Search Results (403)

This paper aims to leverage artificial intelligence (AI) to assist physicians in utilizing advanced deep learning techniques integrated into developed models within electronic health records (EHRs) in medical information systems (MISes), which have been in use for over 15 years in health centers across the Republic of Serbia. This paper presents a human-centered AI approach that emphasizes physician decision-making supported by AI models. This study presents two developed and implemented deep neural network (DNN) models in the EHR. Both models were based on data that were collected during the COVID-19 outbreak. The models were evaluated using five-fold cross-validation. The convolutional neural network (CNN), based on the pre-trained VGG19 architecture for classifying chest X-ray images, was trained on a publicly available smaller dataset containing 196 entries, and achieved an average classification accuracy of 91.83 ± 2.82%. The DNN model for optimizing patient appointment scheduling was trained on a large dataset (341,569 entries) and a rich feature design extracted from the MIS, which is daily used in Serbia, achieving an average classification accuracy of 77.51 ± 0.70%. Both models have consistent performance and good generalization. The architecture of a realized MIS, incorporating the positioning of developed AI tools that encompass both developed models, is also presented in this study. Full article

Background: Biological aging influences cancer outcomes, but its changes during chemotherapy and impact on chemotoxicity in colorectal cancer (CRC) remain underinvestigated. We examined (1) trajectories of biological aging (using Levine Phenotypic Age) during six months of chemotherapy, (2) sociodemographic and clinical risk factors for biological aging, and (3) links between biological aging and chemotoxicity. Methods: Using data from electronic health records (2013–2019) from 1129 adult CRC patients, we computed biological aging (raw Levine Phenotypic Age and its age acceleration [Levine Phenotypic Age–chronological age]) from routine blood tests (e.g., complete blood counts, hepatorenal/inflammatory markers). Chemotoxicity was identified primarily via International Classification of Diseases (ICD-9 and -10) codes. Results: Chemotherapy accelerated biological aging over time. Biological aging at baseline and changes over time predicted chemotoxicity. However, changes in biological aging over time showed stronger associations than baseline biological aging. Advanced cancer stages, higher comorbidity burden, and socioeconomic disadvantage (especially area-level deprivation) were associated with accelerated biological aging at baseline and over time. Biological aging occurred across both young and older adults. Conclusions: Levine Phenotypic Age, computed from routine blood tests in EHRs, offers a feasible clinical tool for aging-related chemotoxicity risk stratification. Validation in diverse cohorts and the development of predictive models are needed. Full article

Healthcare data rapidly increases, and patients seek customized, effective healthcare services. Big data and machine learning (ML) enabled smart healthcare systems hold revolutionary potential. Unlike previous reviews that separately address AI or big data, this work synthesizes their convergence through real-world case studies, cross-domain ML applications, and a critical discussion on ethical integration in smart diagnostics. The review focuses on the role of big data analysis and ML towards better diagnosis, improved efficiency of operations, and individualized care for patients. It explores the principal challenges of data heterogeneity, privacy, computational complexity, and advanced methods such as federated learning (FL) and edge computing. Applications in real-world settings, such as disease prediction, medical imaging, drug discovery, and remote monitoring, illustrate how ML methods, such as deep learning (DL) and natural language processing (NLP), enhance clinical decision-making. A comparison of ML models highlights their value in dealing with large and heterogeneous healthcare datasets. In addition, the use of nascent technologies such as wearables and Internet of Medical Things (IoMT) is examined for their role in supporting real-time data-driven delivery of healthcare. The paper emphasizes the pragmatic application of intelligent systems by highlighting case studies that reflect up to 95% diagnostic accuracy and cost savings. The review ends with future directions that seek to develop scalable, ethical, and interpretable AI-powered healthcare systems. It bridges the gap between ML algorithms and smart diagnostics, offering critical perspectives for clinicians, data scientists, and policymakers. Full article

Background: Unlike most cancers, breast cancer poses a persistent risk of distant recurrence—often years after initial treatment—making long-term risk stratification uniquely challenging. Current tools fall short in predicting late metastatic events, particularly for early-stage patients. Methods: We present an interpretable machine learning (ML) pipeline to predict distant recurrence-free survival at 5, 10, and 15 years, integrating Bayesian network-based causal feature selection, deep feed-forward neural network models (DNMs), and SHAP-based interpretation. Using electronic health record (EHR)-based clinical data from over 6000 patients, we first applied the Markov blanket and interactive risk factor learner (MBIL) to identify minimally sufficient predictor subsets. These were then used to train optimized DNM classifiers, with hyperparameters tuned via grid search and benchmarked against models from 10 traditional ML methods and models trained using all predictors. Results: Our best models achieved area under the curve (AUC) scores of 0.79, 0.83, and 0.89 for 5-, 10-, and 15-year predictions, respectively—substantially outperforming baselines. MBIL reduced input dimensionality by over 80% without sacrificing accuracy. Importantly, MBIL-selected features (e.g., nodal status, hormone receptor expression, tumor size) overlapped strongly with top SHAP contributors, reinforcing interpretability. Calibration plots further demonstrated close agreement between predicted probabilities and observed recurrence rates. The percentage performance improvement due to grid search ranged from 25.3% to 60%. Conclusions: This study demonstrates that combining causal selection, deep learning, and grid search improves prediction accuracy, transparency, and calibration for long-horizon breast cancer recurrence risk. The proposed framework is well-positioned for clinical use, especially to guide long-term follow-up and therapy decisions in early-stage patients. Full article

The increasing complexity and scale of electronic health records (EHRs) demand advanced tools for efficient data retrieval, summarization, and comparative analysis in clinical practice. MERA (Medical Electronic Records Assistant) is a Retrieval-Augmented Generation (RAG)-based AI system that addresses these needs by integrating domain-specific retrieval with large language models (LLMs) to deliver robust question answering, similarity search, and report summarization functionalities. MERA is designed to overcome key limitations of conventional LLMs in healthcare, such as hallucinations, outdated knowledge, and limited explainability. To ensure both privacy compliance and model robustness, we constructed a large synthetic dataset using state-of-the-art LLMs, including Mistral v0.3, Qwen 2.5, and Llama 3, and further validated MERA on de-identified real-world EHRs from the MIMIC-IV-Note dataset. Comprehensive evaluation demonstrates MERA’s high accuracy in medical question answering (correctness: 0.91; relevance: 0.98; groundedness: 0.89; retrieval relevance: 0.92), strong summarization performance (ROUGE-1 F1-score: 0.70; Jaccard similarity: 0.73), and effective similarity search (METEOR: 0.7–1.0 across diagnoses), with consistent results on real EHRs. The similarity search module empowers clinicians to efficiently identify and compare analogous patient cases, supporting differential diagnosis and personalized treatment planning. By generating concise, contextually relevant, and explainable insights, MERA reduces clinician workload and enhances decision-making. To our knowledge, this is the first system to integrate clinical question answering, summarization, and similarity search within a unified RAG-based framework. Full article

Background: Among all ambulatory antibiotic prescriptions, about 20% are non-visit-based (ordered outside of an in-person clinical encounter), and about 30% are not associated with an infection-related diagnosis code. Objective/Methods: To identify the rationale for ambulatory antibiotic prescribing, we queried the electronic health record (EHR) of a single, large health system in the Midwest United States to identify all oral antibiotics prescribed from November 2018 to February 2019 and examined visit, procedure, lab, department, and diagnosis codes. For the remaining antibiotic prescriptions—mostly non-visit-based, no-infectious-diagnosis-documented—we randomly selected and manually reviewed the EHR to identify a prescribing rationale and, if none was present, surveyed prescribers for their rationale. Results: During the study period, there were 47,619 antibiotic prescriptions from 1177 clinicians to 41,935 patients, of which 2608 (6%) were eligible non-visit-based, no-infectious-diagnosis-documented. We randomly selected 2298. There was a documented rationale for 2116 (92%) prescriptions. The most common documented reasons—not mutually exclusive—were patient-reported symptoms (71%), persistence of symptoms after initial management (18%), travel (17%), and responding to lab or imaging results (11%). We contacted 160 clinicians who did not document any prescribing rationale in the EHR and received responses from 62 (39%). Clinicians’ stated reasons included upcoming or current patient travel (19%), the antibiotic was for the prescriber’s own family member (19%), or the clinician made a diagnosis but did not document it in the EHR (18%). Conclusions: Non-visit-based, no-infectious-diagnosis-documented antibiotic prescriptions were most often in response to patient-reported symptoms, though they also occur for a variety of other reasons, some problematic, like in the absence of documentation or for a family member. Full article

This review investigates recent developments in cardiac mechano-electrical-fluid interaction (MEFI) modeling, with a focus on multiphysics simulation platforms and digital twin frameworks developed between 2015 and 2025. The purpose of the study is to assess how computational modeling methods—particularly finite element and immersed boundary techniques, monolithic and partitioned coupling schemes, and artificial intelligence (AI)-enhanced surrogate modeling—capture the integrated dynamics of cardiac electrophysiology, tissue mechanics, and hemodynamics. The goal is to evaluate the translational potential of MEFI models in clinical applications such as cardiac resynchronization therapy (CRT), arrhythmia classification, atrial fibrillation ablation, and surgical planning. Quantitative results from the literature demonstrate <5% error in pressure–volume loop predictions, >0.90 F1 scores in machine-learning-based arrhythmia detection, and <10% deviation in myocardial strain relative to MRI-based ground truth. These findings highlight both the promise and limitations of current MEFI approaches. While recent advances improve physiological fidelity and predictive accuracy, key challenges remain in achieving multiscale integration, model validation across diverse populations, and real-time clinical applicability. The review concludes by identifying future milestones for clinical translation, including regulatory model certification, standardization of validation protocols, and integration of patient-specific digital twins into electronic health record (EHR) systems. Full article

Background/Objectives: Electronic medical record systems play a crucial role in the operation of modern healthcare institutions, enabling the foundational data necessary for advancements in personalized medicine. Despite their importance, the software supporting these systems frequently experiences data availability and integrity issues, particularly concerning patients’ personal information. This study aims to present a decentralized architecture that integrates both clinical and personal patient data, with a provenance mechanism to enable data tracing and auditing, ultimately supporting more precise and personalized healthcare decisions. Methods: A system implementation based on the solution was developed, and a feasibility study was conducted with synthetic medical records data. Results: The system was able to correctly receive data of 190 instances of the entities designed, which included different types of medical records, and generate 573 provenance entries that captured in detail the context of the associated medical information. Conclusions: For the first cycle of the research, the system developed served to validate the main features of the solution, and through that, it was possible to infer the feasibility of a decentralized EHR and PHR health system with formal provenance data tracking. Such a system lays a robust foundation for secure and reliable data management, which is essential for the effective implementation and future development of personalized medicine initiatives. Full article

Background/Objectives: The overuse of short-acting β₂-agonists (SABAs) has been associated with increased asthma morbidity and mortality, prompting changes in treatment guidelines. However, the role of frequent short-acting muscarinic antagonists (SAMAs) use remains poorly defined and unaddressed in current recommendations. This study offers the first real-world analysis of SAMA overuse in asthma, quantifying its association with exacerbation risk and healthcare utilization and comparing its predictive value to that of SABAs. Methods: A retrospective multicenter cohort study analyzed electronic health records (EHRs) from 132 adults with asthma in the Spanish National Health System (SNS). Associations between annual SAMA use and clinical outcomes were assessed using negative binomial regression and 5000-sample bootstrap simulations. Interaction and threshold models were applied to explore how SAMA use affected outcomes and identify clinically actionable cutoffs. Results: SAMA use was independently associated with a 19.2% increase in exacerbation frequency per canister and a nearly sixfold increase in the odds of experiencing ≥1 exacerbation (OR = 5.97; 95% CI: 2.43–14.66). An inflection point at 2.5 canisters/year marked the threshold beyond which annual exacerbations exceeded one. Increased SAMA use was also associated with a higher number of respiratory consultations and with more frequent prescriptions of systemic corticosteroids and antibiotics. The risk increased more sharply with SAMAs than with SABAs, and the lack of correlation between them suggests distinct clinical patterns underlying their use. Conclusions: SAMA use emerges as a digitally traceable and clinically meaningful indicator of asthma instability. While the associations observed are robust and consistent across multiple outcomes, they should be considered provisional due to the study’s retrospective design and limited sample size. Replication in larger and more diverse cohorts is needed to confirm external validity. These findings support the integration of SAMA tracking into asthma management tools—alongside SABAs—to enable the earlier identification of uncontrolled disease and guide therapeutic adjustment. Full article

The integration of artificial intelligence into healthcare has advanced rapidly in recent years, with multimodal approaches emerging as promising tools for improving diagnostic accuracy and clinical decision making. These approaches combine heterogeneous data sources such as medical images, electronic health records, physiological signals, and clinical notes to better capture the complexity of disease processes. Despite this progress, only a limited number of studies offer a unified view of multimodal AI applications in medicine. In this review, we provide a comprehensive and up-to-date analysis of machine learning and deep learning-based multimodal architectures, fusion strategies, and their performance across a range of diagnostic tasks. We begin by summarizing publicly available datasets and examining the preprocessing pipelines required for harmonizing heterogeneous medical data. We then categorize key fusion strategies used to integrate information from multiple modalities and overview representative model architectures, from hybrid designs and transformer-based vision-language models to optimization-driven and EHR-centric frameworks. Finally, we highlight the challenges present in existing works. Our analysis shows that multimodal approaches tend to outperform unimodal systems in diagnostic performance, robustness, and generalization. This review provides a unified view of the field and opens up future research directions aimed at building clinically usable, interpretable, and scalable multimodal diagnostic systems. Full article

Background: The effective use of electronic health records (EHRs) is an essential clinical skill, but medical schools have traditionally provided limited systematic teaching on the topic. Inefficient use of EHRs results in delays in diagnosis, fragmented care, and clinician burnout. This study investigates the impact on medical students’ confidence, efficiency, and proficiency in extracting clinically pertinent information from patient records following an organised EHR teaching programme. Methods: This observational cohort involved 60 final-year medical students from three London medical schools. Participants received a structured three-phase intervention involving an introductory workshop, case-based hands-on practice, and guided reflection on EHR navigation habits. Pre- and post-intervention testing involved mixed-method surveys, simulated case tasks, and faculty-assessed data retrieval exercises to measure changes in students’ confidence, efficiency, and ability to synthesise patient information. Quantitative data were analysed using paired t-tests, while qualitative reflections were theme-analysed to identify shifts in clinical reasoning. Results: All 60 students successfully finished the intervention and assessments. Pre-intervention, only 28% students reported feeling confident in using EHRs effectively, with a confidence rating of 3.0. Post-intervention, 87% reported confidence with a rating of 4.5 (p < 0.01). Efficiency in the recovery of critical patient information improved from 3.2 to 4.6 (p < 0.01). Students also demonstrated enhanced awareness regarding system-related issues, such as information overload and fragmented documentation, and provided recommendations on enhancing data synthesis for clinical decision making. Conclusions: This study emphasises the value of structured EHR instruction in enhancing the confidence and proficiency of medical students in using electronic records. The integration of structured EHR education to medical curricula can better prepare future physicians in managing information overload, improve diagnostic accuracy, and enhance the quality of patient care. Future research should explore the long-term impact of structured EHR training on clinical performance, diagnostic accuracy, and patient outcomes during real-world clinical placements and postgraduate training. Full article

(1) Background: The integration of digital technologies such as electronic health records (EHRs), telepsychiatry, and communication platforms has transformed the mental health sector a lot compared to in previous years. While these tools enhance service delivery, they also introduce unique stressors. Despite growing concerns, there is no validated instrument specifically designed to measure the digital stress experienced by mental health professionals. (2) Methods: This study involved the development and psychometric validation of the Digital Stress Scale (DSC). The process included item generation through a literature review and qualitative interviews, expert panel validation, and a two-phase statistical evaluation. Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) were conducted on responses from 423 licensed mental health professionals using EHRs and digital communication tools. The scale’s reliability and convergent validity were assessed via internal consistency and correlations with established mental health measures. (3) Results: The final DSC included four subscales: digital fatigue, technostress, digital disengagement, and work–life digital boundaries. CFA supported the factor structure (CFI = 0.965, RMSEA = 0.038), and the overall reliability was acceptable (Cronbach’s Alpha = 0.87). Descriptive analysis showed moderate-to-high levels of digital stress (M = 11.94, SD = 2.72). Digital fatigue was the strongest predictor of total stress (β = 1.00, p < 0.001), followed by technostress and work–life boundary violations. All subscales were significantly correlated with burnout (r = 0.72), job dissatisfaction (r = −0.61), and perceived stress (r = 0.68), all with a p < 0.001. (4) Conclusions: The DSC is a valid and reliable tool for assessing digital stress among mental health professionals. Findings point out the urgent need for policy-level interventions to mitigate digital overload, promote healthy work–life boundaries, and enhance digital competency in mental health settings. Full article

The increasing integration of digital technologies in healthcare has expanded the attack surface for privacy violations in critical systems such as electronic health records (EHRs), telehealth platforms, and medical device software. However, current vulnerability detection datasets lack domain-specific privacy annotations essential for compliance with healthcare regulations like HIPAA and GDPR. This study presents C3-VULMAP, a novel and large-scale dataset explicitly designed for privacy-aware vulnerability detection in healthcare software. The dataset comprises over 30,000 vulnerable and 7.8 million non-vulnerable C/C++ functions, annotated with CWE categories and systematically mapped to LINDDUN privacy threat types. The objective is to support the development of automated, privacy-focused detection systems that can identify fine-grained software vulnerabilities in healthcare environments. To achieve this, we developed a hybrid construction methodology combining manual threat modeling, LLM-assisted synthetic generation, and multi-source aggregation. We then conducted comprehensive evaluations using traditional machine learning algorithms (Support Vector Machines, XGBoost), graph neural networks (Devign, Reveal), and transformer-based models (CodeBERT, RoBERTa, CodeT5). The results demonstrate that transformer models, such as RoBERTa, achieve high detection performance (F1 = 0.987), while Reveal leads GNN-based methods (F1 = 0.993), with different models excelling across specific privacy threat categories. These findings validate C3-VULMAP as a powerful benchmarking resource and show its potential to guide the development of privacy-preserving, secure-by-design software in embedded and electronic healthcare systems. The dataset fills a critical gap in privacy threat modeling and vulnerability detection and is positioned to support future research in cybersecurity and intelligent electronic systems for healthcare. Full article

To elucidate the changes in gas exchange across the life course, we estimated the age trajectories of O₂ saturation, CO₂ (as either end-tidal or serum bicarbonate), resting heart rate, and resting respiratory rate from age 2 yr onward in female and male patients separately. We utilized two sources’ data: electronic health records (EHR) representing ambulatory visits of approximately 53,000 individuals and sleep clinic polysomnogram (PSG) records representing an additional ~21,000. We used linear regression to estimate age-group-specific mean response levels for women and men. We compared estimated female–male differences between pre- and post-pubertal children and between pre- and post-menopausal periods among adults. Women between 15 and 45 years had higher O₂ saturation and lower serum bicarbonate levels or end-tidal CO₂ levels than men of similar ages. For O₂ saturation and for both measures of CO₂, the female–male difference was larger on average among adults at pre-menopausal ages than those at post-menopausal ages. Women had higher O₂ saturation throughout their lives than men; however, the difference disappeared in the elderly. Women between menarche and menopause had significantly lower end-tidal CO₂ and serum bicarbonate than men of similar ages. After menopause, however, women appeared to have higher mean levels of both end-tidal CO₂ and serum bicarbonate than men. Full article

Early detection of morbidities that complicate pregnancy improves health outcomes in low- and middle-income countries. Automatic revision of electronic health records (EHRs) can help identify such morbidity risks. There is a lack of corpora to train models in Spanish in specific domains, and there are no models specialized in maternal EHRs. This study aims to develop a fine-tuned model that detects clinical concepts using a built database with text extracted from maternal EHRs in Spanish. We created a corpus with 13.998 annotations from 200 clinical notes in Spanish associated with EHRs obtained from a reference institution of high obstetric risk in Colombia. Using the Beginning–Inside–Outside tagging scheme, we fine-tuned five different transformer-based models to classify between 16 classes associated with eight entities. The best model achieved a macro F1 score of 0.55 ± 0.03. The entities with the best performance were signs, symptoms, and negations, with exact F1 scores of 0.714 and 0.726, respectively. The lower scores were associated with those classes with fewer observations. Even though our dataset is shorter in size and more diverse in entity types than other datasets in Spanish, our results are comparable to other state-of-the-art named entity recognition models fine-tuned in Spanish and the biomedical domain. This work introduces the first fine-tuning of a model for named entity recognition specifically designed for maternal EHRs. Our results can be used as a base to develop new models to extract concepts in the maternal–fetal domains and help healthcare providers detect morbidities that complicate pregnancy early. Full article