共找到 20 条结果
As the public increasingly interacts with artificial intelligence (AI) chatbots, we compared the answers from five AI chatbots to standardized questions that patients might ask about ovarian and cervical cancer. ChatGPT 3.5, Google Gemini 2.0, Reddit Answers, Bootcamp, and DeepSeek were queried with 15 frequently asked questions (FAQs) on cervical cancer and 11 on ovarian cancer. In a blinded, randomized survey, each deidentified response was independently assessed by three gynecologic oncologists using a 4-point scale (1, accurate and comprehensive; 2, accurate but inadequate; 3, accurate but outdated or inaccurate; 4, completely inaccurate). Readability (Flesch-Kincaid grade level) and word count were recorded. For cervical cancer FAQs, ChatGPT 3.5, Google Gemini 2.0, Reddit Answers, Bootcamp, and DeepSeek received scores of 1.4, 1.6, 2.7, 1.5, and 1.5, respectively. For ovarian cancer FAQs, average scores were 1.3, 1.4, 2.5, 1.2, and 1.4, respectively. All AI chatbot responses were written at a reading level above 11th grade, making them generally difficult for the average American to read. Although generally rated as accurate and adequate, the answers were frequently off-topic and not generalizable. Healthcare providers should be aware of unintentionally generated misinformation to better counsel patients.
Research focused on Alzheimer's disease (AD) 'biomarker clocks' seeks to identify ages at which AD pathological landmarks occur (e.g., initiation of amyloid accumulation) and are meaningfully related to disease outcomes (e.g., symptom onset). However, the statistical approach for assessing the association between age at biomarker-clock event and remaining time to clinical symptom onset can create a structural artifact. We term it here the 'countdown paradox', because the remaining time to symptom onset shrinks as the age at biomarker-clock event increases, which may result in inaccurate associations between age at biomarker-clock event and the remaining time. We conducted analyses to examine this issue with simulation studies and theoretical results, and also examined it empirically using five biomarkers in two longitudinal AD-related cohorts (BIOCARD and ADNI): (1) CSF Aβ42/Aβ40, (2) CSF p-tau181, (3) plasma p-tau181, (4) amyloid PET, and (5) plasma p-tau217. As an alternative analytic approach to the standard approach, we used a time-varying effect analysis that evaluates the association between biomarker-clock events and symptom onset on the 'age' time scale, avoiding the structural coupling between predictor and outcome. This analytic approach generates clinically relevant insights on the prognostic value of biomarker-clock events. Under simulated null scenarios in which the biomarker was generated independent of symptom onset, the standard analysis produced false-positive rates up to 100% and hazard ratios above 1, regardless of the true effect direction, whereas the time-varying analysis maintained type I error near the nominal 5%. Moreover, in analyses of both the BIOCARD and ADNI cohorts, the standard analysis produced uniformly significant associations for ages at biomarker-clock events, based on all five biomarkers (hazard ratios 1.94-3.34, all P < 0.01), comparable to the pattern predicted by the countdown paradox and reported in the literature. The time-varying analysis showed a different pattern for the effect of age at biomarker-clock events: for all biomarkers investigated, a younger age at biomarker-clock events is associated with a higher hazard for symptom onset on the age scale, conveying the opposite prognostic message implied by the standard analysis. These findings suggest that the standard biomarker-clock analysis may generate inaccurate associations and even reverse the apparent direction of the age effect, inverting the resulting prognostic message. A time-varying effect analysis avoids this by relating the age at a biomarker-clock event to clinical onset, with important implications for interpreting prior biomarker-clock studies.
Large language models (LLMs) are increasingly embedded in medical education and clinical care settings, yet contemporary Canadian data describing medical students' use and perceptions remain limited. To quantify the prevalence, frequency, and patterns of LLM use among medical students in Canada; to characterize perceptions of utility, accuracy, limitations, and impact; and to describe perceived barriers, challenges, and ethical/privacy concerns. We conducted a national, cross-sectional survey distributed to English-speaking medical students between November and December 2025. Recruitment occurred through medical school channels, student unions, and national/regional student organizations. Among 286 respondents from 10 medical schools, 96.50% reported using at least one LLM. The most commonly used LLMs were ChatGPT (93.36%) and OpenEvidence (57.69%). Daily/weekly use was most frequent for coursework assistance (60.22%) and clinical questions (57.14%). Most respondents reported positive impacts on efficiency (81.62%), learning (77.01%), and academic performance (59.49%). Students commonly reported encountering inaccurate information (90.18%). Formal instruction on LLM use was uncommon (10.95%), though 67.67% of students agreed medical schools should integrate formal instruction on LLMs. Only 21.43% of respondents felt adequately educated on data privacy regulations applicable to these tools. While LLM use among surveyed medical students in Canada was nearly universal and perceived favorably, students reported exposure to inaccurate outputs and substantial gaps in formal training and privacy literacy. These findings support the development of structured curricular guidance on appropriate application of these tools, including information verification practices and ethical, privacy-aware engagement.
The Risk-stratification of Emergency Department suspected Sepsis (REDS) score requires external validation. The six dichotomous variables and lactate can be accurately extracted electronically unlike refractory hypotension (RH), where the recording of the fluid-bolus and blood pressure measurement may be inaccurate. The impact of excluding RH is not known. To develop the quick-REDS (qREDS) score, by removing RH from the REDS score and to validate its prognostic performance in comparison to the original score and the National Early Warning Score (NEWS)2 score, for all-cause in-hospital mortality and 168-hour survival probability (SP). Vital signs, blood results, outcome at 168 hours and at discharge, were extracted from the electronic records of emergency department (ED) adults admitted after intravenous antibiotics for an infection. The REDS, qREDS and NEWS2 scores were calculated, receiver operating characteristic (ROC) curves constructed for in-hospital all-cause mortality, the area under the ROC (AUROC) curves compared and cut-off points noted. Test-characteristics at cut-off points ≥1, ≥3, ≥5 and ≥7 were studied. Kaplan-Meier (KM) curves were constructed for all the scores for SP at 168 hours for score bands 0-2, 3-4 and ≥5. Of 3202 patients, 433 died in hospital and 209 died at 168 hours. AUROC curve: REDS 0.74 (95% CI 0.72 to 0.75) and qREDS 0.73 (95% CI 0.71 to 0.75); both greater than NEWS2 0.66 (95% CI 0.65 to 0.68); p<0.0001. Cut-off points: REDS ≥3, qREDS ≥3 and NEWS2 ≥6. Specificity for in-hospital mortality of score ≥5; REDS 89.2% (95% CI 88.0% to 90.3%), qREDS 90.6% (95% CI 89.4% to 91.6%) and NEWS2 43.9% (95% CI 42.1% to 45.8%). All KM curves were significant, log-rank test p<0.0001. SP at 168 hours: whole population 93.4% (SE 0.4); SP at ≥5points: REDS 78.3% (SE 1.2), qREDS 78.0% (SE 2.0) and NEWS2 90.6% (SE 0.7). SP trends were similar at 24, 48 and 72 hours. qREDS is a valid surrogate for the REDS score and had significantly better prognostic performance than the NEWS2 score for in-hospital mortality and 168-hour survival probability, in ED patients with suspected sepsis.
Universal lesion detection (ULD) in computed tomography plays an essential role in computer-aided diagnosis. However, images in ULD tasks often exhibit substantial variations in quality, including issues such as limited clarity and inaccurate labels. Effectively leveraging training images with heterogeneous qualities thus becomes a significant challenge. Existing training strategies, such as self-paced learning (SCL) and hard example mining (OHEM), attempt to address this by reweighting high-loss samples. However, they often rely solely on loss, which inadequately captures sample difficulty, and may lead to over- or under-utilization of hard samples. To tackle this, we revisit the role of minibatch sampling (MBS) and propose a novel Mixed-order Minibatch Sampling (MoMBS) approach. MoMBS introduces a joint measure based on both loss and uncertainty, moving beyond sole reliance on loss. This enables a finer-grained categorization of high-loss samples by distinguishing between those that are poorly labeled and underrepresented, versus those that are overfitted but well represented. Motivated by human learning behavior, we prioritize underrepresented samples as the main contributors to the minibatch gradient and protect them from being overwhelmed by noisy or overfitted examples via a mixed-order sampling design. This leads to a more accurate estimation of sample difficulty and avoids indiscriminate treatment in the utilization of hard examples. Our primary experiments on DeepLesion for ULD show that MoMBS improves over two state-of-the-art (SOTA) methods across three different dataset settings by 0.97%-7.28%. To further verify its generalizability, we evaluate MoMBS on three additional tasks: Seg-19 (up to 2.3% improvement over five SOTA baselines), CIFAR100-LT (up to 5.3% over nine SOTA methods), and CIFAR100-NL (up to 4.6% over five SOTA methods), consistently outperforming existing approaches. Code: https://github.com/lhlm1994/MoMBS.
Current birth-size reference charts for nursery use may inaccurately categorize newborn size, impacting care. We aimed to develop new charts using data from uncomplicated newborns admitted to US nurseries. We analyzed de-identified data from newborns admitted to US normal newborn nurseries between 2021 and 2024. Statistical outliers were excluded, and gestational age was recorded in days. Birth weight, length, head circumference, and BMI models were developed by a randomly selected subset and validated by the remaining data. Proportions of newborns below the 10th percentile and above the 90th percentile were compared with the 2010 Olsen and 2013 Fenton charts. Uncomplicated newborns born between 35.0 and 41.6 weeks of gestation were included. The new charts did not exhibit overfitting bias and more accurately identified the expected 10% of newborns below the 10th and above the 90th percentiles compared to existing charts. New charts (weight, length, head circumference, BMI) improve birth-size referencing in normal newborn nurseries.
Background: U.S. state licensing restrictions requiring written practice agreements for certified nurse-midwives (CNMs) lack evidence of improved patient safety and instead contribute to maternity care workforce shortages. The objective of this study was to examine restrictive state licensing and practice policies for CNMs to identify the problem they were intended to solve. Methods: This was a qualitative content analysis of state policies that require CNMs to obtain written practice agreements, collected in January 2022, using the "What's the Problem Represented to be" framework. Analysis was completed by two midwife researchers with interpretation assistance from a panel of expert midwives. Results: Variation in written practice agreement requirements across states demonstrated that the regulations echoed historical physician complaints about the "midwife problem" first articulated in 1912. Effects of this problemization are inaccurate workforce data and persistent maternity care shortages. The problemization is maintained by the physician organization's advocacy against midwifery autonomy. The COVID-19 pandemic policy modifications demonstrated that states could eliminate restrictions when prioritizing access without compromising safety. Conclusions: Restrictive CNM licensing regulations perpetuate professional gatekeeping rather than ensuring public safety, directly contributing to maternity care shortages. State legislatures should eliminate written practice agreement requirements for CNMs and adopt evidence-based criteria focused on practitioner qualifications rather than supervisory relationships. National nursing organizations should advocate for regulatory reform, prioritizing healthcare access. Removing these barriers can expand the maternity care workforce and improve access to reproductive healthcare, particularly in underserved communities.
Renewable energy sources are expanding but may negatively affect wildlife in complex ways. Raptors are susceptible to wind turbines; collisions with turbines are well-studied, but the potential for spatial displacement has received less attention. We studied the effects of wind turbines on the distribution and abundance of raptors at a renowned wintering and migration stopover site on Amherst Island, Ontario, Canada. We used standardized surveys involving 3,284 observations of raptors to record their presence and precise locations during winter and spring migration for three years before and three years after the windfarm was built, incorporating both spatial and temporal controls. We found no evidence that wind turbines impacted how Northern Harrier (Circus hudsonicus), Bald Eagle (Haliaeetus leucocephalus), Red-tailed Hawk (Buteo jamaicensis), Rough-legged Hawk (Buteo lagopus), Snowy Owl (Bubo scandiacus), or American Kestrel (Falco sparverius) used Amherst Island during winter and spring migration. Similarly, we found no evidence for declines in abundance of these species, or of Short-eared Owl (Asio flammeus) and Northern Shrike (Lanius borealis), following turbine construction. Our results illustrate how spatial controls, in the absence of temporal controls, can lead to inaccurate assessments of turbine impacts for species that use habitats differentially.
The relevance of the study is as follows. Traditional methods of pedagogy do not allow objective and continuous monitoring of the emotional state of each student. The teacher usually relies on subjective observations - facial expressions behavior in a group - which may be inaccurate. At the same time modern technologies provide new opportunities: cameras and sensors backed by AI algorithms are able to automatically recognize facial expressions voice intonations and other signs of an emotional state. There are already commercial solutions for facial emotion recognition which have been used in marketing and security. Their potential application in education is an important topic as automated emotion analysis could provide educators with objective data on how engaged a group is whether students are overworked and who is experiencing difficulties or stress. It is the problem of mass learning fatigue and emotional burnout according to empirical studies that makes the introduction of such systems modern and practically significant and is the topic of this article. The article notes that the introduction of AI systems that can monitor both behavioral and physiological indicators opens up new opportunities for objective registration of cognitive states. For example sensors and wearable devices can measure heart rate heart rate variability brain activity and other parameters related to the level of concentration and mental stress. Computer vision can track eye gaze direction blinking frequency and posture which indirectly indicates the level of attention or fatigue. The study presents the results of a psychometric testing study using validated tools for emotional involvement in the educational process and a study of the factors of burnout and knowledge fatigue among full-time students at the Novorossiysk branch of the Financial University under the Government of the Russian Federation. The sample consisted of 125 respondents. Традиционные методы педагогики не позволяют объективно и непрерывно отслеживать эмоциональное состояние каждого студента. Преподаватель обычно опирается на субъективные наблюдения — выражения лиц поведение в группе — которые могут быть неточными. Современные технологии дают новые возможности: камеры и датчики подкреплённые алгоритмами ИИ способны автоматически распознавать мимику голосовые интонации и другие признаки эмоционального состояния. Уже существуют коммерческие решения для распознавания эмоций по лицу которые применяются в маркетинге и безопасности. Проблема массовой учебной усталости и эмоционального выгорания делает внедрение подобных систем современным и практически значимым и является темой данной статьи. В статье отмечено что внедрение ИИ-систем способных проводить мониторинг как поведенческих так и физиологических индикаторов открывает новые возможности для объективной регистрации когнитивных состояний. Например датчики и носимые устройства могут измерять частоту сердечных сокращений вариабельность сердечного ритма активность мозга и другие параметры связанные с уровнем концентрации и умственного напряжения. Компьютерное зрение способно отслеживать направление взгляда частоту миганий позу что косвенно свидетельствует об уровне внимания или утомления. В исследовании приведены результаты проведённого исследования психометрического тестирования с использованием валидизированных инструментов эмоциональной включённости в образовательный процесс и исследование факторов выгорания усталости от знаний студентов очной формы обучения Новороссийского филиала Финансового университета при Правительстве РФ. Выборка составила 125 респондентов.
There is a broad consensus that forensic tests for the prediction of externally visible characteristics (EVC) and analysis of biogeographic ancestry (BGA) of an individual are technically reliable. However, interpretation of the results and population-specific genotype distribution patterns remains challenging. EVC and BGA analyses provide valuable information for population genetics studies and as investigative leads for criminal cases, as well as for historical and contemporary identification tests. However, inaccurate or incorrect predictions, for example, from subjective bias in the interpretations made, have the potential to misdirect police investigations. The legal situation regarding EVC and BGA testing varies by country: ranging from countries where it is explicitly prohibited, to those without specific regulations on biogeographic ancestry prediction, and others that have already enacted laws governing its use. The reluctance to utilize these analyses is not only due to legal restrictions and data protection concerns, but also to initial limited sets of sufficiently comprehensive forensic DNA assays. Forensic BGA marker panels typically contain up to ∼300 SNPs. This relatively small number of genetic markers, along with limited reference population data, complicates the interpretation of results from donors of unknown origin. This paper presents the results of a collaborative EDNAP study, which, for the first time, evaluated the approach to reporting EVC and BGA data between international laboratories. For the study, DNA from nine individuals with self-reported ancestry was collected and analysed using various forensic panels differing in the number and composition of ancestry-informative markers genotyped, comprising: the Precision ID mtDNA Whole Genome Panel, the VISAGE Basic Tool and the VISAGE Enhanced Tool for Appearance and Ancestry Prediction, and the Ion AmpliSeq™ PhenoTrivium Panel. To ensure full data protection, all SNP genotypes and uniparental marker haplotypes obtained were not shared with third parties. Instead, the genetic data were analysed using a range of commonly used population analysis software packages. These analysis outcomes were then distributed to twelve European forensic laboratories (both academic and law enforcement institutions), who were asked to prepare reports based on their interpretation of the phenotypes and ancestry they inferred from the analysis data. A questionnaire sent alongside the genetic information, aimed to evaluate which difficulties were encountered by the participants in processing the BGA analysis data they were given.
Quasi-elastic neutron scattering (QENS) probes atomic and molecular motion on length and time scales central to catalysis, energy materials, and gas adsorption. However, conventional analytical fitting of QENS spectra often fails to uniquely determine the underlying dynamics. The flexibility of simplified line-shape models can make spectra generated by distinct physical processes statistically indistinguishable, leading to an ambiguous or inaccurate mechanistic interpretation. By integrating molecular dynamics simulations, physically derived Q-dependent scattering models, Bayesian model discrimination, and polarization analysis, we demonstrate that QENS can, for the first time, resolve anisotropic rotational motion in liquid benzene, a prototypical aromatic molecule relevant to microporous catalysis. The extracted spinning and tumbling diffusion coefficients suggest stronger anisotropy than previously recognized. This integrated, Bayesian evidence-based analytical framework defines a new paradigm for QENS, enabling direct resolution of the rotational and translational dynamics that govern molecular interactions and transport: the fundamental processes and rate-limiting steps in confined hydrocarbon catalysis.
Text-to-video (T2V) generative artificial intelligence (AI) models can produce short videos from natural-language prompts, creating new opportunities for medical visualization and communication. In pediatrics, visual explanations and preparatory media are already used to support learning and to reduce distress during procedures; however, current T2V systems (including proprietary models such as OpenAI's Sora) have not been clinically validated for routine pediatric care. To provide a forward-looking perspective on how T2V generative AI could be used in pediatric education, patient and caregiver communication, and simulation, while clearly distinguishing current evidence from hypothetical applications. We performed a narrative (nonsystematic) review of the literature on (i) generative AI and T2V models in healthcare and (ii) pediatric video-based education, procedural preparation, and digital distraction (PubMed and Google Scholar; search through June 2025; representative keywords included "pediatric" AND "video-based education," "procedural preparation," "anxiety," "pain," and "simulation"). Potential applications include scenario-based training videos for clinicians, tailored educational animations for families, and child-friendly preparatory videos to reduce anxiety and pain. Key limitations and risks include inaccurate or oversimplified content, bias, privacy and data-governance concerns, unclear accountability, commercialization pressures, and inequitable access. Pediatric-specific safeguards (transparent disclosure, clinician review, age-appropriate design, consent/assent practices, and equity planning) are essential. T2V generative AI should be framed as an experimental adjunct rather than a near-term clinical solution. Rigorous validation, governance, and equity-focused implementation strategies are needed before pediatric deployment.
Accurate X-ray computed tomography (CT) image segmentation of the abdominal organs is a key task in automated medical image analysis, with crucial applications in clinical decision-making, computer-aided diagnosis, and surgical planning. However, existing methods still face significant challenges: insufficient capability in modeling long-range contextual dependencies, hindering the adaptability to the complicated morphological variations and spatial relationships of abdominal organs; and inaccurate boundary segmentation due to blurred edges and irregular anatomical structures, particularly in regions with high tissue adhesiveness. To address these issues, we propose an efficient abdominal multi-organ segmentation model, EEA-UNet. Specifically, we design an efficient element-wise adaptive (EEA) attention mechanism integrated into the skip connections to enhance inter-organ feature interactions while maintaining computational efficiency. This module effectively expands the receptive field, improving long-range dependency modeling. An enhanced multi-scale feature fusion (EMF) module is introduced to strengthen decoding capability, coupled with an edge-awareness composite loss function to optimize segmentation accuracy for small organs and boundary regions.Experimental results on the Synapse dataset demonstrate the competitive performance of EEA-UNet, achieving a Dice score of 84.45% and an HD95 of 0.16. Our method demonstrates a favorable trade-off between segmentation accuracy and computational efficiency, showing improved results compared with several existing approaches in both visual comparison and quantitative metrics.
Generative artificial intelligence (AI) tools are increasingly being used by patients seeking cancer-related information, creating new opportunities for accessible and personalized cancer education. Large language models can simplify complex medical concepts, improve access to educational resources, and support patient engagement. However, these benefits are accompanied by growing concerns regarding misinformation, hallucinatory content, outdated recommendations, and the potential for harmful health decisions. As AI-generated information becomes more integrated into cancer education, an important ethical question emerges: who is responsible when AI provides inaccurate cancer information? This commentary examines the shared responsibilities of patients, healthcare professionals, healthcare organizations, and AI developers in ensuring the safe use of AI-generated cancer information. This commentary argues that accountability should not rest with a single stakeholder but instead be viewed as a shared responsibility across the cancer education ecosystem. The commentary further argues that AI literacy should become an essential component of modern cancer education to support informed decision-making and safeguard patient well-being.
Intestinal metaplasia (IM) is a precancerous gastric lesion that commonly develops in the setting of chronic gastric inflammation and represents a key step in gastric carcinogenesis. Delayed or inaccurate identification of IM may delay appropriate surveillance and clinical management, potentially increasing the risk of progression to gastric cancer. In this study, we propose a deep learning-based framework, termed CNXTGeM, for the automated detection of intestinal metaplasia in hematoxylin and eosin (H&E)-stained gastric histopathology images. The model integrates a ConvNeXt-Tiny backbone with Generalized Mean (GeM) pooling and Efficient Channel Attention (ECA) to enhance feature representation and discrimination of histopathological patterns associated with intestinal metaplasia. The framework was evaluated using 1,037 H&E-stained gastric biopsy samples (516 IM and 521 controls) obtained from Elazığ Fethi Sekin City Hospital, with an 80/10/10 stratified patient-level train/validation/test split to prevent data leakage. External validation was further performed using the publicly available GasHisSDB dataset (33,284 image patches). Model interpretability was assessed using three complementary gradient-based visualization techniques: Grad-CAM, Grad-CAM++, and XGrad-CAM. CNXTGeM outperformed the evaluated baseline deep learning models, including VGG16, VGG19, DenseNet121, and MobileNetV2, achieving an accuracy of 99.04%, precision of 98.08%, specificity of 98.11%, and an F1-score of 99.03%. Notably, the proposed framework achieved 100% sensitivity, representing an 8.51% improvement in recall over the baseline ConvNeXt model, which may help reduce missed IM cases in computer-assisted histopathological assessment. On the external GasHisSDB dataset, CNXTGeM maintained robust performance (accuracy = 99.34%, F1-score = 99.31%), suggesting good generalization to an independent external dataset. Gradient-based visualization analyses (Grad-CAM, Grad-CAM++, and XGrad-CAM) indicated that the model consistently focused on histopathological regions relevant to inflammation-related mucosal alterations. The proposed CNXTGeM framework demonstrates the potential to provide a reliable, efficient, and interpretable artificial intelligence-based approach for computer-assisted detection of intestinal metaplasia. By accurately identifying inflammation-associated histopathological features, the model supports computer-assisted histopathological assessment, reduces inter-observer variability, and may facilitate digital pathology workflows for the assessment of intestinal metaplasia.
MRI is the modality of choice for detection, characterization, and noninvasive diagnosis of focal liver observations. It offers excellent tissue contrast using multiple sequences and postcontrast multiphasic imaging. However, its diagnostic accuracy depends on appropriate selection of pulse sequences and imaging parameters. Poorly acquired images or suboptimal sequence parameter selection can lead to medical errors, such as missed lesions and inaccurate characterization of focal liver observations. A solid understanding of applied MRI physics can help optimize image quality and prevent such errors. In this educational review, the authors aim to (a) guide radiologists in assessing MRI quality, (b) promote effective communication using appropriate terminology when collaborating with technologists and medical physicists, and (c) establish a reference liver MRI protocol. First, they detail metrics of MRI quality-contrast-to-noise ratio, spatial resolution, and signal-to-noise ratio-along with the impact of key technical parameters on acquisition time and step-by-step guidance for improving image quality. Second, they review the range of pulse sequences used in liver MRI, discussing their respective strengths and limitations in achieving optimal imaging quality. Third, they address key ancillary considerations for performance of liver MRI, including fat suppression techniques, respiratory motion suppression strategies, and acceleration methods. ©RSNA, 2026 Supplemental material is available for this article.
Human immunodeficiency virus (HIV) remains a major public health challenge, particularly in sub-Saharan Africa, where achieving sustained HIV virologic suppression continues to be difficult despite expanded access to antiretroviral therapy (ART). This challenge is further complicated among people living with HIV-HBV coinfection due to increased disease complexity, treatment burden, and adherence-related barriers. Health education plays an important role in improving adherence and virologic outcomes; however, its implementation remains inconsistent in resource-limited settings. This study explored barriers and enablers to implementing health education for HIV virologic suppression among adults living with HIV-HBV coinfection in Northwest Ethiopia. This study aimed to explore the barriers and enablers influencing the implementation of health education for HIV virological suppression among adults living with HIV-HBV co-infection in North-West Ethiopia. A facility-based interpretive qualitative study was conducted from November 13, 2025, to January 12, 2026, at the University of Gondar Comprehensive Specialized Hospital and Felege Hiwot Comprehensive Specialized Hospital in Northwest Ethiopia. A total of 28 purposively selected participants, including adult people living with HIV-HBV coinfection, healthcare providers, and HIV program managers, participated in in-depth interviews (IDIs), key informant interviews (KIIs), and focus group discussions (FGDs). Data were audio-recorded, transcribed verbatim, translated into English, and analyzed using reflexive thematic analysis with a deductive approach guided by the Consolidated Framework for Implementation Research (CFIR) using MAXQDA Analytics Pro 2024. Trustworthiness was ensured through triangulation, member checking, peer debriefing, and maintaining an audit trail. Ten barriers and eight enablers were identified across all CFIR domains. In addition to inaccurate information or graphics and transportation issues, the main reported hurdles included patient-level issues such as inadequate literacy, misconceptions about ART or viral load, anxiety about being seen attending sessions, and limited patient engagement in instructional design. Staff shortages, lack of commitment to counselling duties, a focus on ART refills rather than education, irregular or shortened sessions, lack of time, and the absence of a dedicated budget were all organizational and system-level barriers. Conversely, teamwork, clearly defined staff roles, the use of viral load monitoring tools, frequent review meetings, alignment with national HIV guidelines, patient motivation to achieve viral suppression, trust in provider advice, and improved adherence linked to education were all perceived enablers. Implementation of health education for HIV virological suppression among HIV-HBV co-infected patients is hindered by key barriers, particularly low patient literacy, misconceptions, stigma-related concerns, and systemic challenges such as staff shortages, limited time, and inadequate resources. However, strong teamwork, patient motivation, provider trust, and alignment with national guidelines serve as important enablers. Findings from this study informed the updating of the existing health education intervention to better address identified barriers and strengthen enabling factors. Testing the revised intervention for feasibility and acceptability is essential before wider implementation to improve adherence and achieve sustained HIV virological suppression in this population.
Electrical impedance tomography (EIT) provides an attractive solution for large-area tactile sensing due to its minimal wiring and geometric flexibility, but its nonlinear inverse problem often leads to severe artifacts and inaccurate reconstruction. This work presents PhyDNN, a physics-driven deep reconstruction framework that embeds the EIT forward model directly into the learning objective. By jointly minimizing the discrepancy between predicted and ground-truth conductivity maps and enforcing consistency with the forward PDE, PhyDNN reduces the black-box nature of deep networks and improves both physical plausibility and generalization. To enable efficient backpropagation, we design a differentiable forward-operator network that accurately approximates the nonlinear EIT response, allowing fast physics-guided training. Extensive simulations and real tactile experiments on a 16-electrode soft sensor demonstrate that PhyDNN consistently outperforms NOSER, TV, and standard DNNs in reconstructing contact shape and location. The proposed method yields fewer artifacts, sharper boundaries, and higher quantitative scores, demonstrating its effectiveness for high-quality tomographic tactile sensing.
Self-assessment is a key requirement for lifelong learning in medicine. Evidence from gender-related research indicates that important moderators affecting self-assessment are influenced by gender. Therefore, systematic gender differences in the accuracy of self-assessment may be assumed. The present study aimed to examine gender differences in medical students' self-assessment. Specifically, this study addressed two research questions: (1) Are there systematic gender differences in medical students' self-assessment accuracy? (2) What is the magnitude of these gender differences when accounting for academic progress and knowledge? Medical students from 3 cohorts at Medical School OWL were surveyed in 3 waves between April 2023 and April 2024 during the Progress Test Medicine. Prior to taking the test, students were asked to indicate the percentage of Progress Test Medicine questions that they expected to answer correctly in 5 knowledge areas. Self-assessment accuracy was calculated as the difference between the subjective self-assessment and the objective test score. Linear mixed models were used to analyze the influence of gender on students' self-assessment accuracy while accounting for academic progress and knowledge. A total of 165 students with 404 data points participated in this study (269/404, 66.6% women; 135/404, 33.4% men; mean age 21.96, SD 3.61 years). Across all models, women rated themselves significantly less accurately than men (adjusted P values ranged from <.001 to .01). The observed gender effect ranged from -6.08 to -3.74 percentage points. The results indicated systematic gender differences in medical students' self-assessment in favor of men, with a magnitude comparable to the average knowledge acquired in an entire semester of study. In view of the potentially negative consequences of inaccurate self-assessment, targeted support for realistic self-assessment during medical studies may be particularly beneficial for women.