Asymmetry of temporal causal reasoning in preschoolers: The role of temporal direction representation.

To characterize clinical reasoning in prioritization and test whether errors are linked to experience or are universal, by examining how information congruence and informativeness influence nurses' prioritization and diagnostic reasoning, and by identifying cognitive mechanisms underlying systematic errors under clinical uncertainty. A concurrent embedded mixed-methods study was conducted with 130 nurses from two university hospitals. Using a think-aloud protocol, participants reasoned through four experimentally controlled clinical scenarios in which information congruence (data aligned vs. misaligned with the most common diagnosis) and informativeness (amount of data) were manipulated. Prioritization accuracy (correct vs. incorrect priority) was the primary outcome. Qualitative analysis identified cognitive biases, which were entered into a logistic regression model to quantify their association with accuracy. Accuracy collapsed when nurses faced incongruent clinical data, falling from 49.3 % in congruent scenarios to 18.4 % in incongruent ones (31-point drop; 95 % CI 20-42 %; p<0.001). This decrement was independent of age, experience, educational level, and ward type. Qualitative analysis showed that most nurses (71.4 %) actively dismissed critical conflicting cues. Confirmation bias (OR=0.048, p=0.015) and information bias (OR=0.082, p=0.010), were strong significant predictors of incorrect prioritization. Nursing prioritization errors are systematic cognitive failures rather than random mistakes or simple knowledge deficits. The core vulnerability appears to be metacognitive: an impaired ability to detect and resolve conflict between an activated mental model and new, incongruent information. Interventions to reduce diagnostic and prioritization errors should explicitly train cognitive and metacognitive skills for managing incongruence and flexibly updating clinical representations.

Temporomandibular disorders (TMDs) affect the temporomandibular joint and masticatory muscles, causing pain and dysfunction. Despite the Commission of Dental Accreditation (CODA) mandates for TMD integration into predoctoral curricula, educational delivery remains variable. This study assessed dental students' clinical decision-making in myofascial pain (MFP) and disc displacement with reduction (DDwR) cases using standardized virtual patient scenarios. Dental students completing the Orofacial Pain and Oral Medicine rotation at Herman Ostrow School of Dentistry engaged with five virtual patient scenarios between October 2024 and August 2025. Data from 149 MFP and 80 DDwR cases were analyzed. Student selections for diagnostic tests, diagnoses, and treatment plans were coded based on expert consensus. Correlation and regression analyses examined relationships between test selection, diagnostic accuracy, and overall performance. Among 238 students, pass rates were 81.9% for MFP and 58.8% for DDwR. Diagnostic accuracy strongly predicted success; students selecting correct diagnoses were significantly more likely to pass (MFP: OR=19.8; DDwR: OR>100). Time spent showed no correlation with performance. Critical errors differed by case type: MFP cases showed diagnostic failures (66.7%) and excessive testing (51.9%), while DDwR cases revealed over-selection across multiple categories (81.8%), indicating distinct cognitive challenges. Virtual patient scenarios provide safe, flexible environments to assess and develop diagnostic reasoning in TMDs, revealing persistent gaps in test prioritization and treatment planning. These findings support simulation-based education for dental curricula and emphasize integrated teaching strategies to enhance clinical reasoning and decision-making skills.

Creativity plays an important role in diagnostic reasoning, particularly in supporting calibration and insights. However, the cognitive mechanisms underlying creativity have not yet been clearly articulated in this context. This study focused on associative thinking, an essential component of creativity, and conducted a systematic examination in the context of diagnostic reasoning. To the best of our knowledge, this is the first such study. Specifically, this study explored how the four types of associative thinking-remote, chained, radial, and dissociative-contribute to the generation of diagnostic hypotheses by examining clinical cases. This article presents a theoretical reinterpretation, grounded in the cognitive perspectives of sophisticated reasoning processes routinely employed by expert clinicians. By illuminating the role of associative thinking in diagnostic hypothesis generation, this study provides a novel conceptual perspective that aims to enhance both creativity and diagnostic accuracy, ultimately contributing to improved quality of patientcare.

The differentials for rapidly progressive dementia are broad, encompassing structural, infectious, inflammatory, neoplastic, and neurodegenerative etiologies. The presence of abnormal movements further complicates the diagnostic approach. We describe a 69-year-old man presenting with a diverse array of neurologic symptoms, starting with rapidly progressive cognitive impairment, later developing abnormal movements, sleep disruption, and constitutional symptoms. Despite extensive investigations and empirical treatment, the diagnosis remained elusive until postmortem evaluation. This case highlights the challenges inherent in neurologic diagnostic odysseys, offering insight into the diagnostic reasoning process and unveiling novel clinical findings that may aid earlier recognition of this rare disorder.

Despite the growing use of Electronic Health Records (EHR) for AI-assisted diagnosis prediction, most data-driven models struggle to incorporate clinically meaningful medical knowledge. They often rely on limited ontologies, lacking structured reasoning capabilities and comprehensive coverage. This raises an important research question: Will medical knowledge improve predictive models to support stepwise clinical reasoning as performed by human doctors? To address this problem, we propose DuaLK, a dual-expertise framework that combines two complementary sources of information. For external knowledge, we construct a Diagnosis Knowledge Graph (KG) that encodes both hierarchical and semantic relations enriched by large language models (LLM). To align with patient data, we further introduce a lab-informed proxy task that guides the model to follow a clinically consistent, stepwise reasoning process based on lab test signals. Experimental results on two public EHR datasets demonstrate that DuaLK consistently outperforms existing baselines across four clinical prediction tasks. These findings highlight the potential of combining structured medical knowledge with individual-level clinical signals to achieve more accurate and interpretable diagnostic predictions. The source code is publicly available on https://github.com/humphreyhuu/DuaLK.

Purpose: To evaluate the diagnostic accuracy of GPT-4.o, Claude 3.7 Sonnet, and Gemini 2.5 Pro for real-world retinal cases. Methods: Forty retina cases from the University of Iowa's EyeRounds repository were assessed under 2 conditions: (1) full clinical context (textual history, examination, and image descriptions); and (2) image-only (raw clinical images). Outputs were evaluated for diagnostic accuracy, symptom/sign recognition, differential diagnosis, and treatment recommendations. Cochran Q, McNemar test, analysis of variance, and Cohen kappa were used to make statistical comparisons. Results: In full-context cases, GPT-4.o achieved the highest diagnostic accuracy (78.4%), followed by Claude (73.0%) and Gemini (29.7%) (P < .001). In image-only cases, Claude was superior (73.7%), outperforming GPT-4.o (63.2%) and Gemini (31.6%) (P < .001). GPT-4.o was best at identifying signs/symptoms in full-context cases (64.0%), while Claude excelled in image-only cases (63.3%). GPT-4.o (42.9%) and Claude (37.0%) outperformed Gemini (25.9%) on differential diagnosis (P ≤ .011). Claude showed superior treatment accuracy in image-only cases (61.1%), while GPT-4.o was superior in full-context cases (57.3%), and both outperformed Gemini (35.8%). GPT-4.o and Claude showed substantial agreement in image-based diagnoses (κ=0.658), while Gemini showed minimal agreement (κ≤0.196). Conclusions: GPT-4.o and Claude demonstrated strong diagnostic and clinical reasoning in retinal cases, with Claude excelling in image-based analysis and GPT-4.o in text-rich contexts. Gemini's lower performance underscores the importance of careful model selection in clinical applications.

Can large language models perform clinical anamnesis? Comparative evaluation of ChatGPT, Claude, and Gemini in diagnostic reasoning through case-based questioning in oral and maxillofacial disorders.

Leveraging human pose estimation for diagnostic feedback: Action research on instructional mediation and sustainable learning in coach education.

Comparative analysis of diagnostic prediction and clinical reasoning of large language models in complex endodontic case scenarios.

Non-syndromic hyper-IgE in children: A practical approach.

Primary care is entering a period of structural strain as CMS expansions, Medicare Advantage verification requirements, and growing EMR and quality metric pressures converge in 2026. This article describes how these simultaneous demands are reshaping the pace, tone, and reliability of clinical encounters, from the moment a patient checks in to the moment a clinician begins diagnostic reasoning. The added layers now exceed the capacity of workflows built for a different era, eroding presence, situational awareness, staff stability, and patient trust. These failures rarely look dramatic; they appear as missed nuance, delayed follow-through, and the quiet loss of continuity. The path forward requires redesign, not endurance — pre-visit planning, clearer task distribution, better information architecture, protected cognitive space, and a renewed understanding that staff morale is operational infrastructure. Quality and patient experience will improve only when the structure around clinicians is rebuilt to support the work itself.

Corrigendum to “Moving Toward Competency: Development of a Simulation-Based Formative Assessment of Nurse Practitioner Student Diagnostic Reasoning,” The Journal for Nurse Practitioners, 2026;22:105598

Interpretable, automated Artificial Intelligence (AI) solutions are essential for accurate 12-lead electrocardiogram (ECG) arrhythmia classification because they remove the time-consuming and inconsistent aspects of manual interpretation. Current models are limited in complexity, data variety, and validation. This paper proposes a novel Deep Learning (DL) architecture that combines Convolutional Neural Networks (CNNs), Bidirectional Long Short-Term Memory (Bi-LSTMs), and transformer layers to jointly extract morphological, temporal, and spatial patterns from ECG signals. The model was trained and evaluated on the PhysioNet/Computing in Cardiology Challenge 2020 dataset, comprising more than 43,000 multi-label ECG recordings across 27 arrhythmia classes. It achieved an accuracy of [Formula: see text], a macro-F1 score of [Formula: see text], and an Area Under the ROC Curve (AUC) exceeding [Formula: see text] for life-threatening arrhythmias such as Ventricular Premature Beats (VPB) and Atrial Fibrillation (AF). To ensure clinical transparency, the model integrates SHAP (SHAPley Additive exPlanations), enabling case-by-case interpretability by attributing predictions to physiologically relevant waveform segments and ECG leads. This approach aligns with cardiologists' diagnostic reasoning and supports real-world decision-making. Additionally, the model is computationally efficient, with a footprint of [Formula: see text] and inference latency of [Formula: see text], enabling deployment in telemedicine, wearable monitoring systems, and critical care settings. The proposed framework achieves high diagnostic performance, robustness to class imbalance, and human-level interpretability simultaneously, providing a reliable, scalable solution for automated ECG analysis. These findings advance the application of explainable DL algorithms in cardiovascular diagnostics.

The debate about whether health visiting, a specialist community public health nursing role, is at the level of advanced practice nurse has gone on for more than a decade. There is little empirical evidence that the role matches the traditional role of an advanced practice nurse, although many of the attributes of advanced practice nursing such as prescribing rights, managing complex cases, caseloads with undifferentiated need and advanced assessment and decision-making are certainly present. The current study aimed to develop, refine and test the Health Visiting Advanced Practice Scale to assess the scope of advanced practice of UK health visitors. A cross-sectional and methodological scale validation design, following classical test theory. The design consisted of three phases; the first involved scale development including item generation, phase two assessed the content validity index, and the third phase involved a cross-sectional survey to establish construct validity, content validity, and internal consistency reliability, and conduct exploratory and confirmatory factor analysis. The initial 44-item scale underwent iterative exploratory and confirmatory factor analyses, leading to a refined 5-factor structure with 29 items covering domains such as family-centred care, leadership, prescribing, diagnostic reasoning, and professional practice. This final version demonstrated strong reliability and construct validity in the EFA but mixed fit indices in the CFA, supporting both internal consistency and validity of the scale. The final scale offers a rigorously validated tool for assessing advanced practice among UK health visitors, capturing core domains such as family-centred care, leadership, prescribing, and diagnostic reasoning. By bridging theoretical frameworks with real-world practice, it fills a critical gap in evaluating and supporting the professional scope of this public health nursing specialty. These findings provide valid and reliable insights for measuring and improving health visitors' advanced practice and developing future professional policies. No patient or public contribution. STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines for cross-sectional studies.

Breast ultrasound is a key diagnostic method for breast cancer and relies heavily on the interpretation of visual cues. At the same time, medical education is increasingly being driven by time constraints, which favors rapid pattern recognition, limiting the scope for reflective image analysis and the diagnostic process. Therefore, the aim of this study was to propose and evaluate an artistic and pedagogical teaching model, inspired by the interpretive practices of Italian High Renaissance painting, as a tool to support the development of diagnostic reasoning in breast ultrasound. This model focuses on careful observation, analysis of the relationship between detail and the overall image, and the conscious transformation of visual cues into clinical meaning. This study was conducted during the four-day ARSA Think Tank Meeting (ARSATTM). Medical students worked in four groups; two groups received methodological training based on visual cue analysis, and two did not. All groups performed identical tasks involving the interpretation of breast ultrasound images and ultrasound examinations on real patients. The results indicate that an artistic-pedagogical teaching model to promote more coherent and reflective diagnostic reasoning in breast ultrasound is feasible. Therefore, integrating this approach may be a valuable addition to medical students' ultrasound education in the realities of limited clinical time.

Clinical decision-making under extreme time pressure represents one of the most demanding cognitive challenges in emergency medicine. This review analyzes how temporal compression, cognitive load, workflow interruptions, and stress exposure interact to influence diagnostic accuracy and patient safety. Drawing on foundational literature in diagnostic reasoning, cognitive psychology, and patient safety research, the study synthesizes evidence demonstrating that increasing interruption frequency and cognitive burden are associated with measurable rises in task error and progressive declines in diagnostic reliability. Temporal constraints appear to accelerate reliance on intuitive processing, increasing susceptibility to premature closure, particularly in ambiguous or evolving clinical presentations. Additionally, elevated stress levels are associated not only with reduced mean diagnostic accuracy but also with greater variability in performance. The findings suggest that diagnostic vulnerability in emergency settings is not random but emerges predictably from the interaction between cognitive architecture and system design. Structured mitigation strategies—including micro-reflective checkpoints, diagnostic prompts, workflow protection from unnecessary interruptions, and simulation-based reasoning training—are associated with reductions in error patterns without compromising response time. The implications are particularly relevant for emergency departments operating under structural constraints, including middle-income health systems in Latin America. Strengthening diagnostic safety requires an integrated approach that aligns clinical expertise, environmental design, and institutional safety culture.

Background: Radiology readouts, which involve student-teacher (mentored) and student-student (peer) interactions, are a cornerstone of medical education; however, communication dynamics in these settings, and students’ perceptions of them, are underexplored. This qualitative study examined medical students’ perceptions of peer and mentored learning during radiology readouts and analyzed associated verbal communication patterns. Methods: This qualitative observational study was conducted at Maastricht University in the Netherlands between April and May 2023. We conducted the study in three stages: (1) observation of student-student interactions, (2) observation of student-teacher interactions, and (3) semi-structured interviews with students about their perceptions of these interactions. We categorized verbal communication data using an adapted Verbal Response Modes (VRM) taxonomy, grouping intents into cognitive structuring, instructing, and questioning. We analyzed the interview data thematically. Results: Verbal communication analysis revealed that cognitive structuring during student–student interactions primarily involved disclosure and confirmation, while student-teacher interactions also included interpretation. Questioning was consistent across both interaction types, but instructing, such as advisement, was more prevalent in student-teacher interactions. We identified two key themes in the interview data. Theme 1: Peer interactions fostered uncertainty, while teacher interactions provided certainty through accurate information. Theme 2: Peer interactions facilitated verbalization of thoughts, whereas teacher interactions enhanced thought processes through meaningful prompts and insights. Conclusion: These findings indicate that student–teacher interactions are more responsive (interpretation) and directive (advisement), promoting certainty and deeper discussion, whereas student–student interactions, though more egocentric (disclosure), support thorough articulation despite perceived uncertainty. This study informs the design of radiology education by highlighting the complementary roles of peer and teacher interactions in fostering diagnostic reasoning and managing uncertainty.

The exponential growth of medical data and complexity in Pulmonary and Critical Care Medicine (PCCM) necessitates a paradigm shift in clinical reasoning and education. This paper proposes a structured framework for collaborative decision-making between physicians and artificial intelligence (AI) that emphasizes the integration of critical thinking with AI-enhanced capabilities. Critical thinking is defined as the systematic and objective analysis of information accompanied by logical reasoning and evidence‑based judgment. Traditional medical education often neglects explicit instruction in these cognitive skills, leaving clinicians vulnerable to diagnostic error and cognitive biases. Meanwhile, AI excels at data aggregation, pattern recognition, and predictive analytics, offering complementary capabilities that can support evidence-based decision-making. By distinguishing two modes of interaction AI‑assisted tools that provide recommendations and AI‑enhanced systems that simulate complex scenarios, we propose a conceptual model in which AI reinforces rather than replaces physician judgment. The article outlines how critical‑thinking skills map onto phases of diagnostic and management reasoning and illustrates the roles of physicians and AI. We discuss applications of AI in diagnosis, radiology, pulmonary function interpretation and decision support, and outline ethical considerations including algorithmic bias, data privacy and HIPAA compliance. This framework reimagines the physician-AI relationship as a cognitive partnership essential for the future of PCCM.

This study provides a descriptive, multi-institutional comparison of dental students' recognition accuracy and management decisions of oral soft tissue pathological entities across four US dental schools. While prior single-institution studies have evaluated diagnostic ability, this work provides a multi-institutional comparison to identify recognition and decision-making variability as well as potential educational implications. A voluntary, anonymous online survey with twenty image-based multiple-choice questions was administered to 160 students. Each question included an image, patient demographics, and key case details. Management questions were provided with reference diagnoses to prevent conflation of diagnostic and management errors. Descriptive statistics, t-tests, logistic regression, chi-square tests, and multi-rater kappa statistics were used to analyze performance, with effect sizes, reference groups, and 95% confidence intervals reported. Model fit for logistic models was assessed using likelihood ratio tests. Students demonstrated a high level of recognition for conditions like candida (84%) and tobacco keratosis (95%). Subtle or clinically variable lesions were more challenging, such as erosive lichen planus (50%) and idiopathic leukoplakia (58%). Management accuracy frequently lagged behind recognition accuracy. Significant inter-school differences were observed for both recognition (χ2=27.66, p<0.0001) and management (χ2=30.80, p<0.0001), with students from School #4 outperforming peers. Kappa values remained low (<0.2 for most items), indicating wide variability and limited internal agreement. Students demonstrated strong theoretical knowledge of common oral pathological entities but variability in identifying and managing rare or diagnostically nuanced conditions. Because this survey evaluates recognition rather than competence, findings highlight the need for case-based, competency-aligned teaching approaches to strengthen diagnostic reasoning and management decision-making.