Search Results (236)

Adverse drug events (ADEs) significantly impact patient safety and health outcomes. Manual ADE detection from clinical narratives is time-consuming, labor-intensive, and costly. Recent advancements in large language models (LLMs), including transformer-based architectures such as Bidirectional Encoder Representations from Transformers (BERT) and Generative Pretrained Transformer (GPT) series, offer promising methods for automating ADE extraction from clinical data. These models have been applied to various aspects of pharmacovigilance and clinical decision support, demonstrating potential in extracting ADE-related information from real-world clinical data. Additionally, chatbot-assisted systems have been explored as tools in clinical management, aiding in medication adherence, patient engagement, and symptom monitoring. This narrative review synthesizes the current state of LLMs in ADE detection from a clinical perspective, organizing studies into categories such as human-facing decision support tools, immune-related ADE detection, cancer-related and non-cancer-related ADE surveillance, and personalized decision support systems. In total, 39 articles were included in this review. Across domains, LLM-driven methods have demonstrated promising performances, often outperforming traditional approaches. However, critical limitations persist, such as domain-specific variability in model performance, interpretability challenges, data quality and privacy concerns, and infrastructure requirements. By addressing these challenges, LLM-based ADE detection could enhance pharmacovigilance practices, improve patient safety outcomes, and optimize clinical workflows. Full article

As an emerging information technology, Wireless Body Area Networks (WBANs) provide a lot of convenience for the development of the medical field. A WBAN is composed of many miniature sensor nodes in the form of an ad hoc network, which can realize remote medical monitoring. However, the data transmission between sensor nodes in the WBAN not only consumes the energy of the node but also causes the temperature of the node to rise, thereby causing human tissue damage. Therefore, in response to the energy consumption problem in the Wireless Body Area Network and the hot node problem in the transmission path, this paper proposes a temperature state awareness-based energy-saving routing protocol (TSAER). The protocol senses the temperature state of nodes and then calculates the data receiving probability of nodes in different temperature state intervals. A benefit function based on several parameters such as the residual energy of the node, the distance to sink, and the probability of receiving data was constructed. The neighbor node with the maximum benefit function was selected as the best forwarding node, and the data was forwarded. The simulation results show that compared with the existing M-ATTEPMT and iM-SIMPLE protocols, TSAER effectively prolongs the network lifetime and controls the formation of hot nodes in the network. Full article

South Africa’s bioeconomy strategy identifies bio-innovation as a key driver of economic growth and social development, with plant breeding playing a central role in improving food security through the development of high-yielding, resilient, and high-quality crops. However, consumer perceptions of recent advances, particularly new breeding techniques (NBTs), remain underexplored. This study examines South African consumer attitudes towards plant breeding innovations, using a mixed-methods approach. The initial focus group interviews informed the development of a structured quantitative survey examining familiarity, perceptions, and acceptance of plant breeding technologies. Consumer awareness of plant breeding principles was found to be limited, with 67–68% of respondents unfamiliar with both conventional and modern plant breeding procedures. Despite this information gap, consumers expressed conditional support for modern breeding techniques, especially when associated with actual benefits like increased nutritional value, environmental sustainability, and crop resilience. When favourable effects were outlined, support for general investment in modern breeding practices climbed from 45% to 74%. Consumer purchase decisions emphasised price, product quality, and convenience over manufacturing techniques, with sustainability ranked last among the assessed factors. Trust in the sources of food safety information varied greatly, with medical experts and scientists being ranked highly, while government sources were viewed more sceptically. The results further suggest that targeted education could improve customer confidence, as there is a significant positive association (R² = 0.938) between familiarity and acceptance. These findings emphasise the significance of open communication strategies and focused consumer education in increasing the adoption of plant breeding breakthroughs. The study offers useful insights for policymakers, researchers, and industry stakeholders working on engagement strategies to facilitate the ethical growth and application of agricultural biotechnology in support of food security and quality in South Africa. This study contributes to a better understanding of South African consumers’ perceptions of plant breeding innovations and food safety. The research findings offer valuable insights for policymakers, researchers, and industry stakeholders in developing effective engagement and communication strategies that address consumer concerns and promote the adoption of products derived from diverse plant breeding technologies. Full article

Neglecting preventive healthcare policies has contributed to the global surge in chronic diseases, increased hospitalizations, declining quality of care, and escalating costs. Non-communicable diseases (NCDs)—notably cardiovascular conditions, diabetes, and cancer—consume over 80% of healthcare expenditure and account for more than 60% of global deaths, which are projected to exceed 75% by 2030. Poor diets, sedentary lifestyles, regulatory loopholes, and underfunded public health initiatives are driving this crisis. Compounding the issue are flawed policies, congressional lobbying, and conflicts of interest that prioritize costly, hospital-based, symptom-driven care over identifying and treating to eliminate root causes and disease prevention. Regulatory agencies are failing to deliver their intended functions. For instance, the U.S. Food and Drug Administration’s (FDA) broad oversight across drugs, devices, food, and supplements has resulted in inefficiencies, reduced transparency, and public safety risks. This broad mandate has allowed the release of unsafe drugs, food additives, and supplements, contributing to the rising childhood diseases, the burden of chronic illness, and over-medicalization. The author proposes separating oversight responsibilities: transferring authority over food, supplements, and OTC products to a new Food and Nutraceutical Agency (FNA), allowing the FDA to be restructured as the Drug and Device Agency (DDA), to refocus on pharmaceuticals and medical devices. While complete reform requires Congressional action, interim policy shifts are urgently needed to improve public health. Broader structural changes—including overhauling the Affordable Care Act, eliminating waste and fraud, redesigning regulatory and insurance systems, and eliminating intermediaries are essential to reducing costs, improving care, and transforming national and global health outcomes. The information provided herein can serve as a White Paper to help reform health agencies and healthcare systems for greater efficiency and lower costs in the USA and globally. Full article

Background/objectives: Social media has significantly enhanced access to medical knowledge by enabling rapid information sharing. With YouTube being the second-most popular website, we intended to evaluate the quality of its content as a source of information for patients and relatives for information about cerebral palsy. Methods: The first 30 videos for search terms “Cerebral palsy”, “Spastic cerebral palsy”, “Dyskinetic cerebral palsy”, “Worster-Drought syndrome”, and “Ataxic cerebral palsy” were selected for inquiry. Out of 150 films, a total of 83 were assessed with a mixed method approach by two independent raters utilizing evidence-based quality scales such as Quality Criteria for Consumer Health Information (DISCERN), the Journal of the American Medical Association instrument (JAMA), and the Global Quality Score (GQS). Furthermore, audience engagement was analyzed, and the Video Power Index (VPI) was calculated for each video. Results: The mean total DISCERN score excluding the final question (subjective assessment of the video) was 30.5 ± 8.7 (out of 75 points), implying that the quality of the videos was poor. The global JAMA score was 2.36 ± 0.57 between the raters. The mean GQS score reached 2.57 ± 0.78. The videos had statistically higher DISCERN scores when they included treatment options, risk factors, anatomy, definition, information for doctors, epidemiology, doctor as a speaker, and patient experience. Conclusions: YouTube seems to be a poor source of information for patients and relatives on cerebral palsy. The analysis can contribute to creating more engaging, holistic, and informative videos regarding this topic. Full article

Background and Objectives: Major trauma ranks among the leading causes of mortality and handicap in both developing and developed countries, consuming substantial healthcare resources. Its unpredictable nature and diverse clinical presentations often lead to rapid and challenging-to-predict changes in patient conditions. An increasing number of models have been developed to address this challenge. Given our access to extensive and relatively comprehensive data, we seek assistance in making a meaningful contribution to this topic. This study aims to leverage artificial intelligence (AI)/machine learning (ML) to forecast potential adverse effects in major trauma patients. Methods: This retrospective analysis considered major trauma patient admitted to Chi Mei Medical Center from 1 January 2010 to 31 December 2019. Results: A total of 5521 major trauma patients were analyzed. Among five AI models tested, XGBoost showed the best performance (AUC 0.748), outperforming traditional clinical scores such as ISS and GCS. The model was deployed as a web-based application integrated into the hospital information system. Preliminary clinical use demonstrated improved efficiency, interpretability through SHAP analysis, and positive user feedback from healthcare professionals. Conclusions: This study presents a predictive model for estimating recovery probabilities in severe burn patients, effectively integrated into the hospital information system (HIS) without complex computations. Clinical use has shown improved efficiency and quality. Future efforts will expand predictions to include complications and treatment outcomes, aiming for broader applications as technology advances. Full article

Chronic pain is a prevalent public health concern in the United States, frequently managed with over-the-counter (OTC) painkillers without professional medical supervision. This study investigates household-level patterns of over-the-counter painkiller use utilizing a nationally representative dataset from NielsenIQ, focusing on how reported health conditions, whether self-identified or professionally diagnosed, affect purchasing behaviors. By linking consumer purchase data with self-reported ailment information, this study analyzed painkiller expenditures across different ailment types and demographic groups. Results show that over-the-counter painkiller purchases were highly symptom-driven, particularly for headache-related products, which were the most frequently purchased category across all household types. Nearly one-third of single-member households purchased over-the-counter painkillers for headaches, regardless of diagnosis type, indicating a strong role of perceived need in driving behavior. Females and older individuals more frequently reported ailments, with consistently higher proportions across both pain-related and other conditions. Nonetheless, a notable share of households reported over-the-counter painkiller use without any reported ailments. The findings suggest that diagnostic status plays a limited role in determining over-the-counter painkiller usage, emphasizing the need for improved public health messaging around safe self-medication. These insights can inform targeted education, labeling regulations, and policy interventions to support safer and more equitable pain management practices at the population level. Full article

Consumers’ growing interest in sustainability and the consideration of purchasing second-hand products present conditions for developing and improving a new method for Remaining Value (RV) estimation. The remaining value refers to the value of an end-of-life product that has been inspected, repaired, if necessary, and prepared for resale. Through the literature review, the main blockers, trustworthiness, price, and quality, were identified as preventing consumers from purchasing used products. Trustworthiness could be ensured by evaluating used products in an automated and model-based manner. To enhance consumers’ confidence, this study proposes a novel framework to assess the remaining value of non-new products by incorporating the diagnostic test results, even in the absence of a gold standard for model comparison and evaluation. This research expands the application of the Hui–Walter method beyond medical diagnostics by adapting it to sustainability-focused estimation. The proposed framework is designed to assist consumers in making data-informed purchase decisions and support retailers in assessing the market price while contributing to the environmental pillar of sustainability by reducing waste and resource consumption and extending the product lifetime. This work aligns with the United Nations Sustainable Development Goals 12 (Responsible Consumption and Production) and 13 (Climate Action) by providing quantifiable methods to extend the product lifecycle and minimize electronic waste. While this study focuses on developing the theoretical framework, future work will apply and validate this framework using empirical case studies and compare it with the remaining value estimation models. Full article

Background/Objectives: The accurate and efficient segmentation of brain tumors from 3D MRI data remains a significant challenge in medical imaging. Conventional deep learning methods, such as convolutional neural networks and transformer-based models, frequently introduce significant computational overhead or fail to effectively represent multi-scale features. Methods: This paper presents a lightweight deep learning framework that uses adaptive discrete wavelet decomposition and iterative axial attention to improve 3D brain tumor segmentation. The wavelet decomposition module effectively captures multi-scale information by breaking it down into frequency sub-bands, thereby the mitigating detail loss often associated with standard downsampling methods. Ablation studies confirm that this module enhances segmentation accuracy, particularly in preserving the finer structural details of tumor components. Simultaneously, the iterative axial factorization attention reduces the computational burden of 3D spatial modeling by processing attention sequentially along individual axes, preserving long-range interdependence while consuming minimal resources. Results: Our model performs well on the BraTS2020 and FeTS2022 datasets with average Dice scores of 85.0% and 88.1%, with our competitive results using only 5.23 million parameters and 9.75 GFLOPs. In comparison to state-of-the-art methods, it effectively balances accuracy and efficiency, making it suitable for resource-constrained clinical applications. Conclusions: This study underscores the advantages of integrating frequency-domain analysis with optimized attention mechanisms, paving the way for scalable, high-performance medical image segmentation algorithms with broader clinical diagnostic applications. Full article

Moderate wine consumption is, generally, the focus of various medical studies, while consumer behaviour research does not specifically centres on moderation in wine consumption. Wine consumption in moderation is an important part of various healthy diets; still, consumers need to make informed choices when purchasing wine and the information printed on wine labels partially contributes to this process. Therefore, the main aims of this paper were to develop a scale for measuring perceptions of the health benefits associated with moderate wine consumption, and to test the effect of dietary habits and non-obligatory wine label information on the perception of the health benefits associated with moderate wine consumption. The data were collected on a sample of wine consumers who participated in an interdisciplinary experiment regarding the impact of moderate wine consumption on human health. Univariate, bivariate, and multivariate statistics were used. The consumers’ socio-demographic characteristics were used as a starting point in the analysis because they influence wine consumption. Gender was identified as a consistently important variable in predicting the perception of health benefits associated with moderate wine consumption. Health behaviour was a significant predictor along with gender, but after introducing non-obligatory wine label information, its significance in explaining the dependent variable was diminished. The results suggest that a consumer’s perception of the scale of moderate wine consumption is a unidimensional construct. Furthermore, the non-obligatory information on wine labels was identified and classified as either wine-related warnings or wine-related health benefits. Full article

Clinical documentation, particularly the hospital discharge report (HDR), is essential for ensuring continuity of care, yet its preparation is time-consuming and places a considerable clinical and administrative burden on healthcare professionals. Recent advancements in Generative Artificial Intelligence (GenAI) and the use of prompt engineering in large language models (LLMs) offer opportunities to automate parts of this process, improving efficiency and documentation quality while reducing administrative workload. This study aims to design a digital system based on LLMs capable of automatically generating HDRs using information from clinical course notes and emergency care reports. The system was developed through iterative cycles, integrating various instruction flows and evaluating five different LLMs combined with prompt engineering strategies and agent-based architectures. Throughout the development, more than 60 discharge reports were generated and assessed, leading to continuous system refinement. In the production phase, 40 pneumology discharge reports were produced, receiving positive feedback from physicians, with an average score of 2.9 out of 4, indicating the system’s usefulness, with only minor edits needed in most cases. The ongoing expansion of the system to additional services and its integration within a hospital electronic system highlights the potential of LLMs, when combined with effective prompt engineering and agent-based architectures, to generate high-quality medical content and provide meaningful support to healthcare professionals. Hospital discharge reports (HDRs) are pivotal for continuity of care but consume substantial clinician time. Generative AI systems based on large language models (LLMs) could streamline this process, provided they deliver accurate, multilingual, and workflow-compatible outputs. We pursued a three-stage, design-science approach. Proof-of-concept: five state-of-the-art LLMs were benchmarked with multi-agent prompting to produce sample HDRs and define the optimal agent structure. Prototype: 60 HDRs spanning six specialties were generated and compared with clinician originals using ROUGE with average scores compatible with specialized news summarizing models in Spanish and Catalan (lower scores). A qualitative audit of 27 HDR pairs showed recurrent divergences in medication dose (56%) and social context (52%). Pilot deployment: The AI-HDR service was embedded in the hospital’s electronic health record. In the pilot, 47 HDRs were autogenerated in real-world settings and reviewed by attending physicians. Missing information and factual errors were flagged in 53% and 47% of drafts, respectively, while written assessments diminished the importance of these errors. An LLM-driven, agent-orchestrated pipeline can safely draft real-world HDRs, cutting administrative overhead while achieving clinician-acceptable quality, not without errors that require human supervision. Future work should refine specialty-specific prompts to curb omissions, add temporal consistency checks to prevent outdated data propagation, and validate time savings and clinical impact in multi-center trials. Full article

Understanding and interpreting medical images, particularly radiology images, is a time-consuming task that requires specialized expertise. In this study, we developed a deep learning-based system capable of automatically assigning multiple standardized medical concepts to radiology images, leveraging deep learning models. These concepts are based on Unified Medical Language System (UMLS) Concept Unique Identifiers (CUIs) and describe the radiology images in detail. Each image is associated with multiple concepts, making it a multilabel classification problem. We implemented several deep learning models, including DenseNet121, ResNet101, and VGG19, and evaluated them on the ImageCLEF 2020 Medical Concept Detection dataset. This dataset consists of radiology images with multiple CUIs associated with each image and is organized into seven categories based on their modality information. In this study, transfer learning techniques were applied, with the models initially pre-trained on the ImageNet dataset and subsequently fine-tuned on the ImageCLEF dataset. We present the evaluation results based on the F1-score metric, demonstrating the effectiveness of our approach. Our best-performing model, DenseNet121, achieved an F1-score of 0.89 on the classification of the twenty most frequent medical concepts, indicating a significant improvement over baseline methods. Full article

Cell nuclei instance segmentation plays a critical role in pathological image analysis. In recent years, fully supervised methods for cell nuclei instance segmentation have achieved significant results. However, in practical medical image processing, annotating dense cell nuclei at the instance level is often costly and time-consuming, making it challenging to acquire large-scale labeled datasets. This challenge has motivated researchers to explore ways to further enhance segmentation performance under limited labeling conditions. To address this issue, this paper proposes a network based on category-adaptive sampling and attention mechanisms for semi-supervised nuclei instance segmentation. Specifically, we design a category-adaptive sampling method that forces the model to focus on rare categories and dynamically adapt to different data distributions. By dynamically adjusting the sampling strategy, the balance of samples across different cell types is improved. Additionally, we propose a strong–weak contrast consistency method that significantly expands the perturbation space. Strong perturbations enhance the model’s ability to discriminate key nuclei features, while weak perturbations improve its robustness against noise and interference. Furthermore, we introduce a region-adaptive attention mechanism that dynamically assigns higher weights to key regions, guiding the model to prioritize learning discriminative features in challenging areas such as blurred or ambiguous cell boundaries. This improves the morphological accuracy of the segmentation masks. Our method effectively leverages the potential information in unlabeled data, thereby reducing reliance on large-scale, high-quality labeled datasets. Experimental results on public datasets demonstrate the effectiveness of our approach. Full article

Background: Despite the widespread use of precautionary allergen labelling (PAL) by manufacturers, PAL is not always used consistently and can be a source of misinterpretation by consumers and allergists. Although its use is not specifically regulated in Canada, some voluntary guidelines exist. The aims of this study were to investigate allergic consumers’ and clinicians’ understanding of PAL, to describe consumers’ attitudes towards products with PAL, and to examine recommendations given by clinicians to their patients about these products. We also compared two groups of consumers enrolled in this study, since the majority of them (72%) were registered in the Food Allergy Canada database and the others (28%) came from representative consumers of the general population. Methods: An online survey was sent from 2 to 28 December 2021 to allergic consumers registered with Food Allergy Canada’s database and to a group of allergic consumers extracted from a panel representative of the general population and not registered with Food Allergy Canada (third-party panel). All consumer participants had a food allergy or were a parent/caregiver of a child with food allergy and had to be diagnosed by a medical professional. Considering that consumers registered via the Food Allergy Canada database could be more informed about labelling than the third-party consumer panel, we conducted a multivariate analysis (logistic regression) on the key variables related to PAL allowing to compare these two groups of participants. In addition, a separate online survey was sent to allergist members of the Canadian Society of Allergy and Clinical Immunology and provincial associations to investigate their understanding of PAL from 12 November 2021 to 16 January 2022. Results: A total of 1080 consumers and 63 allergists (29% of allergists in Canada) responded to the surveys. Fifty percent of consumers were adults with food allergy, and 50% were a parent/caregiver of a child with food allergy. Food allergy was diagnosed most commonly by an allergist in 76% of the cases. Fifty-four percent of consumers purchased products with a PAL statement at least occasionally, and more than half of consumers (53%) considered PAL a very useful tool. Most surveyed individuals (59%) had not heard of the term “individual allergen threshold” or had heard the term but did not know what it meant. The same allergic consumers were reluctant to buy food products with even a small amount of their allergen (i.e., a dose that would not trigger an allergic reaction in the vast majority of them). Half of allergists reported PAL was not useful in its current form, and 83% supported the consumption of foods with PAL to their patients in some circumstances. Conclusion: While most consumers are somewhat confident in the accuracy of ingredient information on pre-packaged foods, interpretation of PAL remains confusing by many allergic consumers. If changes are to be made based on allergen thresholds, a multi-stakeholder approach will be required with greater consumer and allergist education on risk assessment concepts to facilitate the implementation of allergen population thresholds into the application of PAL. Full article

Background/Objectives: YouTube is an open-access video streaming platform with minimal regulation which has led to a vast library of unregulated medical videos. This study assesses the quality of information, understandability and actionability of videos on YouTube pertaining to sacral neuromodulation (SNM). Methods: The first 50 videos on YouTube after searching “sacral neuromodulation for bladder dysfunction” were reviewed. Thirty-eight of these videos met the inclusion criteria. These videos were reviewed by two Urology Registrars and the videos were scored using two standardised tools. The DISCERN tool assesses quality of information and the Patient Education Materials Assessment Tool for Audiovisual Material (PEMAT-A/V) tool assesses user understandability and accessibility. Results: Forty-two percent of videos were deemed to be poor or very poor, with 58% being fair, good or excellent according to the DISCERN standardised tool. For PEMAT-A/V the average score for understandability was 74% (43–100%) and actionability was 38% (0–100%). We found statistical significance comparing the duration of videos to the DISCERN groups (p = 0.02). We also found significance comparing the understandability of videos using the PEMAT-A/V score to the DISCERN groups (p ≤ 0.05). Conclusions: Forty-two percent of videos on SNM are of poor or very poor quality. The actionability score for consumers to seek out further information is also low at 38%. This raises concerns about the quality of information that is widely available on YouTube and how consumers will use this information when making decisions about their health. Full article