Search Results (1,438)

Pharmacists have been recognized as the most accessible healthcare professionals, and research has been carried out on expanded pharmacy services they could provide. Additional pharmacy services are a cost-effective way to prevent medication errors, reduce the number of drug-related problems, and prevent chronic disease progression. Therefore, this study aims to evaluate pharmacy service users’ views of expanded pharmacy services in Croatia. This study included 745 participants. Patients who have a healthcare professional in their family more frequently knew of the existence of e-health records and the option to share it with their pharmacists (134, 56.3% vs. 229, 45.2%, p = 0.005), while persons that have chronic illness more frequently visit the same pharmacy (176, 77.9% vs. 178, 34.3%, p < 0.001). Participants are confident that pharmacists can provide screening services and education on inhaler usage; however, only around 60% agreed that pharmacists can independently lead therapy adjustment, medication substitution, or monitor therapy based on test results. Our findings should be supported with projects evaluating the cost-effectiveness of such services as they would be accepted by a greater number of pharmacy service users if covered by the national health insurer. Full article

Accurate synthetic image generation is crucial for addressing data scarcity challenges in medical image classification tasks, particularly in sensor-derived medical imaging. In this work, we propose a novel method using a Wasserstein Generative Adversarial Network with Gradient Penalty (WGAN-GP) and nearest-neighbor interpolation to generate high-quality synthetic images for diabetic retinopathy classification. Our approach enhances training datasets by generating realistic retinal images that retain critical pathological features. We evaluated the method across multiple retinal image datasets, including Retinal-Lesions, Fine-Grained Annotated Diabetic Retinopathy (FGADR), Indian Diabetic Retinopathy Image Dataset (IDRiD), and the Kaggle Diabetic Retinopathy dataset. The proposed method outperformed traditional generative models, such as conditional GANs and PathoGAN, achieving the best performance on key metrics: a Fréchet Inception Distance (FID) of 15.21, a Mean Squared Error (MSE) of 0.002025, and a Structural Similarity Index (SSIM) of 0.89 in the Kaggle dataset. Additionally, expert evaluations revealed that only 56.66% of synthetic images could be distinguished from real ones, demonstrating the high fidelity and clinical relevance of the generated data. These results highlight the effectiveness of our approach in improving medical image classification by generating realistic and diverse synthetic datasets. Full article

This study explores the potential of natural language models, including large language models, to extract causal relations from medical texts, specifically from clinical practice guidelines (CPGs). The outcomes of causality extraction from clinical practice guidelines for gestational diabetes are presented, marking a first in the field. The results are reported on a set of experiments using variants of BERT (BioBERT, DistilBERT, and BERT) and using newer large language models (LLMs), namely, GPT-4 and LLAMA2. Our experiments show that BioBERT performed better than other models, including the large language models, with an average F1-score of 0.72. The GPT-4 and LLAMA2 results show similar performance but less consistency. The code and an annotated corpus of causal statements within the clinical practice guidelines for gestational diabetes are released. Extracting causal structures might help identify LLMs’ hallucinations and possibly prevent some medical errors if LLMs are used in patient settings. Some practical extensions of extracting causal statements from medical text would include providing additional diagnostic support based on less frequent cause–effect relationships, identifying possible inconsistencies in medical guidelines, and evaluating the evidence for recommendations. Full article

Background: Diabetes is a metabolic disorder characterized by increased blood sugar levels. Early detection of diabetes could help individuals to manage and delay the progression of this disorder effectively. Machine learning (ML) methods are important in forecasting the progression and diagnosis of different medical problems with better accuracy. Although they cannot substitute the work of physicians in the prediction and diagnosis of disease, they can be of great help in identifying hidden patterns based on the results and outcome of disease. Methods: In this research, we retrieved the PIMA dataset from the Kaggle repository, the retrieved dataset was further processed for applied PCA, heatmap, and scatter plot for exploratory data analysis (EDA), which helps to find out the relationship between various features in the dataset using visual representation. Four different ML algorithms Random Forest (RF), Decision Tree (DT), Naïve Bayes (NB), and Logistic regression (LR) were implemented on Rattle using Python for the prediction of diabetes among the female population. Results: Results of our study showed that RF performs better in terms of accuracy of 80%, precision of 82%, error rate of 20%, and sensitivity of 88% as compared to other developed models DT, NB, and LR. Conclusions: Diabetes is a common problem prevailing across the globe, ML-based prediction models can help in the prediction of diabetes much earlier before the worsening of the condition. Full article

Background/Objectives: Several population pharmacokinetic models of vedolizumab (VDZ) are available for inflammatory bowel disease (IBD) patients. However, their predictive performance in real-world clinical settings remains unknown. This study aims to externally evaluate the published VDZ pharmacokinetic models, focusing on their predictive performance and simulation-based clinical applicability. Methods: A literature search was conducted through PubMed to identify VDZ population pharmacokinetic models. A total of 114 VDZ concentrations from 106 IBD patients treated at the University Medical Center “Zvezdara”, Republic of Serbia, served as the external evaluation cohort. The predictive performance of the models was assessed using prediction- and simulation-based diagnostics. Furthermore, the models were utilized for Monte Carlo simulations to generate concentration–time profiles based on 24 covariate combinations specified within the models. Results: Four published pharmacokinetic models of VDZ were included in the evaluation. Using the external dataset, the median prediction error (MDPE) ranged from 13.82% to 25.57%, while the median absolute prediction error (MAPE) varied between 41.64% and 47.56%. None of the models fully met the combined criteria in the prediction-based diagnostics. However, in simulation-based diagnostics, pvcVPC showed satisfactory results, despite wide prediction intervals. Analysis of NPDE revealed that only the models by Rosario et al. and Okamoto et al. fulfilled the evaluation criteria. Simulation analysis further demonstrated that the median VDZ concentration remains above 12 μg/mL at week 22 during maintenance treatment for approximately 45–60% of patients with the best-case covariate combinations and an 8-week dosing frequency. Conclusions: None of the published models satisfied the combined criteria (MDPE, MAPE, percentages of prediction error within ±20% and ±30%), rendering them unsuitable for a priori predictions. However, two models demonstrated better suitability for simulation-based applications. Full article

Background: Significant intraoperative and postoperative blood loss are rare but possibly life-threatening complications after lung resection surgery either during open or minimally invasive procedures. Microporous Polysaccharide Haemospheres (ARISTA™AH) have demonstrated time-efficient haemostasis, lower postoperative blood volumes and a lower blood transfusion requirement, without any identified adverse events across other specialities. The primary aim of our study was to evaluate the impact of ARISTA™AH on short-term postoperative outcomes in thoracic surgery. Our secondary aim was to compare ARISTA™AH with other commonly used haemostatic agents. Methods: We retrospectively reviewed a prospectively collected database of consecutive early-stage lung cancer patients surgically treated in two European centres (October 2020–December 2022). Exclusion criteria included open surgery, patients with coagulopathy/anticoagulant medication, major intraoperative bleeding, non-anatomical lung resection and age <18 years. The cohort was divided into five groups according to the haemostatic agent that was used. Propensity score matching was used to estimate the effect of ARISTA™AH on various intra- and postoperative parameters (continuous and binary outcome modelling). Results: A total of 482 patients (M/F:223/259; VATS 97/RATS 385) with a mean age of 68.9 (±10.6) years were analysed. In 253 cases, ARISTA™AH was intraoperatively used to control bleeding. This cohort of patients had a significant reduction in total drain volume by 135 mls (standard error 53.9; p = 0.012). The use of ARISTA™AH did reduce the average length of a hospital stay (−1.47 days) and the duration of chest drainage (−0.596 days), albeit not significantly. In the ARISTA™AH group, we observed no postoperative bleeding, no blood transfusion requirement, no 30-day mortality and no requirement for redo surgery. The use of ARISTA™AH significantly reduced the odds of postoperative complications, as well as the need for transfusion and redo surgery. Conclusions: Our data showed that Microporous Polysaccharide Haemospheres are a safe and effective haemostatic device. Their use has a positive effect on the short-term postoperative outcomes of patients surgically treated for early-stage lung cancer. Full article

Background: Since 2023, ChatGPT-4 has been impactful across several sectors including healthcare, where it aids in medical information analysis and education. Electronic patient-reported outcomes (ePROs) play a crucial role in monitoring cancer patients’ post-treatment symptoms, enabling early interventions. However, managing the voluminous ePRO data presents significant challenges. This study assesses the feasibility of utilizing ChatGPT-4 for analyzing side effect data from ePROs. Methods: Thirty cancer patients were consecutively collected via a web-based ePRO platform, reporting side effects over 4 weeks. ChatGPT-4, simulating oncologists, dietitians, and nurses, analyzed this data and offered improvement suggestions, which were then reviewed by professionals in those fields. Results: Two oncologists, two dieticians, and two nurses evaluated the AI’s performance across roles with 540 reviews. ChatGPT-4 excelled in data accuracy and completeness and was noted for its empathy and support, enhancing communication and reducing caregiver stress. It was potentially effective as a dietician. Discussion: This study offers preliminary insights into the feasibility of integrating AI tools like ChatGPT-4 into ePRO cancer care, highlighting its potential to reduce healthcare provider workload. Key directions for future research include enhancing AI’s capabilities in cancer care knowledge validation, emotional support, improving doctor-patient communication, increasing patient health literacy, and minimizing errors in AI-driven clinical processes. As technology advances, AI holds promise for playing a more significant role in ePRO cancer care and supporting shared decision-making between clinicians and patients. Full article

Introduction: To report the role of prostaglandin-associated periorbitopathy (PAP) severity on the surgical efficacy of Ahmed Glaucoma Valve (AGV) implantation. Subjects and Methods: Retrospective observational case series. Participants were the consecutive 102 eyes from 102 Japanese subjects (55 males, 47 females; mean age ± standard deviation, 74.9 ± 7.8 years) who underwent AGV implantation for primary open-angle glaucoma (POAG), completed full postoperative visits for 12 months, and had information on PAP severity graded by the Shimane University PAP Grading System (SU-PAP). Data were collected via medical chart review. Comparison of surgical success rates among groups stratified by SU-PAP grades (grades 0–3) using survival curve analysis. Failure was defined based on additional glaucoma surgery, IOP reduction in less than 20%, postoperative IOP exceeding 18 mmHg (definition A) or 15 mmHg (definition B), or postoperative visual acuity reduced to no light perception. Results: At 12 months postoperatively, the success rates for grades 0, 1, 2, and 3 were 47%, 43%, 42%, and 73%, respectively, for definition A (p = 0.35) and 35%, 26%, 19%, and 27%, respectively, for definition B (p = 0.64, log-rank test). For definition A, younger age was associated with surgical failure (Hazard ratio = 0.97/year, p = 0.049, Wald test), while no other factors, including gender, preoperative IOP, medications, refractive error, history of conjunctival manipulation procedures, or SU-PAP grade, were associated with surgical failure. For definition B, no factors were found to influence surgical outcomes. Conclusions: The preoperative severity of PAP might not affect the postoperative outcomes of AGV. Given that the success rate of trabeculectomy is influenced by PAP severity, in cases with severe PAP, physicians are advised to consider long-tube shunt surgery as an initial filtration procedure or as a rescue procedure when filtration surgery is unsuccessful. Full article

Accurate and early detection of brain tumors is essential for improving clinical outcomes and guiding effective treatment planning. Traditional segmentation techniques in MRI often struggle with challenges such as noise, intensity variations, and complex tumor morphologies, which can hinder their effectiveness in critical healthcare scenarios. This study proposes an innovative hybrid methodology that integrates advanced metaheuristic optimization and entropy-based fuzzy clustering to enhance segmentation precision in brain tumor detection. This method combines the nature-inspired Elephant Herding Optimization (EHO) algorithm with Entropy-Driven Fuzzy C-Means (EnFCM) clustering, offering significant improvements over conventional methods. EHO is utilized to optimize the clustering process, enhancing the algorithm’s ability to delineate tumor boundaries, while entropy-based fuzzy clustering accounts for intensity inhomogeneity and diverse tumor characteristics, promoting more consistent and reliable segmentation results. This approach was evaluated using the BraTS challenge dataset, a benchmark in the field of brain tumor segmentation. The results demonstrate marked improvements across several performance metrics, including Dice similarity, mean squared error (MSE), peak signal-to-noise ratio (PSNR), and the Tanimoto coefficient (TC), underscoring this method’s robustness and segmentation accuracy. By managing image noise and reducing computational demands, the EHO-EnFCM approach not only captures intricate tumor structures but also facilitates efficient image processing, making it suitable for real-time clinical applications. Overall, the findings reveal the potential of this hybrid approach to advance MRI-based tumor detection, offering a promising tool that enhances both accuracy and computational efficiency for medical imaging and diagnosis. Full article

Brain tumor detection is crucial in medical research due to high mortality rates and treatment challenges. Early and accurate diagnosis is vital for improving patient outcomes, however, traditional methods, such as manual Magnetic Resonance Imaging (MRI) analysis, are often time-consuming and error-prone. The rise of deep learning has led to advanced models for automated brain tumor feature extraction, segmentation, and classification. Despite these advancements, comprehensive reviews synthesizing recent findings remain scarce. By analyzing over 100 research papers over past half-decade (2019–2024), this review fills that gap, exploring the latest methods and paradigms, summarizing key concepts, challenges, datasets, and offering insights into future directions for brain tumor detection using deep learning. This review also incorporates an analysis of previous reviews and targets three main aspects: feature extraction, segmentation, and classification. The results revealed that research primarily focuses on Convolutional Neural Networks (CNNs) and their variants, with a strong emphasis on transfer learning using pre-trained models. Other methods, such as Generative Adversarial Networks (GANs) and Autoencoders, are used for feature extraction, while Recurrent Neural Networks (RNNs) are employed for time-sequence modeling. Some models integrate with Internet of Things (IoT) frameworks or federated learning for real-time diagnostics and privacy, often paired with optimization algorithms. However, the adoption of eXplainable AI (XAI) remains limited, despite its importance in building trust in medical diagnostics. Finally, this review outlines future opportunities, focusing on image quality, underexplored deep learning techniques, expanding datasets, and exploring deeper learning representations and model behavior such as recurrent expansion to advance medical imaging diagnostics. Full article

The rapid advancement of artificial intelligence is transforming the computer-aided design of microfluidic chips. As a key component, microfluidic mixers are widely used in bioengineering, chemical experiments, and medical diagnostics due to their efficient mixing capabilities. Traditionally, the simulation of these mixers relies on the finite element method (FEM), which, although effective, presents challenges due to its computational complexity and time-consuming nature. To address this, we propose a nine-grid network (NGN) model theory with a centrally symmetric structure.The NGN uses a symmetric structure similar to a 3 × 3 grid to partition the fluid space to be predicted. Using this theory, we developed and trained an artificial neural network (ANN) to predict the fluid dynamics within microfluidic mixers. This approach significantly reduces the time required for fluid evaluation. In this study, we designed a prototype microfluidic mixer and validated the reliability of our method by comparing it with predictions from traditional FEM software. The results show that our NGN model completes fluid predictions in just 40 s compared to approximately 10 min with FEM, with acceptable error margins. This technology achieves a 15-fold acceleration, greatly reducing the time and cost of microfluidic chip design. Full article

Background and Objectives: Brain tumors are complex diseases that require careful diagnosis and treatment. A minor error in the diagnosis may easily lead to significant consequences. Thus, one must place a premium on accurately identifying brain tumors. However, deep learning (DL) models often face challenges in obtaining sufficient medical imaging data due to legal, privacy, and technical barriers hindering data sharing between institutions. This study aims to implement a federated learning (FL) approach with privacy-preserving techniques (PPTs) directed toward segmenting brain tumor lesions in a distributed and privacy-aware manner.Methods: The suggested approach employs a model of 3D U-Net, which is trained using federated learning on the BraTS 2020 dataset. PPTs, such as differential privacy, are included to ensure data confidentiality while managing privacy and heterogeneity challenges with minimal communication overhead. The efficiency of the model is measured in terms of Dice similarity coefficients (DSCs) and 95% Hausdorff distances (HD95) concerning the target areas concerned by tumors, which include the whole tumor (WT), tumor core (TC), and enhancing tumor core (ET). Results: In the validation phase, the partial federated model achieved DSCs of 86.1%, 83.3%, and 79.8%, corresponding to 95% values of 25.3 mm, 8.61 mm, and 9.16 mm for WT, TC, and ET, respectively. On the final test set, the model demonstrated improved performance, achieving DSCs of 89.85%, 87.55%, and 86.6%, with HD95 values of 22.95 mm, 8.68 mm, and 8.32 mm for WT, TC, and ET, respectively, which indicates the effectiveness of the segmentation approach, and its privacy preservation.Conclusion: This study presents a highly competitive, collaborative federated learning model with PPTs that can successfully segment brain tumor lesions without compromising patient data confidentiality. Future work will improve model generalizability and extend the framework to other medical imaging tasks. Full article

Background: The fourth universal definition of myocardial infarction (MI) introduced the differentiation of acute myocardial injury from MI. In this study, we developed a computational phenotype for distinct identification of acute myocardial injury and MI within electronic medical records (EMRs). Methods: Two cohorts were used from a Calgary-wide EMR system: a chart review of 3042 randomly selected inpatients from Dec 2014 to Jun 2015; and 11,685 episodes of care that included cardiac catheterization from Jan 2013 to Apr 2017. Electrocardiogram (ECG) reports were processed using natural language processing and combined with high-sensitivity troponin lab results to classify patients as having an acute myocardial injury, MI, or neither. Results: For patients with an MI diagnosis, only 64.0% (65.7%) in the catheterized cohorts (chart review cohort) had two troponin measurements within 6 h of each other. For patients with two troponin measurements within 6 h; of those with an MI diagnosis, our phenotype classified 25.2% (31.3%) with an acute myocardial injury and 62.2% (55.2%) with an MI in the catheterized cohort (chart review cohort); and of those without an MI diagnosis, our phenotype classified 12.9% (12.4%) with an acute myocardial injury and 10.0% (13.1%) with an MI in the catheterized cohort (chart review cohort). Conclusions: Patients with two troponin measurements within 6 h, identified by our phenotype as having either an acute myocardial injury or MI, will at least meet the diagnostic criteria for an acute myocardial injury (barring lab errors) and indicate many previously uncaptured cases. Myocardial infarctions are harder to be certain of because ECG report findings might be superseded by evidence not included in our phenotype, or due to errors with the natural language processing. Full article

Background and Objectives: On a local and national scale, genitourinary malformations (GUMs) are the second most encountered congenital anomaly in children. GUMs are linked to several risk factors, including maternal co-morbidities and insufficient folic acid. They may also be related to maternal health and care during pregnancy. Expanding our knowledge about these factors is necessary for the development of preventative measures, which could reduce GUM incidence. This study evaluated the local youth’s understanding and perceptions of genitourinary anomalies. Materials and Methods: This cross-sectional, qualitative, anonymous, questionnaire-based study involved members of the local population, aged 18 years or over. Based on a 5% type I error rate (α = 0.05) and an 80% response rate, a sample size of 481 was determined. The questionnaire was completed by 902 people. The data were analyzed using SPSS version 25 (IBM). Results: Over half (57%) of respondents believed that hormonal therapy during pregnancy could increase GUM risk. Moreover, 46% thought that maternal chronic diseases could be another risk factor, while 43% believed that pregnancy-related conditions, such as pre-eclampsia, increased GUM risk. Women had higher odds of high perception scores than men, according to the univariate and multivariate analyses. Most participants (74%) strongly agreed that proper and ongoing prenatal follow-ups are necessary, 69% agreed that premarital medical check-ups are necessary, and 67% believed that optimal nutrition throughout pregnancy is necessary to reduce GUM risk. Conclusions: The results emphasize the necessity of developing healthcare strategies specifically designed to increase knowledge about GUMs and overcome incorrect community perceptions of risk factors that could also help improve attitudes towards prevention and ultimately reduce the incidence of GUMs. Full article

Hospital inpatient care relies on constant monitoring and reliable real-time data. Continuous improvement, adaptability, and state-of-the-art technologies are critical for ongoing efficiency, productivity, and readiness growth. When appropriately used, technologies, such as blockchain and IoT-enabled devices, can change the practice of medicine and ensure that it is performed based on correct assumptions and reliable data. The proposed electronic health record (EHR) can obtain context information from beacons, change the user interface of medical devices according to their location, and provide a more user-friendly interface for medical devices. The data generated, which are associated with the location of the beacons and devices, were stored in Hyperledger Fabric, a permissioned distributed ledger technology. Overall, by prompting and adjusting the user interface to context- and location-specific information while ensuring the immutability and value of the data, this solution targets a decrease in medical errors and an increase in the efficiency in healthcare inpatient care by improving user experience and ease of access to data for health professionals. Moreover, given auditing, accountability, and governance needs, it must ensure when, if, and by whom the data are accessed. Full article