International Journal Of Medical Informatics[JOURNAL]
Sun
200 papers
The role of artificial intelligence in virtual emergency care: a systematic review.
Shankar R, Wang L, Hoe HS … +5 more , Yee IL, Fong LM, Kumar Gollamudi SP, Wong TC, Wong S
Int J Med Inform · 2026 Jul · PMID 41880917 · Publisher ↗
BACKGROUND: The integration of artificial intelligence (AI) into virtual emergency care represents a potentially transformative approach to healthcare delivery, yet the evidence base remains poorly characterized. This sy...
BACKGROUND: The integration of artificial intelligence (AI) into virtual emergency care represents a potentially transformative approach to healthcare delivery, yet the evidence base remains poorly characterized. This systematic review comprehensively evaluates the current state of AI applications in virtual emergency care settings. METHODS: We systematically searched eight databases (Embase, PsycINFO, MEDLINE, PubMed, Scopus, Web of Science, CINAHL, Cochrane Library) from inception through March 2025. Of 7,098 records identified and 4,935 screened after deduplication using Covidence, 8 studies met inclusion criteria following exclusion of one study lacking AI components. Studies were assessed using PROBAST + AI for risk of bias and quality assessment, TRIPOD + AI for reporting quality, and GRADE for certainty of evidence. RESULTS: The eight included studies (total participants: approximately 0.5 million) evaluated diverse AI applications including decision trees, machine learning ensembles, and graph neural networks across multiple virtual emergency contexts. Performance varied widely (accuracy 77.5-100%, sensitivity 63-100%, specificity 60% in single study reporting). All clinical studies demonstrated serious risk of bias. TRIPOD + AI compliance averaged only 36.9% (range 30.9-48.1%). GRADE assessment revealed very low to low certainty evidence across all outcomes, with no studies measuring actual clinical outcomes. CONCLUSIONS: Current evidence is insufficient to support widespread clinical implementation of AI in virtual emergency care. While preliminary results suggest potential benefits in triage accuracy and resource efficiency, critical gaps exist in validation, clinical outcome assessment, and reporting standards. Future research must prioritize prospective controlled trials with real patient data, clinical outcome measurements, and adherence to reporting guidelines.
Beyond the conventional: Artificial intelligence in identifying risk factors in sports injuries. A scoping review.
Mora JSM, Medina RAB, Molina VM … +1 more , Hernandez Rincon EH
Int J Med Inform · 2026 Jul · PMID 41880916 · Publisher ↗
PURPOSE: To map contemporary uses of artificial intelligence (AI) to identify conventional and unconventional risk factors for sports injuries in athletes. METHODS: Following Joanna Briggs Institute guidance and PRISMA-S...
PURPOSE: To map contemporary uses of artificial intelligence (AI) to identify conventional and unconventional risk factors for sports injuries in athletes. METHODS: Following Joanna Briggs Institute guidance and PRISMA-ScR, we conducted an open-access search (January 2015-September 2025) in PubMed/MEDLINE, Scopus, and ScienceDirect using MeSH/DeCS terms for AI and injury risk. Records were managed in Rayyan with de-duplication, independent triple screening, standardized data extraction, and narrative synthesis to address heterogeneity. RESULTS: Fifty-nine studies met inclusion criteria. Frequently modelled predictors included sleep quality and psychological state, external and internal load metrics, and prior injury. AI approaches spanned classical machine learning and deep learning applied to multimodal sensor and clinical data. Across studies, generalizability was limited by heterogeneous populations, outcomes, and measurement/reporting practices; few works reported consistent measurement standards or external validation. Where assessed, observer agreement and classification performance were acceptable but variable. CONCLUSION: AI for sports-injury risk is expanding rapidly, led by classical machine learning on multimodal sensor data. Key gaps are external validity and reproducibility. Progress will require multicenter prospective cohorts, standardized measurement and reporting, sport-specific meta-analyses, transparent model sharing, and deliberate clinical integration into athlete care pathways.
Obstacle and the medium: a leadership compass for digital health strategy.
Moura LA
Int J Med Inform · 2026 Jun · PMID 41876300 · Publisher ↗
This editorial draws on two intellectual traditions - Stoic philosophy and media theory - to propose a leadership framework for digital health strategy. From Marcus Aurelius comes the principle that obstacles are not int...
This editorial draws on two intellectual traditions - Stoic philosophy and media theory - to propose a leadership framework for digital health strategy. From Marcus Aurelius comes the principle that obstacles are not interruptions to action but the terrain through which action unfolds. From Marshall McLuhan comes the insight that digital health creates environments that shape clinical behaviour, resource allocation, and organisational imagination - not neutral containers for data. The editorial distinguishes two levels of leadership responsibility. At the personal level, the leader studies and navigates obstacles. At the systemic level, the leader documents what is learned, transforms governance and policy where authority exists, and builds capacity that outlasts any single tenure. Navigation is necessary; transformation is the mandate. Digital health obstacles are increasingly structural: fragmented governance, misaligned funding models, unenforced interoperability standards, workforce resistance, and evaluation frameworks that measure inputs rather than impact. These function as architectures - active forces that determine what becomes possible. The Stoic leader treats them as conditions to transform, not constraints to circumvent. Applying McLuhan, the editorial argues that digital health strategy must shift from designing what systems *do* to designing what systems *enable*. The architecture is the strategy. Workflows, governance structures, and information systems shape behaviour before any content reaches users. The editorial closes with a practical framework - study, navigate, document, transform, build capacity - and a generational standard: leaders will be judged not by how well they navigated their obstacles, but by how many they removed for those who follow.
Enhancing hospital drug and medical supply request processes through a clinically prioritized and demand-variability-aware adaptive decision support framework.
İşli D, Aydın L
Int J Med Inform · 2026 Jul · PMID 41875651 · Publisher ↗
BACKGROUND: Rule-based replenishment thresholds in hospital information systems (HIS) are feasible and auditable but may respond slowly to demand variability and clinical criticality. OBJECTIVE: To evaluate a decision-su...
BACKGROUND: Rule-based replenishment thresholds in hospital information systems (HIS) are feasible and auditable but may respond slowly to demand variability and clinical criticality. OBJECTIVE: To evaluate a decision-support framework that preserves a three-level request structure (minimum/critical/maximum) while adapting thresholds using short-horizon demand forecasts and uncertainty, stratified by VEN clinical priority (Vital-Essential-Non-essential), and to benchmark its forecasting component against a naïve rolling 180-day mean predictor (μ180). METHODS: A retrospective unit-item-day panel was constructed from routine hospital records and classified by VEN. XGBoost forecasted 14- and 30-day cumulative demand; uncertainty was modeled using a rolling 30-day consumption standard deviation scaled by VEN-specific coefficients. Adaptive targets were mapped to existing HIS request levels and backtested against the baseline rule (7×/15×/30 × the rolling 180-day mean daily consumption) on a 30-day ML-available panel (35,801 unit-item-day records; 1269 series; 2025-09-30-2026-01-22). Forecast accuracy was compared with μ180 using RMSE and WAPE, supported by calibration and Diebold-Mariano testing. Policies were evaluated using paired Wilcoxon tests and lead-time sensitivity analyses. RESULTS: The proposed policy reduced stock-out days (54 to 10; -81.5%), days below critical (11,612 to 6396; -44.9%), ordering days (1936 to 550; -71.6%), and ordered quantity (853,119 to 579,795; p < 0.001). Stock-outs were eliminated for Vital items. XGBoost reduced WAPE by 62%-71% versus μ180 across horizons, with absolute-error superiority confirmed by Diebold-Mariano testing. Inventory exposure increased (366,406 to 802,049), indicating a measurable safety-inventory trade-off. CONCLUSIONS: VEN-prioritized, forecast- and volatility-aware threshold adaptation improved service continuity and ordering efficiency within an auditable HIS workflow while quantifying trade-offs under alternative lead-time assumptions.
Characterizing nursing home care team communication via text messaging: A social network analysis.
Powell KR, Farmer MS, Popescu M … +2 more , Mehr DR, Alexander GL
Int J Med Inform · 2026 Jul · PMID 41865476 · Publisher ↗
PURPOSE: Communication breakdowns among healthcare teams contribute substantially to preventable adverse events in nursing homes. Although text messaging platforms are increasingly used to support care coordination, litt...
PURPOSE: Communication breakdowns among healthcare teams contribute substantially to preventable adverse events in nursing homes. Although text messaging platforms are increasingly used to support care coordination, little is known about how these technologies are embedded within care team communication networks. The purpose of this study was to characterize patterns of text-message communication among nursing home care teams and to examine how network structures reflect underlying social system dynamics. METHODS: Social network analysis was applied to text messages (n = 5,092) linked to (n = 585) nursing home-to-hospital resident transfers over 5 years (2015-2020). Message metadata and content were used to construct communication networks and to calculate network measures, including density, centralization (in-degree and out-degree), and reciprocity. Networks were qualitatively classified into communication models based on shared structural characteristics and interpreted using Rogers' Diffusion of Innovations theory. All analysis and visualization was conducted using Python (3.13.9). RESULTS: Three distinct communication models were identified. The Integrated model exhibited high message volume, high connectivity, and low hierarchy, consistent with mature adoption of text messaging as a general collaboration tool. The Hub-and-Spoke model showed moderate message volume with centralized information flow, reflecting protocol-driven, hierarchical communication. The Siloed model demonstrated high message volume but low role diversity, indicating niche use within specific professional roles rather than facility-wide coordination. Across models, individuals occupying central or bridging positions appeared to function as informal opinion leaders, influencing communication through frequent interaction rather than formal authority. CONCLUSIONS: The structure and quality of communication networks shape how information flows, influence is exercised through electronic messaging, and innovations are integrated into care processes. Social network analysis offers a rigorous approach for evaluating implementation and guiding strategies to support effective, team-based communication in nursing home settings. Future work should examine whether increasingly integrated communication networks are associated with improved resident outcomes.
Comparing large language models and human experts in interpreting MRI reports for personalized patient education.
Du K, Li A, Zuo QH … +6 more , Zhang CY, Guo R, Chen P, Du WS, Zuo YL, Li SM
Int J Med Inform · 2026 Jul · PMID 41865475 · Publisher ↗
RATIONALE AND OBJECTIVES: Knee osteoarthritis (OA) is a prevalent global condition. While MRI guides clinical decisions, its technical complexity hinders patient understanding and engagement. Translating these findings i...
RATIONALE AND OBJECTIVES: Knee osteoarthritis (OA) is a prevalent global condition. While MRI guides clinical decisions, its technical complexity hinders patient understanding and engagement. Translating these findings into comprehensible, personalized patient education remains challenging. Large language models (LLMs) show promise in automating this process. To evaluate and compare the effectiveness of advanced large language models against experienced clinicians in generating comprehensible, personalized patient education materials derived from knee MRI reports. MATERIALS AND METHODS: This study compared performance of two LLMs, GPT-4o and Claude 3.5 Sonnet, with experienced clinicians in generating personalized patient education materials from 150 anonymized knee MRI reports. To assess their effectiveness, we developed a comprehensive, multidimensional evaluation framework. This included readability evaluation, using both validated linguistic metrics and expert assessments of clarity, emphasis, and coherence; content personalization, quantified with a novel structured scoring system focused on specificity, practicality, and actionability of recommendations; and generation efficiency, measured in words per minute. RESULTS: Both LLMs significantly outperformed clinicians across key metrics, with GPT-4o showing superior performance. Compared to clinicians, GPT-4o and Claude 3.5 Sonnet demonstrated higher expert-rated understandability (72[IQR 6] vs 60[IQR 6] vs 50[IQR 12], P < 0.001), better personalization scores (68[IQR 2] vs 62[IQR 4] vs 64[IQR 9], P < 0.001), and markedly higher generation efficiency (1348.5 ± 202.2 vs 1160.8 ± 137.2 vs 142.6 ± 29.8 WPM, P < 0.001). Readability indices consistently favored LLM-generated content. CONCLUSIONS: Advanced LLMs, particularly GPT-4o, showed strong performance in translating knee MRI reports into comprehensible and personalized patient education materials, with advantages in readability, personalization, and efficiency over clinician-generated outputs in this study setting. These findings support the potential role of LLMs as clinician-supervised tools for scalable patient education, while highlighting the need for further validation across institutions, models, and clinical workflows before deployment.
Research using smartwatches for the measurement of physical activity, sedentary behavior, and sleep: A scoping review.
da Silva JF, Germano-Soares AH, de Oliveira TV … +4 more , Barbosa Silva LC, Barbosa Filho VC, de Aguiar Silva TC, Tassitano RM
Int J Med Inform · 2026 Jul · PMID 41864152 · Publisher ↗
OBJECTIVE: The objective of this scoping review is to map the evidence on using smartwatches to objectively measure physical activity (PA), sedentary behavior (SB), and sleep in children, adolescents, adults, and older a...
OBJECTIVE: The objective of this scoping review is to map the evidence on using smartwatches to objectively measure physical activity (PA), sedentary behavior (SB), and sleep in children, adolescents, adults, and older adults. METHODS: This review followed the Joanna Briggs Institute (JBI) guidelines for scoping reviews, following a previously published protocol. The searches were conducted in January 2022 on Medline, Scopus, Web of Science, IEEE Xplore Digital Library, Scielo, LILACS, Health Technology Assessment Database, Cochrane clinical trials, and clinical trials. The screening was performed independently by two authors. A narrative synthesis was used. RESULTS: After the electronic search, 5,925 records were identified and 2,008 duplicates removed. Screening of 3,917 titles and abstracts resulted in 491 full-text articles assessed for eligibility. A total of 427 studies were excluded for not meeting inclusion criteria, with one study added after author contact. Thus, 64 of these studies were included. Sample sizes ranged widely from 4 to 6,454 participants, with most studies (80.9%) including 100 or fewer participants. Most studies were conducted in the United States (46.9%), followed by China (9.4%). A total of 18 different smartwatch brands were identified, with Apple® being the most investigated (46.9%), followed by Samsung® (15.6), Fitbit® (14.1%), Motorola® (10.9%), Polar® (10.9%), Garmin® (9.4%), Asus® (3.1%), and others (1.6% each. Smartwatches were primarily used to validate measurements of PA, SB, and/or sleep parameters, followed by their use in assessing these variables as exposures or outcomes, testing feasibility, supporting self-monitoring, and other purposes. The most frequently measured variables were steps, total sleep time, and SB. CONCLUSION: This review can help studies, interventions, and health professionals that aim to use this technology as a measurement instrument, as well as help end users who use it on a daily basis.
Automated Flow and local LLM-Driven clinical Context Engineering: Precision colorectal cancer recurrence registry.
Yang YW, Chang CY, Lin YZ … +10 more , Cheng HH, Huang SC, Lin HH, Lin CC, Lan YT, Wang HS, Chang SC, Yang SH, Chen WS, Jiang JK
Int J Med Inform · 2026 Jun · PMID 41849918 · Publisher ↗
OBJECTIVE: This study develops and validates a deployable, privacy-preserving automated tool to address the labor-intensive nature of colorectal cancer (CRC) recurrence registration. Our tool utilizes a reproducible, two...
OBJECTIVE: This study develops and validates a deployable, privacy-preserving automated tool to address the labor-intensive nature of colorectal cancer (CRC) recurrence registration. Our tool utilizes a reproducible, two-stage 'Clinical Context Engineering' workflow to mimic expert clinical reasoning and overcome the limitations of handling longitudinal clinical data ambiguity. METHODS: We retrospectively studied 3053 CRC patients (2010-2018). A local Large Language Model (LLM) (Qwen3:14B) analyzed ∼ 19,900 pathology and ∼ 43,900 imaging reports using iterative, patch-based analysis ("raw LLM"). Clinical validation rules were applied to generate "rule-based LLM" output, enhancing explainability and trustworthiness. Both automated methods and the manual Taiwan Cancer Registry (TCR) database were compared against a 20% manual reference standard (N = 602). Full prompts and validation code are provided for complete reproducibility. RESULTS: Under a strict 60-day temporal tolerance, the Rule-Based LLM achieved 90.7% accuracy, comparable to standard TCR processes (92.0%), and 77.2% sensitivity. The application of clinical validation rules significantly improved specificity from 87.7% (Raw LLM) to 93.9% (Rule-Based LLM). In time window analysis, the Rule-Based LLM identified 87.1% of recurrences within 60 days of the reference date. CONCLUSION: Our locally deployed, privacy-preserving, and explainable Clinical Context Engineering framework offers a viable, non-inferior alternative to standard TCR processes, reducing workload while maintaining data quality and fostering trust in AI-assisted cancer registry automation.
Considerations for enhancing the clinical translational potential of LLM-Based TBI mortality prediction models.
Wang L, Zhao S
Int J Med Inform · 2026 Jun · PMID 41846057 · Publisher ↗
This study explores the use of GPT-5 and traditional machine learning by scholars such as Tu et al. to predict the risk of emergency death in traumatic brain injury (TBI), and affirms the value of their experimental desi...
This study explores the use of GPT-5 and traditional machine learning by scholars such as Tu et al. to predict the risk of emergency death in traumatic brain injury (TBI), and affirms the value of their experimental design and method exploration in promoting AI assisted TBI triage. At the same time, it is pointed out that there are still four shortcomings in this study: a lack of clinical experts to validate LLM output, no exploration of subgroup dynamic threshold adaptation, no evaluation of model inference delay, and no sensitivity analysis of prompt words. The study suggests that further research is needed to validate the clinical effectiveness of LLM, optimize the practicality and reproducibility of the model, and further enhance the clinical translational potential and methodological rigor of the LLM based TBI prediction model.
Evaluation of a user-oriented inter-physician exchange portal within a decision support system for primary care - a qualitative study.
Neff MC, Schulz MR, Storf H … +1 more , Schaaf J
Int J Med Inform · 2026 Jun · PMID 41846056 · Publisher ↗
BACKGROUND: General practitioners are often the first point of contact for patients with rare or unclear conditions. While clinical decision support systems (CDSS) based on artificial intelligence can provide valuable as...
BACKGROUND: General practitioners are often the first point of contact for patients with rare or unclear conditions. While clinical decision support systems (CDSS) based on artificial intelligence can provide valuable assistance, effective communication between physicians remains crucial. Therefore, integrating digital tools to facilitate communication between physicians is considered an important additional pillar of a CDSS for primary care. OBJECTIVES: This study examined the design of a portal to facilitate collaboration between physicians within a CDSS for primary care. The objective was to identify usability issues in the inter-physician exchange portal at an early stage to optimise the design of future portals within CDSSs. METHODS: A first prototype of the inter-physician exchange portal was evaluated and further developed in two stages. Heuristic and user evaluations were conducted to optimise usability and better understand challenges. RESULTS: The design process resulted in a first interactive prototype of an inter-physician exchange portal. The heuristic evaluation identified 36 usability issues, which informed subsequent refinements to the prototype. User evaluations indicated positive overall perceptions of usability, along with specific recommendations for enhancing the display of status information. The portal achieved an average System Usability Scale score of 81.5, which corresponds to an 'excellent' rating. CONCLUSION: Developing an inter-physician exchange portal within a CDSS for primary care facilitates the exchange of information between physicians and can thus improve the care for patients with rare and unclear diseases. General practitioners' motivation to engage with the portal is influenced by several factors, notably the existence of an active user community. Future developments should address usability issues and evaluate l acceptance levels, impact on patient care, and long-term user retention in a real-world clinical setting. The integration of existing solutions could also be considered.
Unsupervised identification of muscle phenotypes in adults with obesity: a data-driven framework for the identification of sarcopenia in absence of a gold standard.
Sofia LC, Annalisa B, Gabriele C … +3 more , Paolo B, Maria Grazia C, Alessandra C
Int J Med Inform · 2026 Jun · PMID 41846055 · Publisher ↗
BACKGROUND: Sarcopenia is characterized by progressive loss of skeletal muscle mass and strength and is associated with increased disability and mortality. However, the diagnosis of sarcopenia remains challenging due to...
BACKGROUND: Sarcopenia is characterized by progressive loss of skeletal muscle mass and strength and is associated with increased disability and mortality. However, the diagnosis of sarcopenia remains challenging due to the absence of a universally accepted gold standard and validated cut-off values for skeletal muscle indices. Data-driven approaches based on unsupervised clustering may overcome these limitations by identifying muscle-related phenotypes directly from anthropometric and body composition data. METHODS: In this study, 600 adults with obesity were analyzed and stratified by sex. The dataset was randomly divided into a training set (80%) and a testing set (20%). After data standardization, principal component analysis (PCA) was applied separately in males and females. Unsupervised clustering was then performed on the preserved principal components, and the optimal number of clusters was determined using internal validation indices. Linear Discriminant Analysis (LDA) was applied to assign patients in the test set, and posterior probabilities were correlated with Skeletal Muscle Index (SMI). RESULTS: Clustering consistently identified two distinct groups in both sexes: one with higher SMI and another with lower SMI, consistent with reduced muscle status. Stepwise LDA accurately classified individuals, and posterior probabilities of belonging to the pathological cluster were negatively correlated with SMI in both sexes, despite SMI not being used in clustering or classification. Individuals in the pathological group exhibited significantly lower SMI, particularly among females. CONCLUSIONS: The combined use of unsupervised clustering and LDA allows reliable identification of distinct muscle-related phenotypes in adults with obesity. This framework provides reproducible classifications, correlates with skeletal muscle index, and offers a quantitative approach to stratify patients by muscle status, even in the absence of predefined diagnostic criteria. These findings support the potential of data-driven phenotyping to improve early detection of sarcopenic obesity.
Study of bladder cancer detection in standard white light versus AI-supported endoscopy-01 (RAISE-01) - Development and validation of an AI-based support tool.
Hjort PB, Jensen JE, Jensen JB … +1 more , Ernst A
Int J Med Inform · 2026 Jun · PMID 41833082 · Publisher ↗
BACKGROUND: Diagnosis and surveillance of bladder cancer rely on white-light cystoscopy (WLC). However, this modality is operator-dependent and associated with a risk of missed lesions, contributing to high recurrence ra...
BACKGROUND: Diagnosis and surveillance of bladder cancer rely on white-light cystoscopy (WLC). However, this modality is operator-dependent and associated with a risk of missed lesions, contributing to high recurrence rates, especially in non-muscle invasive bladder cancer. Recent advances in artificial intelligence (AI) enable software-based decision support for bladder lesion detection, with potential for vendor-independent deployment and broad integration into routine clinical workflows. OBJECTIVE: To develop and externally validate an AI-based clinical decision support system for real-time bladder lesion detection during cystoscopy. METHODS: CystoAID, a convolutional neural network-based object detection system, was trained on prospectively collected video recordings from flexible cystoscopies and transurethral resections of bladder tumors. Diagnostic accuracy was evaluated using a retrospective external validation dataset representative of routine clinical practice, in accordance with STARD-AI recommendations. RESULTS: In the external validation cohort, CystoAID achieved a sensitivity of 1.00 (95% CI 0.95-1.00). Precision was 88.1% (95% CI 81.3-92.7), exceeding published estimates for WLC. Precision-recall analysis showed consistently high precision (>0.8) across clinically relevant recall levels, with declining precision at higher recall, reflecting the expected trade-off between sensitivity and false-positive detections. The system operated with low processing latency, supporting feasibility for real-time clinical use. Sensitivity was prioritized to mitigate the clinical risk associated with false-negative findings. CONCLUSIONS: CystoAID is a real-time, AI-based decision support tool for cystoscopy that demonstrated high sensitivity and favorable precision in external validation. These findings support its potential role as an assistive technology in routine urologic practice. Prospective studies are warranted to evaluate clinical impact, workflow integration, and performance in detecting challenging lesion subtypes, including flat lesions and carcinoma in situ.
"Clicking without understanding": A mixed-methods analysis of user agreements in digital mental health services.
Siew Keong GC
Int J Med Inform · 2026 Jun · PMID 41833081 · Publisher ↗
OBJECTIVES: This study evaluates privacy policies and terms of service agreements from digital mental health platforms, focusing on accessibility, comprehensibility, and alignment with informed consent principles in heal...
OBJECTIVES: This study evaluates privacy policies and terms of service agreements from digital mental health platforms, focusing on accessibility, comprehensibility, and alignment with informed consent principles in healthcare informatics. MATERIALS AND METHODS: We applied mixed methods combining content analysis and computational linguistic assessment to 139 user agreements from international mental health applications and Singaporean providers, including commercial platforms and social service agencies serving vulnerable populations. We evaluated readability, communicative practices, regulatory compliance, and power asymmetries. Only 1.67% of services implemented comprehension verification for informed consent. User agreements required approximately 16 years of education for comprehension and exhibited significant linguistic power asymmetries favoring providers. Privacy policies comprehensively addressed data collection but systematically neglected post-service communication regarding data retention and deletion. Among local services, only 8.33% adequately communicated data breach notification procedures as required by Singapore's Personal Data Protection Act. Terms of service failed to establish bidirectional communicative exchange necessary for meaningful healthcare informed consent. Findings reveal fundamental misalignment between digital mental health agreements and collaborative communication principles essential to therapeutic relationships and healthcare informatics best practices. Communication barriers pose particular risks for individuals with serious mental illness requiring accessible health information for decision-making. Results have implications for health informatics policy, consumer health technology design, and digital health regulatory frameworks. Digital mental health platforms demonstrate significant user communication deficiencies. Our findings point to the need for user agreements that are written in plain language, that incorporate essential informed consent components, that balance linguistic power between providers and users, and that accommodate the cognitive needs of vulnerable populations seeking mental health support.
Can AI safely choose antibiotics over the knife? A STROBE-guided benchmark of GPT-4, GPT-5, and Gemini for non-operative acute appendicitis management.
Calışkan YK, Başak F, Erdem O
Int J Med Inform · 2026 Jun · PMID 41831386 · Publisher ↗
BACKGROUND: Non-operative management (NOM) with antibiotics is increasingly used for imaging-confirmed uncomplicated acute appendicitis (AA), but appropriate selection remains clinically delicate-especially in "gray-zone...
BACKGROUND: Non-operative management (NOM) with antibiotics is increasingly used for imaging-confirmed uncomplicated acute appendicitis (AA), but appropriate selection remains clinically delicate-especially in "gray-zone" scenarios such as appendicolith, older age, or borderline imaging findings. As large language models (LLMs) are increasingly queried for clinical guidance by both patients and clinicians, their reliability in distinguishing NOM candidates from patients requiring urgent operative care warrants formal evaluation. This question is clinically relevant not only for academic benchmarking but also because generative AI tools are already being used in health-information seeking and decision-support contexts, where overconfident but unsafe triage advice could influence real-world care pathways. METHODS: We conducted a cross-sectional, in-silico benchmarking study using 50 standardized clinical vignettes spanning uncomplicated AA (n = 20), complicated AA (n = 20), and atypical/high-risk presentations (n = 10; older age, pregnancy, immunocompromise). Three LLMs (GPT-4, GPT-5, Gemini) were queried with a uniform zero-shot prompt requesting NOM candidacy determination and structured risk-benefit communication. Two blinded surgeons scored outputs against predefined criteria anchored to international guideline principles. The primary outcome was Management Accuracy Score (MAS; correct/incorrect). Secondary outcomes included risk-stratification nuance, safety warnings, and evidence-use quality (Likert 1-5). Fleiss' kappa quantified inter-model agreement. The vignette set was intentionally balanced to stress-test models across routine and safety-critical scenarios rather than to estimate population prevalence. RESULTS: Overall MAS differed across models (χ2 = 6.34, p = 0.042): GPT-5 92% (46/50), GPT-4 84% (42/50), and Gemini 76% (38/50). The widest gap occurred in complicated AA (p = 0.028), driven by Gemini's over-recommendation of antibiotic "trials" in appendicolith-positive or abscess-suggestive scenarios. GPT-5 generated the most consistent recurrence counseling and safety framing; no critical safety failures were observed. All three models performed less reliably in atypical/high-risk vignettes; however, because this subgroup contained only 10 cases, these findings should be interpreted cautiously and not over-read as evidence of true between-model equivalence or superiority. CONCLUSION: LLMs demonstrate strong baseline awareness of contemporary AA strategies, yet clinically meaningful variability persists-particularly for contraindications to NOM. GPT-5 performed best overall, while Gemini's over-generalization in high-risk contexts highlights the need for domain-constrained training and guardrails before clinical integration. These findings support supervised, educational use at most, rather than autonomous emergency triage deployment.
Corrigendum to 'Success and failure of human-AI collaboration in clinical reasoning: An experimental study on challenging real-world cases' [Int. J. Med. Inf. 211 (2026) 106342].
Ong KT, Seo J, Kim H … +5 more , Kim J, Kim J, Kim S, Yeo J, Choi EY
Int J Med Inform · 2026 Jun · PMID 41826173 · Publisher ↗
Abstract loading — click title to view on PubMed.
Inclusive safeguards for predictive resilience models in medical education.
Umasugi MT
Int J Med Inform · 2026 Jun · PMID 41825337 · Publisher ↗
Abstract loading — click title to view on PubMed.
A mixed reality framework for interpretable and explainable joint replacement assessment.
Ulrich L, Innocente C, Marullo G … +5 more , Audisio A, Aprato A, Massè A, Moos S, Vezzetti E
Int J Med Inform · 2026 Jun · PMID 41812464 · Publisher ↗
OBJECTIVE: Joint replacement surgery, also known as arthroplasty, is a common procedure that restores mobility and relieves pain in patients with severe joint pathologies. Despite being considered routine, arthroplasties...
OBJECTIVE: Joint replacement surgery, also known as arthroplasty, is a common procedure that restores mobility and relieves pain in patients with severe joint pathologies. Despite being considered routine, arthroplasties are complex interventions with potential complications and variable clinical outcomes. Accurate evaluation of replaced joint mobility to ensure implant stability within the patient's functional range of motion (ROM) is a major challenge in postoperative care. However, the reliability of current assessment methods is limited due to their lack of standardized and quantitative tools. This study presents a patient-specific Mixed Reality (MR) framework designed to enhance postoperative evaluation in joint replacement with a focus on total hip arthroplasty (THA). METHODS: The proposed system enables objective quantification and MR visualization of prosthesis biomechanics by integrating ROM simulation and 3D modeling, promoting explainability and interpretability of surgery outcomes. A retrospective analysis of 67 THAs was performed to compare simulated ROM results with clinical assessments and literature benchmarks. Additionally, surgeons evaluated the system's clinical relevance and usability through a preliminary study, including completion of the System Usability Scale (SUS). RESULTS: Simulated ROM measurements showed good agreement with both clinical assessments and established literature reference values across ten movements commonly examined in orthopedic practice. The MR tool demonstrated high accuracy, repeatability, and potential to support postoperative decision-making, with usability testing yielding a favorable median SUS score of 82.5, indicating strong acceptance among clinicians. CONCLUSION: The patient-specific MR framework provides a reliable, quantitative, and interpretable method for assessing prosthetic joint performance after replacement, supporting its integration into postoperative workflows for improved surgical outcome assessment.
Machine learning approaches to identifying neurodevelopmental disorders using social media data: a systematic review.
Mohammad SR, Kadaei F, Mohkam M
Int J Med Inform · 2026 Jun · PMID 41812463 · Publisher ↗
BACKGROUND: Neurodevelopmental disorders (NDDs), such as autism spectrum disorder (ASD) and attention-deficit/hyperactivity disorder (ADHD), often emerge in early childhood and can significantly impact social, academic,...
BACKGROUND: Neurodevelopmental disorders (NDDs), such as autism spectrum disorder (ASD) and attention-deficit/hyperactivity disorder (ADHD), often emerge in early childhood and can significantly impact social, academic, and emotional functioning. Early identification is critical to improving long-term outcomes, yet traditional diagnostic processes are time-consuming and resource-intensive. As social media becomes an integral part of daily life, user-generated content offers a novel source of behavioral and linguistic data that may support early detection. Advances in artificial intelligence (AI) and machine learning (ML) now make it possible to analyze these large-scale data streams for clinical insights. METHODS: A comprehensive search was performed across five databases-PubMed, Scopus, Web of Science, IEEE Xplore, and ACM Digital Library-from inception up to July 2025. Studies were included if they used AI or ML methods to analyze social media data for detecting NDDs. Data extraction focused on platform type, targeted disorder, dataset size, ML technique, and diagnostic performance. RESULTS: Nineteen studies met the inclusion criteria. Most focused on ASD and ADHD, using platforms such as Reddit, Twitter, YouTube, and Facebook. ML models achieved moderate to high classification performance, with F1-scores ranging from approximately 0.48 to 0.89 depending on the data type and disorder. Video-based models showed particular promise in identifying nonverbal behavioral markers. However, research on other NDDs remains limited, and methodological heterogeneity, small sample sizes, and ethical challenges persist. CONCLUSION: AI-driven analysis of social media data holds significant promise for scalable, non-invasive screening of neurodevelopmental disorders. While current work largely focuses on ASD and ADHD, future research should extend to underrepresented NDDs and address concerns related to data validity, bias, and privacy. With continued advancement, these tools may serve as valuable complements to traditional diagnostic methods.
Critical appraisal of a multicenter clinical-radiological machine learning model for peripheral artery disease diagnosis.
Tahir HN, Hassan A, Bibi S … +3 more , Tariq U, Yusuf MZ, Ali Y
Int J Med Inform · 2026 Jun · PMID 41812462 · Publisher ↗
Abstract loading — click title to view on PubMed.
Development of a complex digital lifestyle intervention for individuals with prediabetes - Using the Medical research Council (MRC) framework.
Holm TF, Barington PF, Kronborg T … +2 more , Jensen MH, Hangaard S
Int J Med Inform · 2026 Jun · PMID 41812461 · Publisher ↗
BACKGROUND: Prediabetes raises the risk of developing type 2 diabetes (T2D) and related complications, making management vital to lessen diabetes-related burdens. Intensive lifestyle interventions can reduce the risk of...
BACKGROUND: Prediabetes raises the risk of developing type 2 diabetes (T2D) and related complications, making management vital to lessen diabetes-related burdens. Intensive lifestyle interventions can reduce the risk of T2D, but their high resource demands limit widespread use. In prediabetes, digital interventions are effective and may improve accessibility to preventive healthcare by reducing costs. However, findings vary, likely due to insufficient systematic development and reporting of complex interventions, which are crucial for effectiveness and replication. Many health interventions are considered complex, e.g., due to multiple interacting components. This study reports the development process of a complex digital lifestyle intervention for individuals with prediabetes. METHODS: The Medical Research Council (MRC) framework guided the development process. This paper reports on the first phase of the MRC framework; the intervention development. The development was divided into four phases, integrating evidence from current literature and stakeholder engagement. A theoretical framework, combining the Theoretical Domains Framework (TDF) and the Behavior Change Technique (BCT) taxonomy, was applied to determine mechanisms of behavior change and intervention content. The structure and components of the interventions were developed to incorporate the identified mechanisms and content. RESULTS: The developed intervention is structured around a starting period, a 12-week active intervention period, a closing period, and a follow-up period. The active intervention comprises four main components: 1) Individual contact, 2) Education, 3) Exercise, and 4) Social community, each with several subcomponents. The intervention integrates 29 unique BCTs targeting 12 unique mechanisms of behavior change identified from the TDF framework. CONCLUSION: This study developed a complex digital lifestyle intervention for individuals with prediabetes, integrating evidence, theory, and stakeholder engagement. While further development and testing are needed, the study provides a systematic description of the development process, addressing gaps in transparency and supporting effective outcomes, replication, and implementation in clinical practice. ABBREVIATIONS: BCT, Behavior Change Techniques; CReDECI, Criteria for Reporting the Development and Evaluation of Complex Interventions in Healthcare; MRC, Medical Research Council; PROs, Patient-related outcomes; TDF, Theoretical Domain Framework; T2D, type 2 diabetes.
About
- Frequency
- Sun
- Papers found
- 200
- RSS feed
- Subscribe