambient AI scribes represent a novel frontier in documentation of structured team discussions in acute care settings, with the potential to strengthen communication reliability and systems learning of these vital conversations. Future research should evaluate their impact on patient safety, workforce well-being, and patient outcomes in acute care settings.
Towards responsible artificial intelligence in healthcare-getting real about real-world data and evidence.
Koski E, Das A, Hsueh PS … +22 more , Solomonides A, Joseph AL, Srivastava G, Johnson CE, Kannry J, Oladimeji B, Price A, Labkoff S, Bharathy G, Lin B, Fridsma D, Fleisher LA, Lopez-Gonzalez M, Singh R, Weiner MG, Stolper R, Baris R, Sincavage S, Naumann T, Williams T, Bui TTT, Quintana Y
J Am Med Inform Assoc · 2025 Nov · PMID 40999782 · Full text
BACKGROUND: The use of real-world data (RWD) in artificial intelligence (AI) applications for healthcare offers unique opportunities but also poses complex challenges related to interpretability, transparency, safety, ef...
BACKGROUND: The use of real-world data (RWD) in artificial intelligence (AI) applications for healthcare offers unique opportunities but also poses complex challenges related to interpretability, transparency, safety, efficacy, bias, equity, privacy, ethics, accountability, and stakeholder engagement. METHODS: A multi-stakeholder expert panel comprising healthcare professionals, AI developers, policymakers, and other stakeholders was assembled. Their task was to identify critical issues and formulate consensus recommendations, focusing on the responsible use of RWD in healthcare AI. The panel's work involved an in-person conference and workshop and extensive deliberations over several months. RESULTS: The panel's findings revealed several critical challenges, including the necessity for data literacy and documentation, the identification and mitigation of bias, privacy and ethics considerations, and the absence of an accountability structure for stakeholder management. To address these, the panel proposed a series of recommendations, such as the adoption of metadata standards for RWD sources, the development of transparency frameworks and instructional labels likened to "nutrition labels" for AI applications, the provision of cross-disciplinary training materials, the implementation of bias detection and mitigation strategies, and the establishment of ongoing monitoring and update processes. CONCLUSION: Guidelines and resources focused on the responsible use of RWD in healthcare AI are essential for developing safe, effective, equitable, and trustworthy applications. The proposed recommendations provide a foundation for a comprehensive framework addressing the entire lifecycle of healthcare AI, emphasizing the importance of documentation, training, transparency, accountability, and multi-stakeholder engagement.
A communication-efficient federated learning algorithm to assess racial disparities in post-transplantation survival time.
Wang Y, Zhang D, Tong J … +7 more , He X, Li L, Sun L, Shukla AM, Bian J, Asch DA, Chen Y
J Am Med Inform Assoc · 2025 Dec · PMID 40990064 · Full text
OBJECTIVE: Patients of different race have different outcomes following renal transplantation. Patients of different race also undergo renal transplantation at different hospitals. We used a novel decentralized multisite...
OBJECTIVE: Patients of different race have different outcomes following renal transplantation. Patients of different race also undergo renal transplantation at different hospitals. We used a novel decentralized multisite approach to quantitatively assess the effect of site of care on racial disparities between non-Hispanic Black (NHB) and non-Hispanic White (NHW) patients in post-transplantation survival times. MATERIALS AND METHODS: In this study, we develop a communication-efficient federated learning algorithm to assess site-of-care associated racial disparities based on decentralized time-to-event data, called Communication-Efficient Distributed Analysis for Racial Disparity in Time-to-event Data (CEDAR-t2e). The algorithm includes 2 modules. Module I is to estimate the site-specific proportional hazards model for time-to-event outcomes in a distributed manner, in which the Poissonization is used to simplify the estimation procedure. Based on the estimated results from Module I, Module II calculates how long the kidney failure time of NHB patients would be extended had they been admitted to transplant centers in the same distribution as NHW patients were admitted. RESULTS: With application to United States Renal Data System data covering 39 043 patients across 73 transplant centers, we found no evidence suggesting the presence of site-of-care associated racial disparities in post-transplantation survival times. In particular, restricting to one year after transplantation, the counterfactual graft failure time would have been extended by only 0.61 days on average if NHB had the same admission distribution to transplant centers as NHW patients. DISCUSSION: The proposed approach offers a quantitative measure to evaluate site-of-care associated racial disparities. CONCLUSION: Our approach has the potential to be extended to investigate site-of-care related disparities in other time-to-event outcomes, thus promoting health equity and improving patient health in various fields.
Including AI in diffusion-weighted breast MRI has potential to increase reader confidence and reduce workload.
Bounias D, Simons L, Baumgartner M … +15 more , Ehring C, Neher P, Kapsner LA, Kovacs B, Floca R, Jaeger PF, Eberle J, Hadler D, Laun FB, Ohlmeyer S, Maier-Hein L, Uder M, Wenkel E, Maier-Hein KH, Bickelhaupt S
J Am Med Inform Assoc · 2025 Dec · PMID 40985463 · Full text
OBJECTIVES: Breast diffusion-weighted imaging (DWI) has shown potential as a standalone imaging technique for certain indications, eg, supplemental screening of women with dense breasts. This study evaluates an artificia...
OBJECTIVES: Breast diffusion-weighted imaging (DWI) has shown potential as a standalone imaging technique for certain indications, eg, supplemental screening of women with dense breasts. This study evaluates an artificial intelligence (AI)-powered computer-aided diagnosis (CAD) system for clinical interpretation and workload reduction in breast DWI. MATERIALS AND METHODS: This retrospective IRB-approved study included: n = 824 examinations for model development (2017-2020) and n = 235 for evaluation (01/2021-06/2021). Readings were performed by three readers using either the AI-CAD or manual readings. BI-RADS-like (Breast Imaging Reporting and Data System) classification was based on DWI. Histopathology served as ground truth. The model was nnDetection-based, trained using 5-fold cross-validation and ensembling. Statistical significance was determined using McNemar's test. Inter-rater agreement was calculated using Cohen's kappa. Model performance was calculated using the area under the receiver operating curve (AUC). RESULTS: The AI-augmented approach significantly reduced BI-RADS-like 3 calls in breast DWI by 29% (P =.019) and increased interrater agreement (0.57 ± 0.10 vs 0.49 ± 0.11), while preserving diagnostic accuracy. Two of the three readers detected more malignant lesions (63/69 vs 59/69 and 64/69 vs 62/69) with the AI-CAD. The AI model achieved an AUC of 0.78 (95% CI: [0.72, 0.85]; P <.001), which increased for women at screening age to 0.82 (95% CI: [0.73, 0.90]; P <.001), indicating a potential for workload reduction of 20.9% at 96% sensitivity. DISCUSSION AND CONCLUSION: Breast DWI might benefit from AI support. In our study, AI showed potential for reduction of BI-RADS-like 3 calls and increase of inter-rater agreement. However, given the limited study size, further research is needed.
Development and application of desiderata for automated clinical ordering.
Saleh SN, Johnson KB
J Am Med Inform Assoc · 2025 Dec · PMID 40985459 · Full text
INTRODUCTION: Automation of clinical orders in electronic health records (EHRs) has the potential to reduce clinician burden and enhance patient safety. However, determining which orders are appropriate for automation re...
INTRODUCTION: Automation of clinical orders in electronic health records (EHRs) has the potential to reduce clinician burden and enhance patient safety. However, determining which orders are appropriate for automation requires a structured framework to ensure clinical validity, transparency, and safety. OBJECTIVE: To develop and validate a framework of desiderata for assessing the appropriateness of automating clinical orders in EHRs and to demonstrate its operational value in a live health system dataset. MATERIALS AND METHODS: The study comprised 4 phases to move from concept generation to real-world demonstration. First, we conducted focus group analyses using ground theory to identify themes and developed desiderata informed by these themes and existing literature. We validated the desiderata by surveying clinicians at a single institution, presenting 10 use cases to and assessing perceived appropriateness, cognitive support, and patient safety using a 4-point Likert scale. Survey results were compared to a priori appropriateness designations using t-tests. To evaluate operational impact, we analyzed one year of order-based alerts and orders (1.4 million firings alert and 44.1 million orders, respectively) using filtering rules and association rule mining to identify candidate orders for automation and their impact. RESULTS: We identified 8 desiderata for automated order appropriateness: logical consistency, data provenance, order transparency, context permanence, monitoring plans, trigger consistency, care team empowerment, and system accountability. Use cases deemed appropriate based on these criteria received significantly higher scores for appropriateness (3.13 ± 0.84 vs 2.30 ± 0.99), cognitive support (3.08 ± 0.82 vs 2.25 ± 0.94), and patient safety (3.08 ± 0.86 vs 2.21 ± 0.98) (all P < .001) compared to those considered inappropriate. Operational analysis revealed an alert firing 19 109 times annually, with a 96% signed order rate, where automation could save an estimated 26.5 provider hours per year. Additionally, an association rule with 16 628 occurrences (68.4% confidence) suggested automation could save 15.8 hours annually and yield 8000 additional appropriate orders. DISCUSSION: The desiderata align with clinician perceptions and provide a structured approach for evaluating automated orders. Our findings highlight the potential for automation of certain clinical orders to improve cognitive support while maintaining patient safety. CONCLUSION: Healthcare systems should use these desiderata, coupled with data mining techniques, to systematically identify and govern appropriate automated orders. Further research is needed to validate operational scalability.
Harnessing data to advance health and health equity.
Bakken S
J Am Med Inform Assoc · 2025 Oct · PMID 40981527 · Full text
Abstract loading — click title to view on PubMed.
About
- Frequency
- Sun
- Papers found
- 200
- RSS feed
- Subscribe