International Journal Of Medical Informatics[JOURNAL]
Sun
200 papers
Utilizing virtual reality to improve ICU visual displays.
Zhang N, Zou J, Kwon S … +5 more , Essani S, Patel H, Tong A, Shive J, Schlesinger J
Int J Med Inform · 2026 Jun · PMID 41795495 · Publisher ↗
OBJECTIVE: Intensive care unit (ICU) clinicians respond to over 900 alarms per day, resulting in sensory overload that leads to delayed and inaccurate clinician responses. Integrating vital signs into object-based visual...
OBJECTIVE: Intensive care unit (ICU) clinicians respond to over 900 alarms per day, resulting in sensory overload that leads to delayed and inaccurate clinician responses. Integrating vital signs into object-based visual displays has been shown to yield quicker and more accurate responses than standard ICU alarm displays. We designed a ring-based visual alarm system that integrates heart rate (HR), blood pressure (BP), and peripheral oxygen sat-uration (SpO) and evaluated its effectiveness using a virtual reality (VR) testing environment. VR offers significant advantages over traditional setups, notably by providing highly controlled, immersive, and replicable testing en-vironments. METHODS: This display was compared against an existing graph-based dis-play that had been previously proven more effective than standard numerical ICU alarms. Using an HTC VIVE Pro Eye headset, we created controlled virtual environments to test participant speed and accuracy in identifying simulated clinical events. Tests were conducted with the graph-based, ring-based, and both displays shown, using head orientation software to determine which display participants viewed. RESULTS: When shown individually, the ring-based display had 33% higher accuracy and a 1.00 s faster mean response time than the graph-based dis-play. When both displays were present, subjects viewed the ring-based dis-play 82.2% of the time. Display location in the visual field had no significant impact on performance. CONCLUSION: The ring-based display yielded improved alarm identification accuracy and response time compared to the graph-based alarm display. VR provided a robust, controlled method for testing medical interfaces. Future work should implement eye-tracking into alarm evaluation. This work's end goal is to combine our visual alarm with auditory alarms to create a multi-sensory alarm system.
An AI-assisted framework for the ethical use of machine learning in healthcare.
Georgiou GP
Int J Med Inform · 2026 Jun · PMID 41795494 · Publisher ↗
This study develops and evaluates ETHICS, a concise, clinician-facing ethical protocol for the routine use of machine learning (ML) in healthcare. Using ChatGPT for first-stage drafting, we generated six actionable princ...
This study develops and evaluates ETHICS, a concise, clinician-facing ethical protocol for the routine use of machine learning (ML) in healthcare. Using ChatGPT for first-stage drafting, we generated six actionable principles - Equity and Fairness, Transparency and Patient-Centered Care, Human Oversight and Clinical Integrity, Information Privacy and Data Governance, Continuous Improvement and Sustainability, and Support and Education for Professionals - structured as a mnemonic to support uptake in time-constrained clinical settings. Outputs were treated as provisional and refined through mandatory human source verification, iterative readability optimization, multidisciplinary expert review, and scenario-based stress testing. Readability analysis showed substantial improvement from high complexity to clinician-accessible language. Expert ratings indicated strong endorsement with excellent inter-rater reliability. Implementation readiness was confirmed across five clinical scenarios, all passing predefined adequacy thresholds. The findings suggest generative AI can accelerate ethical protocol drafting, but validity and practicality depend on structured human oversight and real-world testing.
Artificial intelligence applications in emergency triage for suspected acute coronary syndrome: a systematic review.
Fatai A, Sattayarom C
Int J Med Inform · 2026 Jun · PMID 41795351 · Publisher ↗
BACKGROUND: Acute coronary syndromes (ACS) are time-critical conditions requiring rapid and accurate triage in the emergency department. Traditional triage may lead to delays, whereas artificial intelligence (AI) has the...
BACKGROUND: Acute coronary syndromes (ACS) are time-critical conditions requiring rapid and accurate triage in the emergency department. Traditional triage may lead to delays, whereas artificial intelligence (AI) has the potential to enhance triage accuracy and efficiency. OBJECTIVE: This study aimed to synthesize and evaluate the effectiveness of AI applications in emergency triage or early clinical decision-making for suspected ACS, focusing on their impact on diagnostic accuracy, time to treatment, and operational outcomes. METHODS: This systematic review followed PRISMA guidelines. PubMed, Scopus, CINAHL, and IEEE Xplore were searched for studies published between 2020 and 2025. Eligible articles examined AI applications for triage of suspected ACS in the emergency department. Data on study design, AI models, comparators, inputs, and outcomes were extracted, and study quality was assessed using the Mixed Methods Appraisal Tool (MMAT). RESULTS: Fifteen studies from multiple countries and study designs were included. Convolutional neural networks and ensemble learning methods were the most used models. AI models generally demonstrated high diagnostic performance (AUROC 0.82-0.99), were associated with reductions in treatment times such as door-to-balloon and catheterization intervals, and showed potential to improve operational outcomes, including resource utilization and patient flow. However, findings varied across studies depending on model type, data inputs, and study design. CONCLUSIONS: AI-assisted triage for suspected ACS shows promise in supporting clinical decision-making and improving workflow efficiency. However, substantial heterogeneity and limited prospective validation suggest that findings should be interpreted cautiously. Further research is needed to confirm clinical effectiveness, generalizability, and safe implementation.
Accuracy, Efficiency, and usability of a Semi-Automated SMART on FHIR venous thromboembolism risk assessment App: A randomised crossover simulation study.
Cahill M, O'Shaughnessy F, Boland F … +4 more , Ainle FN, Donnelly J, Cullinan S, Cleary BJ
Int J Med Inform · 2026 Jun · PMID 41785738 · Publisher ↗
BACKGROUND: Venous thromboembolism (VTE) is the leading cause of maternal death. Current VTE risk assessment (VTERA) tools often require manual data entry, which can reduce accuracy and affect thromboprophylaxis decision...
BACKGROUND: Venous thromboembolism (VTE) is the leading cause of maternal death. Current VTE risk assessment (VTERA) tools often require manual data entry, which can reduce accuracy and affect thromboprophylaxis decisions. Building on 'Thrombocalc', an established postpartum manual VTERA tool, we developed a semi-automated VTERA application ('app') using SMART on FHIR. This app integrates with the electronic health record (EHR) to automatically extract 11 risk factors. OBJECTIVES: The primary objective was to evaluate the accuracy of a semi-automated VTERA app. The secondary objectives were to evaluate efficiency and usability. PATIENTS/METHODS: A randomised crossover study was conducted at a tertiary maternity hospital. Healthcare professionals (HCPs) completed ten simulated patient assessments using both manual and semi-automated VTERA tools. Accuracy was evaluated based on the inclusion of appropriate risk factors, accuracy of risk scores, and appropriateness of thromboprophylaxis recommendations. These recommendations pertained to thromboprophylaxis indication, duration, and dosage. Mouse clicks, keystrokes, and completion times were logged to assess efficiency. Usability of the semi-automated tool was evaluated using the System Usability Scale (SUS). Linear and logistic mixed-effects models were used to analyse accuracy and efficiency outcomes. RESULTS: Thirty-one HCPs participated. The proportion of accurate risk scores increased from 63% with the manual tool, to 87% with the semi-automated tool. Accurate thromboprophylaxis recommendations increased from 79% to 91%. Mixed-effects models demonstrated higher odds of producing accurate risk scores (OR 6.99, 95% CI 3.31-14.67) and recommendations (OR 3.59, 95% CI 1.64-7.88) with the semi-automated tool. Task completion time decreased by 49 s (95% CI 32.8-65). Mouse clicks were reduced by 12 clicks (95% CI 9.01-15.37), and keystrokes by 94% (95% CI 93.4-94.2%). The semi-automated tool achieved a mean SUS score of 89.8, indicating excellent usability. CONCLUSION: Automation can enhance postpartum VTE risk assessment. However, optimal performance requires reliable data and HCP oversight.
From algorithmic innovation to clinical deployment: A systematic review of methodological gaps limiting federated learning in healthcare.
Vijayasarathy S
Int J Med Inform · 2026 Jun · PMID 41785737 · Publisher ↗
OBJECTIVE: Federated learning (FL) is a distributed machine learning paradigm designed to enable model training across decentralized data sources without requiring data centralization. This review critically examines FL...
OBJECTIVE: Federated learning (FL) is a distributed machine learning paradigm designed to enable model training across decentralized data sources without requiring data centralization. This review critically examines FL as a methodological approach in biomedical informatics, summarizing its conceptual foundations, methodological variants, successes, and limitations, and identifying directions for future research. METHODS: We conducted a systematic methodological review following PRISMA 2020 guidelines. Searches were performed in PubMed, Web of Science, Scopus, IEEE Xplore, and Google Scholar for studies published between January 2017 and March 2025. Articles were included if they proposed, extended, or critically analyzed FL methods for biomedical or life science data. Methods were categorized by federation topology, optimization strategy, data heterogeneity handling, privacy mechanisms, evaluation design, and translational readiness. RESULTS: From 8,412 records, 97 articles met inclusion criteria. Rapid methodological innovation exists in optimization and privacy mechanisms, while support for design-time evaluation and governance remains limited. CONCLUSION: Federated learning represents a significant methodological advance in biomedical informatics but current implementations address only a subset of translational challenges. Future work must integrate study design, evaluation, interpretability, and governance into FL methods.
Accuracy and completeness of large language models in Epidemic keratoconjunctivitis Queries: A Comparative study.
Es'haghi A, Aliyariparand M, Jamalipour Soufi K … +1 more , Aghaei H
Int J Med Inform · 2026 Jun · PMID 41747323 · Publisher ↗
BACKGROUND: Large language models (LLMs) are increasingly applied in clinical contexts, yet their reliability in disease-specific ophthalmic domains remains insufficiently characterized. Epidemic keratoconjunctivitis (EK...
BACKGROUND: Large language models (LLMs) are increasingly applied in clinical contexts, yet their reliability in disease-specific ophthalmic domains remains insufficiently characterized. Epidemic keratoconjunctivitis (EKC), a highly contagious adenoviral ocular infection, represents a critical area where accurate diagnostic and therapeutic information is essential. This study aimed to assess and compare the accuracy and completeness of responses generated by two widely used LLMs, namely ChatGPT and Meta AI, to standardized ophthalmology questions related to EKC. METHODS: This cross-sectional expert evaluation study was conducted between January and March 2025. Thirty-four structured questions covering five EKC domains (etiology, clinical features, diagnosis, treatment, and prognosis) were developed using authoritative ophthalmology sources. Responses generated by ChatGPT (Plus subscription, web interface) and Meta AI (web interface) were independently scored by two fellowship-trained cornea specialists using a five-point Likert scale for accuracy and completeness. Inter-rater agreement was assessed using Cohen's κ and mean absolute error (MAE). RESULTS: Both models demonstrated strong overall performance, with mean accuracy and completeness scores of 4.87 and 4.90 for ChatGPT and 4.87 and 4.93 for Meta AI, respectively. Agreement between expert raters was substantial (κ range 0.78-0.88). Domain-specific analysis revealed consistently high scores across all categories, with Meta AI showing slightly higher completeness, particularly in diagnostic and symptom-related responses. CONCLUSION: LLMs exhibit consistently strong performance and domain consistency when addressing ophthalmic queries related to EKC. Their performance supports their potential use as adjunct informational and educational tools for ophthalmic conditions such as EKC, particularly for structured knowledge retrieval, treatment explanation, and patient counseling, provided that use remains under appropriate professional oversight.
Machine learning for site risk prediction in clinical trials: development, external validation, and operational application in site qualification.
Yang Z
Int J Med Inform · 2026 May · PMID 41741318 · Publisher ↗
BACKGROUND: Site selection and qualification represent critical operational challenges in clinical trials, particularly in rare diseases like transthyretin amyloid cardiomyopathy (ATTR-CM)-a progressive cardiac disease c...
BACKGROUND: Site selection and qualification represent critical operational challenges in clinical trials, particularly in rare diseases like transthyretin amyloid cardiomyopathy (ATTR-CM)-a progressive cardiac disease caused by extracellular deposition of misfolded transthyretin protein in the myocardium affecting <50,000 US patients. Current Site Qualification Risk Assessment (SQRA) approaches-expert judgment, rule-based scoring, or historical performance-lack systematic validation. From a health informatics perspective, clinical trial data stored in disparate systems (eTMF, CTMS, ClinicalTrials.gov) remain underutilized for proactive risk prediction. We developed and validated an ML-based clinical decision support tool for site risk prediction in ATTR-CM trials with external validation in Duchenne muscular dystrophy (DMD), designed for integration into SQRA workflows to guide Clinical Project Managers (CPMs) in site selection and Site Qualification Visit (SQV) decisions. METHODS: We analyzed 460 sites from 42 ATTR-CM studies and 761 sites from 89 DMD studies using data from ClinicalTrials.gov. A hybrid risk scoring system (expert consensus + LASSO optimization) captured enrollment, data quality, compliance, and complexity dimensions. Four ML algorithms were trained (n = 368) and validated (n = 92) using 5-fold cross-validation and temporal validation. External validation used DMD sites from different disease context and time period. SHAP analysis with bootstrap confidence intervals (1000 iterations) and permutation tests ensured interpretability. We designed an operational workflow integrating the tool into SQRA processes and projected impact based on literature-reported trial parameters. RESULTS: Support Vector Machine achieved 98.91% accuracy (95% CI: 93.8%-99.9%) in ATTR-CM, correctly classifying 91 of 92 sites. Subgroup analysis showed 100% accuracy across all geographic regions and site types, confirming no algorithmic bias. External validation in DMD yielded 81.21% accuracy (95% CI: 73.3%-88.5%), demonstrating cross-disease generalizability. SHAP identified data quality risk (0.112, p < 0.001), screen failure rate (0.090, p < 0.001), and enrollment risk (0.084, p < 0.001) as key predictors. Feature importance rankings showed high consistency between diseases (Spearman ρ = 0.905, p = 0.002). CONCLUSIONS: We developed an operational-ready ML-based tool for site risk prediction, achieving near-perfect accuracy in ATTR-CM with demonstrated cross-disease generalizability in DMD. The tool is designed for integration into SQRA workflows to support CPMs in site selection and SQV decisions. Feature importance consistency (Spearman ρ = 0.905) supports applicability across diverse trial portfolios. However, critical limitations remain: lack of validation against ground truth monitoring outcomes, DMD features estimated using proxies, and projected impact not yet validated through operational deployment. This study demonstrates how health informatics can provide decision support for clinical research operations, with pilot implementation essential to validate projected benefits.
Human in the loop artificial intelligence in healthcare: applications, outcomes, and implementation challenges.
Olawade DB, Plabon SB, Ojo A … +3 more , Ogunbona MA, Makanjuola BD, Olasilola OR
Int J Med Inform · 2026 Jun · PMID 41740273 · Publisher ↗
BACKGROUND: The integration of artificial intelligence in healthcare has transformed clinical practice and research methodologies. However, concerns regarding algorithmic accountability, interpretability, and safety have...
BACKGROUND: The integration of artificial intelligence in healthcare has transformed clinical practice and research methodologies. However, concerns regarding algorithmic accountability, interpretability, and safety have necessitated human oversight in AI systems. Human in the loop artificial intelligence represents a collaborative paradigm where human expertise and machine intelligence converge to enhance decision making while maintaining ethical standards and clinical safety. AIM: This review synthesizes current evidence on human in the loop AI in healthcare delivery and research, examining implementation frameworks, clinical outcomes, comparative advantages over fully automated and clinician-only approaches, and challenges. METHOD: A comprehensive narrative review was conducted using PubMed, Scopus, Web of Science, and IEEE Xplore databases covering studies from 2018 to 2025. Data were thematically synthesized to identify patterns, frameworks, and outcomes. This narrative approach enables comprehensive conceptual synthesis across diverse HITL-AI applications and contexts. RESULTS: Human in the loop AI demonstrates significant applications across diagnostic imaging, clinical decision support, patient monitoring, drug discovery, and research data analysis. Evidence indicates improved diagnostic accuracy, reduced medical errors, enhanced patient safety, and increased clinician trust compared to both automated AI and traditional approaches. Implementation requires EHR interoperability, clear liability frameworks, adaptive training protocols, and quantum-safe cryptographic security. Challenges include workflow integration, regulatory gaps for adaptive systems, and sustainability concerns. CONCLUSION: This review advances the field by synthesizing cross-domain implementation patterns, mapping collaboration models to risk-stratified contexts, identifying regulatory gaps for adaptive systems, and proposing future directions including post-quantum cryptographic integration, AI-driven adaptive architectures, and multi-center scalability frameworks for optimizing human-machine collaboration in healthcare.
Enhancing fairness and standardization in AI-versus-physician diagnostic comparisons: A scoping review.
Chen X, Xu H, Huang Y … +6 more , Zhu F, Tang J, Li H, Min X, Hu L, Lu L
Int J Med Inform · 2026 Jun · PMID 41734515 · Publisher ↗
OBJECTIVE: The growing number of studies directly comparing artificial intelligence (AI) to physicians in diagnostic tasks often focuses on performance outcomes, overlooking fundamental methodological rigor. This scoping...
OBJECTIVE: The growing number of studies directly comparing artificial intelligence (AI) to physicians in diagnostic tasks often focuses on performance outcomes, overlooking fundamental methodological rigor. This scoping review aims to critically appraise the methodological quality of this body of literature, identifying key challenges and proposing a framework to enhance the fairness, standardization, and clinical relevance of future comparisons. MATERIALS AND METHODS: We conducted a systematic search of PubMed, Scopus, and Web of Science for studies published between January 1, 2020, and October 31, 2025, following the PRISMA-ScR guidelines. From 8,851 screened records, 120 studies met the inclusion criteria for direct AI-physician comparison. Data on study characteristics, dataset quality, task design, physician configuration, and reporting transparency were extracted and synthesized narratively. RESULTS: Our analysis of 120 studies revealed a field characterized by significant methodological heterogeneity. Key issues include a predominant focus on retrospective studies (75.8%), frequent information asymmetry between AI and physicians (20.8%), limited clinical relevance in task design despite superficial fidelity, and insufficient physician sample sizes (60.8% had ≤ 10 readers). Furthermore, we found a widespread neglect of time constraints (absent in 50.8% of studies) and a critical lack of transparency regarding code and data availability. CONCLUSION: Current research on AI-physician diagnostic comparisons is often hampered by methodological weaknesses that undermine the validity and generalizability of its findings. To ensure the generation of reliable and clinically meaningful evidence, future studies must prioritize prospective designs, ensure fairness in experimental conditions, and adhere to higher standards of transparency. We propose the AI vs. Physician Study Checklist (AIPSC) as a practical tool to guide the design and reporting of more robust and systematic evaluations, ultimately fostering the responsible integration of AI into clinical practice.
Transforming nursing documentation data into the Observational Medical Outcomes Partners common data model.
Jung H, Yoo S, Kim S … +2 more , Chung J, Lee HY
Int J Med Inform · 2026 Jun · PMID 41734431 · Publisher ↗
BACKGROUND: Electronic health records (EHRs) provide clinical evidence for observational studies. Of these, nursing documentation data reflect patients' problems or situations and nursing services that are not available...
BACKGROUND: Electronic health records (EHRs) provide clinical evidence for observational studies. Of these, nursing documentation data reflect patients' problems or situations and nursing services that are not available from other data sources; however, they have not been actively utilized in research owing to their low quality of documentation. OBJECTIVE: The objectives of this study were to 1) transform nursing documentation data into the Observational Medical Outcomes Partnership (OMOP) common data model (CDM) format and 2) generate a cohort of inpatients with nausea by utilizing transformed nursing documentation data to present the effectiveness of standardization. METHODS: A total of 4006 unique nursing statements used in nursing documentation were extracted from the EHRs of a tertiary general hospital in a South Korean metropolitan area. They were standardized primarily using Systematized Nomenclature Of Medicine Clinical Terms (SNOMED CT), one of the OMOP vocabularies, according to the mapping principles and guidelines. After converting the data into the OMOP CDM format, a cohort of inpatients with nausea was generated by utilizing nursing statements mapped into the "nausea," "nausea care," and "nausea care education" concepts. We then compared the size and demographic characteristics of the cohort with those of a cohort generated based on the diagnosis and chief complaint of nausea. RESULTS: Of the 4006 unique nursing statements, 98.9% were mapped to SNOMED CT concepts. In total, almost 200 million nursing statements from 2,537,310 cases were standardized and converted into OMOP CDM data. They were stored in the observation, procedure_occurrence, and measurement tables, according to their respective mapping domains. Of the hospitalization cases from May 2003 to December 2022, the cohort generated using standardized nursing statements related to nausea consisted of 214,830 cases, whereas the cohort generated using diagnosis and chief complaints consisted of 12,381 cases. CONCLUSION: To the best of our knowledge, this is the first study to convert nursing documentation data into the OMOP CDM format. As a follow-up study, it will be necessary to expand the standardization methods and principles established in this study to other institutions participating in the OMOP CDM project.
Comparison of model Predictive control (MPC) algorithms to optimise blood glucose in fully closed loop (FCL) systems.
Vaughan N, Rashid A
Int J Med Inform · 2026 Jun · PMID 41723939 · Publisher ↗
BACKGROUND AND AIMS: Model Predictive Control (MPC) is emerging within fully closed loop (FCL) systems to offer a promising advancement, by automating glucose regulation for people with Type 1 Diabetes. This article asse...
BACKGROUND AND AIMS: Model Predictive Control (MPC) is emerging within fully closed loop (FCL) systems to offer a promising advancement, by automating glucose regulation for people with Type 1 Diabetes. This article assesses the clinical effectiveness of FCL systems and explores future optimisations by comparison of recent developed systems. METHODS AND RESULTS: Evidence suggests that MPC-based FCL systems outperform hybrid closed-loop (HCL) models using Proportional-Integral-Derivative (PID) control, achieving higher time-in-range (TIR, 74.4% vs. 63.7%, P = 0.020) and better postprandial glucose regulation. However, no system has consistently surpassed the clinical TIR target (>70%), with postprandial hyperglycaemia and insulin absorption delays remaining key challenges. Three recent emerging FCL advancements include nonlinear MPC (NMPC) for dual-hormone systems, integrating glucagon to reduce hypoglycaemia, λ-Policy Iteration (λ-PI), an adaptive reinforcement learning model, and pulse-modulated artificial pancreas (PMCL) systems, which mimic natural insulin secretion. We compare features of these three emerging solutions and propose a novel hybrid model which combines benefits from these algorithms, to improve accuracy. CONCLUSION: While these innovations show promise in in-silico models, clinical validation is lacking. Key barriers include glucagon instability, CGM inaccuracies, cost, and patient adherence. Future research must prioritise long-term trials incorporating real-world factors such as exercise and dietary variability. By integrating predictive control, adaptive learning, and dual-hormone regulation, FCL systems could transform diabetes management, bridging the gap between technology and full automation.
Healthcare practitioner involvement in data-driven clinical decision support development and evaluation: Critical narrative review of recommendations.
Evans RP, Bryant LD, Russell G … +2 more , Wong DC, Absolom K
Int J Med Inform · 2026 Jun · PMID 41723938 · Publisher ↗
OBJECTIVE: While healthcare practitioner (HCP) involvement is widely acknowledged as essential for the development and evaluation of trustworthy data-driven clinical decision support systems (CDSS), practical guidance re...
OBJECTIVE: While healthcare practitioner (HCP) involvement is widely acknowledged as essential for the development and evaluation of trustworthy data-driven clinical decision support systems (CDSS), practical guidance remains limited. This critical narrative review examines existing frameworks, highlights HCP-related considerations, and identifies areas where further guidance is warranted. METHODS: We combined searches of Ovid Medline, the EQUATOR Network Library, and relevant reviews to September 2024, seeking frameworks for developing or evaluating data-driven CDSS. Framework characteristics, coverage across the data-driven CDSS lifecycle, and details of HCP-related recommendations were extracted for analysis. RESULTS: 165 publications were screened, and 32 met inclusion criteria. Nine frameworks made no recommendations relating to HCP involvement. In the other 23, HCP-related recommendations were found for most phases of the data-driven CDSS development and evaluation lifecycle. Recommendations relating to HCP end users included themes of acceptability, communication, and human-AI interaction. Expert clinical input was suggested for various phases, but not required by any reporting guidelines. DISCUSSION: Existing guidance lacks comprehensive methods for including HCPs throughout data-driven CDSS development and evaluation. Reporting guidelines do not position HCPs as experts, which may lead to clinical expertise being overlooked. Frameworks lack detail on complex challenges such as risk communication. No frameworks suggested HCP involvement in data preparation or post-market surveillance, yet HCPs could usefully contribute to these phases. CONCLUSION: HCPs should be included in data-driven CDSS development and evaluation, but there is scope to better understand how to incorporate more clinical insight, and how this might improve trustworthiness of these tools.
Validity of a novel web application for measuring active range of motion and its reliability as a self-measurement method in telerehabilitation.
Aguiló-Furió J, Tronchoni-Crespo B, Moreno-Segura N … +1 more , San Agustín RM
Int J Med Inform · 2026 Jun · PMID 41723937 · Publisher ↗
PURPOSE: As a result of the emergence of Artificial Intelligence (AI), new applications for measuring active range of motion (AROM) in telerehabilitation (TR) are being developed. The main objectives of the present study...
PURPOSE: As a result of the emergence of Artificial Intelligence (AI), new applications for measuring active range of motion (AROM) in telerehabilitation (TR) are being developed. The main objectives of the present study were to evaluate the validity of the TRAK® web application for measuring AROM, to assess its reliability as a self-measurement tool for subjects undergoing TR, and to examine the influence of the subjects' technological skills on the self-measurement process. METHODS: To this end, a cross-sectional observational study was conducted with healthy subjects. Sixty-five volunteer subjects were recruited and divided into two groups based on their technological skills (≤35 years and ≥ 55 years). Thirteen active joint movements were measured by TRAK® in two sessions, with each session being at least one week apart. Validity was assessed in the first session, during which the AROM data obtained by TRAK® were compared with the data obtained from a subsequent video analysis by Kinovea®. In this first session, the physiotherapist supervised the correct execution of the movement. In the second session, the participants repeated the AROM measurements independently and autonomously, following the instructions given by TRAK®, to analyse the reliability of the tool for TR self-measurement. RESULTS: Regarding validity, TRAK® showed good to excellent correlation (ŕs range = 0.739 to 0.987) and root mean square error (RMSE) < 4.74°for eleven out of thirteen movements in the younger group. In the older group, TRAK® obtained good to excellent correlation (ŕs range = 0.703 to 0.928) and RMSE < 4.95°for ten movements. Concerning reliability, however, TRAK® showed SEM percentages above 10% for multiple movements in both populations in its TR modality. CONCLUSION: TRAK® proved to be a valid tool for measuring multiple joint movements regardless of the subject's technological abilities, but was unreliable for assessing AROM in TR.
Digital twin applications in healthcare for people living with disability.
Olawade DB, Almarzook S, Ogunbona MA … +3 more , Makanjuola BD, Olawuyi OF, Wada OZ
Int J Med Inform · 2026 Jun · PMID 41722355 · Publisher ↗
BACKGROUND: Digital twin technology represents a transformative innovation in healthcare, creating virtual replicas of physical entities that enable real-time monitoring, prediction, and personalised intervention. People...
BACKGROUND: Digital twin technology represents a transformative innovation in healthcare, creating virtual replicas of physical entities that enable real-time monitoring, prediction, and personalised intervention. People living with disability face multifaceted healthcare challenges requiring continuous monitoring, adaptive assistive technologies, and individualised treatment approaches. The convergence of digital twin technology with disability healthcare presents unprecedented opportunities for enhancing quality of life, independence, and clinical outcomes. AIM: This review examines the current applications, benefits, challenges, and future directions of digital twin technology in healthcare delivery for people living with disability. METHOD: A narrative review methodology was employed, synthesising literature from academic databases including PubMed, IEEE Xplore, Scopus, and Web of Science. The review encompassed peer reviewed articles, conference proceedings, and technical reports published between 2015 and 2025, focusing on digital twin implementations in disability healthcare contexts. RESULTS: Digital twin applications in disability healthcare span multiple domains, including rehabilitation, assistive device optimisation, cognitive support systems, mobility enhancement, and chronic condition management. The technology demonstrates significant potential in personalising interventions, predicting health deteriorations, optimising assistive technologies, and facilitating remote monitoring. Key applications include virtual prosthetic fitting, wheelchair optimisation, rehabilitation progress tracking, and predictive analytics for secondary complications. However, implementation faces challenges including data privacy concerns, technological accessibility, interoperability issues, and cost barriers. CONCLUSION: Digital twin technology offers transformative potential for disability healthcare, enabling personalised, predictive, and preventive care models. Successful implementation requires addressing technological, ethical, and accessibility challenges whilst ensuring equitable access for diverse disability populations. Critical research priorities include large-scale clinical trials, cost-effectiveness analyses, longitudinal outcomes studies, and ethical frameworks balancing surveillance concerns with care benefits.
GAST-NET: A multi-modal and multi-task deep learning framework for preoperative prediction of perineural invasion and prognostic risk in gastric cancer.
Miao S, Dong H, Feng J … +6 more , Jiang Y, Sun M, Liu Z, Wang Q, Ding X, Wang R
Int J Med Inform · 2026 Jun · PMID 41719850 · Publisher ↗
BACKGROUND: Preoperative imaging prediction of perineural invasion in gastric cancer (GC-PNI) mainly relies on tumour characteristics and clinical variables, while the potential of non-tumour-derived multimodal features...
BACKGROUND: Preoperative imaging prediction of perineural invasion in gastric cancer (GC-PNI) mainly relies on tumour characteristics and clinical variables, while the potential of non-tumour-derived multimodal features remains underexplored. METHOD: We retrospectively enrolled 777 patients from three medical centers and divided them into a training cohort, an internal testing cohort (I-T), and two external testing cohorts (E-T1, E-T2). We developed an end-to-end multimodal and multitask deep learning framework, termed GAST-NET, that integrates tumour CT features, visceral adipose tissue characteristics, and clinical variables to jointly predict perineural invasion (PNI) and five-year prognostic survival risk (PR). The model incorporates an Adaptive Multi-scale Feature Fusion Module (AMFM) and a Cross-Scale Fusion Pooling (CSF Pooling) module to capture hierarchical semantic information and enhance discriminative cross-modal representation. Model performance was assessed using the area under the receiver operating characteristic curve (AUC), and decision curve analysis (DCA). Furthermore, five radiologists were invited to participate in the image reading experiment to verify the clinical interpretability and diagnostic gain of the model. RESULT: The proposed model achieved AUCs of 0.923 (95% CI: 0.865-0.969), 0.868 (95% CI: 0.791-0.934), and 0.871 (95% CI: 0.806-0.930) for PNI prediction across the internal and two external cohorts, respectively. For prognostic risk prediction, the AUC reached 0.873 (95% CI: 0.835-0.922). When used as a decision-support tool, GAST-NET significantly improved diagnostic accuracy and reduced misclassification compared with radiologists. CONCLUSION: GAST-NET demonstrated strong generalizability and potential clinical utility in predicting perineural invasion (PNI) and prognosis in gastric cancer. Notably, visceral adipose tissue features provided complementary value for PNI prediction beyond conventional tumour characteristics.
Management of digital competence sharing in healthcare: Exploring managers' self-assessment and associated factors in a cross-sectional study.
Hammarén M, Pölkki T, Kanste O
Int J Med Inform · 2026 Jun · PMID 41719849 · Publisher ↗
AIM: To describe healthcare managers' self-assessment of management of digital competence sharing and associated background factors. BACKGROUND: Digital competence is essential for the effective use of information and co...
AIM: To describe healthcare managers' self-assessment of management of digital competence sharing and associated background factors. BACKGROUND: Digital competence is essential for the effective use of information and communication technologies (ICT) in healthcare. As digital systems become increasingly embedded in clinical workflows, healthcare managers play a critical role in supporting the development and sharing of digital expertise among healthcare professionals to ensure the successful adoption and utilisation of ICT and new digital solutions. DESIGN: A descriptive cross-sectional study. METHODS: The data were collected from healthcare managers (n = 156) representing five public and one private healthcare organisation in Finland using the Self-assessed Management of Digital Competence Sharing (Sa-MDCS) instrument. The data was analysed using descriptive statistical methods. The study was reported in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) checklist. RESULTS: Healthcare managers generally self-assessed the management of digital competence sharing as good. The highest-rated subscale was the creation of a friendly and safe digital organisational atmosphere, whereas the lowest-rated subscale was the provision of resources and opportunities for digital competence sharing. Significant associations were found between self-assessments and background factors, including age, management experience, digital competence training, organisational context, and management level. CONCLUSION: Although healthcare managers perceive themselves as proficient in promoting digital competence sharing, gaps remain in resource allocation and strategic utilisation of existing digital expertise. Strengthening managerial support is essential for enhancing organisational readiness and the effective integration of ICT into clinical practice. To support this, healthcare organisations should invest in targeted leadership training and ensure adequate resources. Fostering a culture of continuous digital learning and sharing can enhance the adoption of ICT and new digital solutions, streamline clinical workflows, and ultimately improve patient care outcomes.
Predicting neoadjuvant immunotherapy efficacy with machine learning models in non-small cell lung cancer: A systematic review and meta analysis.
Liu W, Feng Z, Zhang M … +2 more , Mao R, Li J
Int J Med Inform · 2026 Jun · PMID 41719848 · Publisher ↗
BACKGROUND: The response of resectable non-small cell lung cancer (NSCLC) to neoadjuvant immunotherapy is heterogeneous. Machine learning can integrate multimodal data to construct predictive models, but the methodologic...
BACKGROUND: The response of resectable non-small cell lung cancer (NSCLC) to neoadjuvant immunotherapy is heterogeneous. Machine learning can integrate multimodal data to construct predictive models, but the methodological quality, risk of bias and clinical applicability of such models have not been systematically evaluated. OBJECTIVE: This study aims to systematically evaluate the methodological quality, risk of bias, and diagnostic performance of machine learning models for predicting neoadjuvant immunotherapy response in resectable NSCLC. METHODS: As of August 22, 2025, 11 databases were retrieved. Two researchers independently extracted the data, and a third researcher resolved the data differences. The quality of the model, the development process and the quality of radiomics reports were evaluated respectively by probast + AI, IJMEDI checklist and RQS. Meta-analysis of the AUC, sensitivity and specificity of the model was conducted using R software, and subgroup analysis was performed according to predictors, algorithms and outcomes. RESULTS: Seventeen studies involving 44 models were included. Eighty-nine percent of models had relatively low quality and all had a high risk of bias - key flaws included unreasonable sample size, improper handling of missing data and defects in validation procedures - but the overall applicability was good. IJMEDI scores ranged 26.5-37 (4 high-quality, others medium); average RQS of 12 radiomics studies was 14.58 (22.22%-52.78%), with multiple deficiencies. Ten internal validation models showed that the combined internal AUC was 0.786 (95% CI: 0.740-0.826, I2 = 0%), there was no publication bias (Egger's test), and the sensitivity was 0.763 (95% CI: (0.56-0.89), with a specificity of 0.908 (95% CI: 0.471-0.991). The predicted AUCs of MPR and PCR were 0.805 and 0.761, respectively. SVM achieved the highest AUC (0.841), and the non-radiomics model (0.869) was superior to the radiomics model (0.775). The combined external validation AUC was 0.760, among which the AUC predicted by MPR was 0.754. CONCLUSION: ML models show potential for predicting neoadjuvant immunotherapy efficacy in resectable NSCLC, with SVM and non-radiomics models superior. However, low methodological quality and high bias risk require cautious interpretation. Future work should refine methodology, address radiomics gaps, and promote clinical translation.
Improvements in EHR usability during eleven years: Cross-sectional surveys on specialist physicians' experiences in Finland.
Viitanen J, Nisula S, Tuovinen T … +1 more , Lääveri T
Int J Med Inform · 2026 Jun · PMID 41719847 · Publisher ↗
BACKGROUND: Usability not only impacts user satisfaction but also the efficiency and effectiveness of electronic health record (EHR) use. Moreover, poor EHR usability is associated with physician burnout and stress. Desp...
BACKGROUND: Usability not only impacts user satisfaction but also the efficiency and effectiveness of electronic health record (EHR) use. Moreover, poor EHR usability is associated with physician burnout and stress. Despite an abundance of research and investments in improving usability, long-term monitoring studies are scarce. Our aim was to explore how the usability of EHR systems has evolved over an 11-year period in Finland from specialist physicians' perspectives and to compare user experiences (UXs) between specialties. METHODS: Nationwide usability-focused cross-sectional surveys were conducted among Finnish physicians in 2010, 2014, 2017, and 2021. To measure usability, we selected six usability-related statements from the validated National Usability-focused Health Information System Scale (NuHISS). The responses of specialist physicians with more than six months of system use (n = 2473 in 2021; 2137 in 2017; 2162 in 2014; and 2312 in 2010) were selected and analyzed by specialty group and study year. RESULTS: During the study years, UXs improved across all specialty groups. A positive change was seen particularly in occupational healthcare (OH) and nonsurgical medical specialties. However, the Net EHR Experience Scores (NEESs) remained negative for all groups, except for physicians working in anesthesiology and intensive care. In 2021, physicians working in OH and general practice were more positive about EHR usability, whereas those in psychiatric specialties gave more negative assessments than other groups. CONCLUSION: Despite improvements over the study years, most physicians remained dissatisfied with EHRs. Interestingly, the positive development was not linear; during 2014-17, NEESs decreased, coinciding with the implementation of the national health information exchange services. Our findings suggest that efforts to improve EHR usability and enhance end-user skills only slowly translate into better UXs. To distinguish the effect of factors such as national requirements or recently implemented EHR brands, continuous monitoring of UXs using context-adapted surveys like NuHISS is crucial.
Agentic memory-augmented retrieval and evidence grounding for medical question-answering tasks.
Jia S, Bit S, Jasodanand VH … +2 more , Liu Y, Kolachalama VB
Int J Med Inform · 2026 Jun · PMID 41713127 · Full text
OBJECTIVE: To evaluate whether a tool-using agent-based system built on large language models (LLMs) outperforms standalone LLMs on medical question-answering tasks. METHODS: We developed a unified, open-source LLM-based...
OBJECTIVE: To evaluate whether a tool-using agent-based system built on large language models (LLMs) outperforms standalone LLMs on medical question-answering tasks. METHODS: We developed a unified, open-source LLM-based agentic system that integrates document retrieval, reranking, evidence grounding, and diagnosis generation to support dynamic, multi-step medical reasoning. Our system features a lightweight retrieval-augmented generation pipeline for efficient evidence retrieval and reranking, coupled with a cache-and-prune memory bank, enabling efficient long-context inference beyond standard LLM limits. The system autonomously invokes specialized tools, eliminating the need for manual prompt engineering or brittle multi-stage templates. We compared the agentic system against standalone language models on various medical question-answering benchmarks. RESULTS: Evaluated on five well-known benchmarks, our system outperforms or closely matches state-of-the-art proprietary and open-source models in multiple-choice and open-ended formats. Specifically, it achieved accuracies of 82.98% on the United States Medical Licensing Examination (USMLE) Step 1 and 86.24% on Step 2, surpassing GPT-4's 80.67% and 81.67%, respectively, while closely matching on Step 3 (88.52% vs. 89.78%). CONCLUSION: Our findings highlight the value of combining tool-augmented and evidence-grounded reasoning strategies to build reliable and scalable medical artificial intelligence systems.
Corrigendum to "Advantages and challenges of tracking st-segment elevation myocardial infarction patients with a real-time dashboard: A single-centre experience" [Int. J. Med. Inform. 209 (2026) 106261].
de Mattia E, Paoletti F, Pedicino D … +11 more , Angioletti C, Perilli A, d'Aiello A, Donno R, Adduci A, Arcuri G, Meneschincheri E, Ruffo B, D'Agostino M, Liuzzo G, de Belvis AG
Int J Med Inform · 2026 May · PMID 41708405 · Publisher ↗
Abstract loading — click title to view on PubMed.
About
- Frequency
- Sun
- Papers found
- 200
- RSS feed
- Subscribe