JOURNAL OF MEDICAL INTERNET RESEARCH医学互联网研究杂志
JOURNAL OF MEDICAL INTERNET RESEARCH(英文缩写 J MED INTERNET RES),ISSN 1439-4456,eISSN 1438-8871,中文译名:医学互联网研究杂志 是一本学术期刊。本页汇总该期刊的最新影响因子、分区信息以及最新收录于 PubMed 的文献,帮助您快速了解期刊全貌。
发文量统计区间：2025-11-01 至 2026-10-31，按本站收录文献的发表日期统计。
期刊介绍
JOURNAL OF MEDICAL INTERNET RESEARCH 最新收录文献
-
1. Exploring Conversational Dynamics in Scientific and Pseudoscientific Health Communities on YouTube: Process Mining and Network Analysis Study.
PMID:DOI:日期:2026-10-02Social media platforms, particularly YouTube (Google LLC), are important sources of health information, but also significant vectors for misinformation and pseudoscience. While many studies analyze the content and sentiment of this information, the dynamic, sequential nature of user interactions, which can shape belief formation and community dynamics, remains poorly understood. This study aimed to explore the applicability of network analysis and process mining techniques for identifying and comparing structural and emotional patterns of conversational flow within a retrieved corpus of YouTube comments from videos about scientific or pseudoscientific health treatments. We conducted an exploratory observational study using publicly available YouTube comment threads posted between 2011 and 2025 from videos categorized as either "scientific" (20,387 comments) or "pseudoscientific" (32,025 comments) using an automated pipeline that combined API-based data extraction, large language model-based video classification, and natural language processing techniques for multilingual sentiment and thematic classification of comments. We then applied process mining to model the temporal sequences of interactions, and network analysis to map the relationships between conversational topics. Network analysis revealed divergent conversational cores. In the scientific corpus, negative expressions of feelings and negative comparisons had greater normalized node strength, while medical treatment and advice requests were also relatively more prominent. In the pseudoscientific corpus, positive expressions of feelings, thanking, compliments, and emoji-only or brief acknowledgments showed greater prominence. Process mining showed a more heterogeneous combination of negative, positive, and neutral activities in the scientific corpus, whereas the most frequent pathways in the pseudoscientific corpus were concentrated around positive expressions of feelings, thanking, compliments, and brief messages of acknowledgment, or simply emojis. The application of network analysis and process mining techniques revealed distinct patterns in the conversational dynamics of health-related YouTube discussions. Within the sampled corpora, scientific discussions appeared more compatible with mixed-valence evaluation and treatment-related exchanges, whereas pseudoscientific discussions showed patterns more consistent with interpersonal affirmation and socioemotional bonding. These preliminary findings suggest that network analysis and process mining are promising approaches for investigating online health communication and misinformation ecosystems. They also suggest that public health strategies may benefit from considering the affective and community-bonding dimensions of engagement in misinformation communities alongside information provision.
-
2. Biomedical Research Images Manipulated by Generative AI to Alter Scientific Outcomes: Diagnostic Study of Human and Automated Detection.
PMID:DOI:日期:2026-10-02In a diagnostic study of 104 western blot and subcutaneous xenograft tumor images, a high-fidelity generative model produced forgeries that could alter the conclusions of a study; 24 PhD-level expert reviewers could not reliably distinguish the forgeries from authentic figures (mean accuracy 50.5%, SD 6.9%), while the best-performing commercial AI detector achieved only moderate discrimination (area under the curve 0.790, 95% CI 0.695-0.885), revealing critical vulnerabilities in current research-integrity safeguards.
-
3. Periodontitis Risk Assessment and Prevention Planning: Comparative Study of Multimodal Large Language Models and Periodontist Evaluations.
PMID:DOI:日期:2026-10-02Periodontitis is one of the most prevalent yet preventable oral diseases, as indicated by multiple clinical and radiographic factors. As these factors are recorded in electronic health records (EHRs), their reuse offers opportunities for personalized risk assessment and targeted prevention. Predictive AI and traditional machine learning models support fragmented detection tasks but lack the integration of textual and imaging predictors. Emerging multimodal large language models (M-LLMs) show promise in combining these data sources for clinical assessment. Evaluating the capabilities of M-LLMs and comparing them against the current clinical standard are therefore essential to determine their potential as digital assistants. This study aimed to evaluate the ability of M-LLMs to assess periodontitis risk and suggest prevention strategies, based on EHR data and radiographic findings. Each M-LLM was individually evaluated by periodontal experts, benchmarked against other models, and compared with a periodontist as a reference. A vignette study was conducted following TRIPOD (Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis) guidelines for the evaluation of LLMs. Ten periodontal vignettes were created, each including a panoramic radiograph and textual EHR data. Three LLMs capable of reasoning and handling multimodal data were compared to a periodontist who generated outputs manually, based on the same prompts and input data. Periodontal experts rated all outputs across 6 predefined criteria on a 5-point Likert scale. Statistical analyses evaluated overall performance per model and tested whether performance varied per model, scenario complexity, or rater. GPT o1 Pro and Claude Sonnet 4 showed strong performance, with 86.7% and 85.6% of ratings deemed acceptable-comparable to the periodontist's output (87.8%). Gemini 2.5 Pro was rated significantly lower than both the periodontist and the other models (59.4% acceptable; <.002). Radiographic interpretation consistently received lower scores than other abilities across all models and the periodontist, with Gemini rated below the acceptable threshold. The time required for completion ranged from approximately 10 seconds for Claude to 37 seconds for Gemini; 3 minutes, 22 seconds, for GPT; and 5 minutes, 57 seconds, for the periodontist. M-LLMs demonstrated strong reasoning abilities in periodontal assessment. Across all models, unacceptable elements were consistently related to errors in radiographic interpretation, though refined prompting or newer model versions may improve this. Notably, even when radiographic findings were incorrect and plaque-retentive factors were absent, outputs were still rated well, indicating that EHR data alone provide a substantial basis. For clinical applicability, M-LLMs must at least perform comparably to a periodontist and meet the quality standards set by periodontal experts-a bar that GPT and Claude appear to approach.
-
4. Voice-Enabled Virtual Patients for Interactive Training in Standardized Clinical Assessment: Mixed Methods Pilot Study.
PMID:DOI:日期:2026-10-02Training mental health clinicians to conduct standardized clinical assessments is challenging due to a lack of scalable, realistic practice opportunities. Traditional methods often fail to prepare trainees for the variability and complexity of real-world patient interactions, potentially impacting data quality in clinical trials. This paper introduces a novel approach to address this training gap using large language model (LLM)-based interview simulations. This study aimed to develop and validate a voice-enabled virtual patient simulation system as a proof of concept. We described the development of the system and evaluated whether it could generate virtual patients that (1) accurately adhered to predefined clinical profiles, (2) maintained a coherent and consistent narrative, and (3) produced dialogue that is perceived as realistic. We implemented a system that used an LLM to simulate patients with specified symptom profiles, demographic backgrounds, and distinct communication styles. The system's performance was analyzed through a mixed methods evaluation, which included a formal assessment by 5 experienced clinical raters who conducted simulated structured Montgomery-Åsberg Depression Rating Scale (MADRS) interviews with 4 virtual patient personae, scored them on the scale, and provided qualitative feedback on the system's clinical plausibility, narrative cohesion, and dialogue realism. Across 20 interviews, the virtual patients demonstrated reasonable adherence to their configured clinical profiles, with human rater scores falling near their predefined score configurations. The mean item difference between MADRS rater scores and configured scores was 0.52 (SD 0.75); interrater reliability for the total score was 0.90 (95% CI 0.68-0.99). Expert raters consistently gave average ratings of "agree" to "strongly agree" when asked to evaluate the qualitative realism and cohesion of the virtual patients. LLM-powered virtual patient simulations offered a promising, scalable tool for training clinicians in standardized clinical assessment. This pilot study provides initial evidence for the system's ability to produce clinically relevant practice scenarios with reasonable fidelity.
-
5. Refined Exclusion in Medical AI: Data Justice and Patient Safety Governance.
PMID:DOI:日期:2026-10-01Medical AI is often evaluated using aggregate measures of discrimination, calibration, and accuracy. However, these measures can obscure clinically important variation across patient groups, institutions, devices, and workflows. This viewpoint defines refined exclusion as a governance condition in which an AI system appears successful in aggregate, while uncertainty, error, or reduced clinical reliability is concentrated in populations that are insufficiently represented, measured, validated, or monitored. The concept does not replace algorithmic fairness, hidden stratification, dataset shift, or subgroup performance analysis. It connects these mechanisms to a distinct consequence: an unequal distribution of safety that remains inadequately detected or corrected. Drawing on purposively selected, illustrative evidence from population health management, chest radiography, dermatology, computational pathology, medical foundation models, and clinical measurement, we distinguish model-level disparity, patient safety signals, and documented patient harm. We then frame data justice as a complementary governance approach with distributional, procedural, and substantive dimensions. The proposed lifecycle decision gates address intended use, subgroup learnability, data provenance, validation, procurement, local deployment, monitoring, updates, and patient feedback. Each gate links minimum evidence to decision authority and 1 of 4 actions: proceed, enrich or validate, restrict use, or pause or retire. Governance intensity should be proportionate to clinical risk and evidentiary uncertainty. By linking subgroup evidence gaps to institutional decisions and corrective action, the framework shifts attention from whether a model performs well on average to whether its safety is demonstrable for the populations and settings in which it will be used.
-
6. Effects of Generative AI-Supported Virtual Reality Dental Skill Training on Learning Performance, Visual Behavior, and Cortical Activation Among Dental Students: Stratified Randomized Controlled Trial.
PMID:DOI:日期:2026-10-01The rapid advancement of AI, particularly generative AI (GenAI), is reshaping educational paradigms. However, its utility in preclinical digital medical skills education remains insufficiently explored. This study aims to investigate whether GenAI‑assisted training yields superior performance relative to traditional instruction in virtual reality (VR)-supported preclinical dental skill education. A total of 123 eligible students from a top‑tier Chinese university completed the trial. Participants were stratified by baseline theoretical scores and randomly assigned within each stratum into two arms: the GenAI-assisted group (n=62) received ChatGPT guidance without human tutoring, with research assistants monitoring AI outputs solely for safety and error checking; the teacher-led control group (n=61) received on-site instructor guidance. Both groups underwent standardized VR-based dental skill training for 7 days. Assessments included dental operative skill examinations, functional near-infrared spectroscopy recording, continuous eye-tracking monitoring, visual-spatial ability tests, and questionnaires measuring self-regulated learning, learning emotions, and engagement. As the primary outcome, there was a statistically significant intergroup difference in operative test scores between the control group (mean 67.39, SD 16) and the GenAI‑assisted group (mean 77.25, SD 11.86), with the GenAI‑assisted group exhibiting higher operative test scores relative to the control group (t=3.88, Cohen d=0.70; P<.001). For secondary outcomes, the GenAI‑assisted group also demonstrated significantly lower cognitive‑load indicators (mean 0.180, SD 0.136) than the control group (mean 0.346, SD 0.150; P<.001), accompanied by more efficient visual attention allocation patterns and stronger activation in prefrontal cortex, motor cortex, visual association cortex, and temporoparietal junction brain regions. The GenAI‑assisted group showed better performance across low (P=.003), medium (P=.004), and high (P=.004) level visual-spatial ability tests. Among low-achieving participants, the largest between-group difference in self-regulated learning was found in favor of the GenAI-assisted group (P=.02). Better scores were also observed for the GenAI‑assisted group in learning enjoyment, as well as behavioral, emotional, and cognitive engagement. Students randomized to GenAI-assisted training had higher immediate operative scores than students receiving teacher-led instruction. Secondary cognitive, behavioral, and neurophysiological findings were exploratory and pointed toward a link between GenAI exposure and variations in skill‑acquisition metrics among dental students receiving VR-based training. The results offer preliminary insights regarding the potential integration of GenAI into preclinical dental skill education. Chinese Clinical Trial Registry ChiCTR2500110063; https://tinyurl.com/2hkv5wt9.
-
7. Implementing Telemedicine for Neonatal Care: Tutorial on a Practical Toolkit Based on Multinational Experience.
PMID:DOI:日期:2026-09-30Real-time audiovisual connections between health care providers (HCPs) in neonatal care, known as TeleNeonatology (TeleNeo), can improve neonatal care. For patients and families, TeleNeo was found to improve patient outcomes and facilitate family-integrated and patient-centered care by ensuring timely access to expert involvement regardless of location, which can strengthen trust and reassurance. For clinicians and health care organizations, TeleNeo enables expert decision-making, fosters continuous professional development, promotes knowledge exchange between hospitals, and increases staff confidence in managing complex medical cases. However, organizational, technical, and infrastructural requirements can hinder successful implementation and sustained adoption of technological interventions, such as TeleNeo. Implementation can be time-consuming and may fail due to the challenges encountered during the implementation process, particularly when incorporating the intervention into existing workflows across multiple hospitals. Nevertheless, there are case studies that demonstrate successful implementation of TeleNeo into routine care. Informed by international experience and theoretical underpinnings of implementation science, this tutorial presents a toolkit to provide step-by-step guidance for health care institutions considering TeleNeo implementation. The objective of the toolkit is to help health care organizations effectively and efficiently implement TeleNeo in their neonatal care pathways. The toolkit provides a structured guide that encompasses the entire implementation process, from initial ideas to stakeholder engagement, workflow design, and program evaluation. It includes checklists, planning guidelines, and tools to help teams design a customized implementation strategy for their institution.
-
8. Behavior Change and Compliance Outcomes of a Novel Web-Based Health Behavior Tool Integrated Into Certified Exercise Practitioner Services: Twelve-Week, Two-Arm, Randomized Pre-Post Trial.
8. 一种新型基于网络的健康行为工具整合到认证运动从业者服务中的行为改变和依从性结果：为期十二周、双臂、随机前后对照试验PMID:DOI:日期:2026-09-30Healthy lifestyle behaviors, including a balanced diet, regular physical activity, adequate sleep, avoiding tobacco, moderating alcohol consumption, and stress management, are associated with reduced risk of chronic disease and improved well-being. In recent years, digital interventions have emerged as cost-effective platforms for promoting these behaviors, yet few have been integrated into services delivered by certified exercise practitioners (CEPs) or focused on multiple domains of health. This study aimed to examine changes in health behaviors during implementation of the Food, Exercise, Sleep, and Stress (FESS) web-based questionnaire as an adjunct to traditional face-to-face services provided by CEPs and compare these changes across 2 prompting frequencies. A 2-arm, randomized pre-post intervention design was conducted with 41 participants (23 women), with a median age of 65 (IQR 58-70) years, who were recruited from CEP clinics. Participants were randomized to complete the FESS questionnaire either 3 times per week (3-day group, n=21) or daily (everyday group, n=20) for 12 weeks. The questionnaire tracked health behaviors across 4 domains and provided health tips. Outcomes were assessed using the assessment version of the FESS questionnaire at baseline and at 6 and 12 weeks. Changes in FESS questionnaire responses were examined over time using Friedman ANOVA with Wilcoxon signed-rank post hoc tests. Group differences between prompting frequencies were assessed by calculating changes in questionnaire responses from baseline to 6 and 12 weeks and analyzing them using the Mann-Whitney test. Compliance was assessed by calculating the proportion of completed FESS questionnaires relative to the total number of questionnaires prescribed for each participant within each intervention arm. Validity of the exercise domain was evaluated by comparing self-reported physical activity and sedentary time from daily FESS responses with accelerometer-derived measures in a subsample of participants. Positive or no changes were observed across most behaviors, with significant improvements in vegetable consumption, water intake, stress, and sleep. Group comparisons indicated larger changes in stress and sleep behaviors in the everyday group; however, compliance was higher in the 3-day group. Compliance declined over time in both groups. Accelerometer validation of the exercise domain showed low rates of overreporting exercise (3/59, 5.1%) and underreporting sedentary behavior (11/59, 18.6%). The FESS questionnaire demonstrated feasibility and was associated with observed changes in health behaviors, particularly in diet, stress, and sleep, when integrated into CEP services. While daily use was associated with larger observed changes, reduced compliance highlights the importance of flexible engagement strategies. These preliminary findings suggest that CEP services may benefit from embedding structured digital tools into traditional face-to-face interactions, as such tools may support education, promote reflection and accountability between sessions, and facilitate attention to a broader range of lifestyle factors within routine CEP care. Further controlled studies are required to confirm these findings.
-
9. Explainable Machine Learning Predictive Models for Surgical Site Infections: Scoping Review.
PMID:DOI:日期:2026-09-30Surgical site infections (SSIs) remain a major cause of health care-associated infections, and early prediction is essential for improving patient outcomes. Machine learning (ML) has shown potential for SSI prediction; however, clinical implementation requires models that are both accurate and explainable. Despite recent progress in explainable ML, its clinical application to SSI prediction remains limited. This study aimed to map explainable ML models for SSI prediction from a clinical perspective and examine their use of structured and unstructured data across the dimensions of data, methodology, and explanation output. We conducted a scoping review following PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) and Joanna Briggs Institute (JBI) guidance, and registered the protocol in PROSPERO. Six databases were searched for eligible studies published from January 2010 onward, without language restrictions. The search was conducted on August 9, 2025, and updated on July 14, 2026. We included studies that developed or validated an explainable ML model for predicting SSI in adults. Two reviewers (RS and YL) independently screened studies and extracted data. Findings were narratively synthesized and presented in evidence maps. Methodological quality was assessed using the PROBAST+AI (Prediction model Risk Of Bias Assessment Tool for prediction models using regression or artificial intelligence) methods. Overall, 77 studies reporting 98 ML models were included. Most models were prognostic (72/98, 73.5%), whereas 26 focused on postoperative SSI diagnosis. A total of 81.8% (63/77) of studies addressed a single surgical specialty, most commonly gastrointestinal surgery (27/63, 42.9%). Overall, 51.9% (40/77) of studies addressed composite SSI predictions. Among all models, % (40/98) were black-box models explained by post hoc methods; SHAP combined with ensemble learning was the leading approach (18/40, 45%). Regression models accounted for half of the inherently interpretable models, interpreted using coefficients. Prognostic models commonly included health and lifestyle (60/72, 83.3%), individual characteristics, and surgical process details (both 57/72, 79.2%); health and lifestyle factors were most frequently important across SSI types. Diagnostic models commonly included surgical process details (13/26, 50%), administrative codes, and individual characteristics (both 10/26, 38.5%). Key diagnostic predictors varied by SSI types: postoperative clinical interventions predominated for composite SSI; vital signs, postoperative interventions, and administrative codes for superficial SSI; postoperative recovery status for deep SSI; and vital signs for organ-space SSI. Extending previous reviews focusing on model performance, this review mapped explainability methods and important features in SSI prediction, identifying recurring predictor patterns and substantial methodological heterogeneity across prognostic and diagnostic settings. Incomplete reporting of feature definitions and explanatory rationale, together with limited clinical relevance, constrained clinical interpretation and actionability. Clinician-informed reporting frameworks and validation of explanation fidelity and clinical relevance are needed to improve the trustworthiness and utility of SSI prediction models.
-
10. Electroencephalography in Subjective Cognitive Decline and Mild Cognitive Impairment: Systematic Review of Biomarkers, Classification, and Prognostic Evidence.
PMID:DOI:日期:2026-09-30Subjective cognitive decline (SCD) and mild cognitive impairment (MCI) are heterogeneous clinical states that may represent early or at-risk stages of Alzheimer disease (AD) and other dementias in some individuals. Improved characterization and risk stratification in these populations may facilitate timely evaluation and intervention. Electroencephalography (EEG), a noninvasive, cost-effective neurophysiological technique with high temporal resolution, holds significant potential for elucidating neural mechanisms and providing candidate neurophysiological markers associated with SCD and MCI. The study aims to systematically synthesize evidence on group-level EEG biomarkers, EEG-based classification models, the evaluation of EEG in screening or diagnostic pathways, and EEG-based prediction of progression in SCD and MCI. A search was conducted across PubMed, Web of Science, Cochrane Library, MEDLINE (via Ovid), Scopus, Wanfang Data, CQVIP, Yiigle, and CNKI databases to include reports published in English or Chinese, which reported group-level EEG differences, EEG-based classification, screening or diagnostic evaluations, or prognostic outcomes in SCD and MCI. A total of 2 independent reviewers screened the titles and abstracts. Prediction-model reports were assessed using PROBAST+AI (Prediction Model Risk of Bias Assessment Tool and Applicability Assessment for AI), prognostic-factor reports using QUIPS (Quality in Prognosis Studies), and other observational reports using design-specific Joanna Briggs Institute checklists. Certainty of evidence was assessed using the GRADE (Grading of Recommendations Assessment, Development, and Evaluation) system. A total of 20 reports were included, representing a maximum of 3853 confirmed-deduplicated participants after correction of a nested subsample and removal of confirmed duplicate participants across reports, of whom 2927 diagnosed with SCD/MCI/AD or other dementias. Of these, 10 reports were assessed as prediction or AI model reports, 9 as nondiagnostic observational reports, and 1 as a prognostic-factor study. Although potentially informative between-group EEG differences and preliminary model performance were reported, no conventional diagnostic test-accuracy study was identified. The certainty of evidence was very low across all 4 evidence bodies. The available evidence suggests that EEG is a promising noninvasive modality for characterizing neurophysiological alterations associated with SCD and MCI. EEG-based classification and prognostic models have also shown encouraging preliminary performance. However, the current evidence is more supportive of biomarker discovery and model development than of established clinical screening or diagnosis. Translation into routine practice will require prospective reports with standardized EEG procedures, representative clinical populations, prespecified thresholds, participant-level data separation, and independent external validation. These findings provide a basis for evaluating EEG as a potential adjunctive or triage tool in future clinical pathways.