DIGITAL HEALTH最新文献

Background: Mobile health (mHealth) applications are increasingly seen as essential components of healthcare digitalization, yet many national initiatives struggle to progress beyond pilot phases. In Czechia, the EZKarta mobile application was launched by the Ministry of Health as a secure digital gateway to vaccination records and preventive check-ups, marking a first step toward a national eHealth platform.

Objective: This study provides an exploratory, analytically grounded insight into early user perceptions and stakeholder views on EZKarta during its pilot phase, focusing on institutional, governance, and user-level factors influencing sustainability and integration.

Methods: A mixed-methods design was applied. Quantitative data from a national online survey (n = 209) were analyzed using nonparametric tests. The qualitative component included semistructured interviews with key stakeholders. Findings were integrated through joint interpretation and thematic triangulation.

Results: The mean usability score (UMUX = 33.2 ± 6.5) was significantly below the international benchmark (p < 2.2 × 10^-16). Only 38% of users reported satisfaction, while 72% indicated willingness to use the application if integrated with their provider's clinical system. Triangulation of survey and interview data suggests that low engagement was driven primarily by limited functionality, lack of clinical system integration, and unclear perceived added value. Stakeholders highlighted fragmented governance as key barriers, while recognizing EZKarta's potential role in national digital health coordination.

Conclusions: EZKarta exemplifies both the opportunities and constraints of mHealth adoption in transitional health systems. Stronger institutional coordination and transparent communication are essential for long-term relevance. The findings may inform policymakers in Central and Eastern Europe.

Background: Digital health misinformation on social media and emerging AI platforms poses a growing challenge for healthcare systems. Community pharmacists are often the first point of contact for patients influenced by online health content, yet limited evidence exists from Saudi Arabia, where social media engagement is among the highest globally.

Objective: To examine how community pharmacists in Saudi Arabia experience digital health misinformation, including its sources, patient impacts, confidence in addressing misinformation, and training needs.

Methods: A descriptive cross-sectional online survey was conducted among licensed community pharmacists in Saudi Arabia between March and April 2025. Using a convenience-plus-snowball sampling strategy, A total of 768 completed survey responses from licensed community pharmacists were included in the analysis. These responses were obtained through an anonymous, nationally distributed electronic questionnaire. The survey explored exposure to misinformation, pharmacists' confidence, and preferred training modalities. Descriptive statistics and regression analyses were used to identify predictors of key outcomes.

Results: Most pharmacists (89.6%) reported encountering patient-derived digital health misinformation, with 51.6% experiencing such encounters weekly or daily. Facebook (39.3%) and WhatsApp (27.7%) were identified as the most common sources, while 10% of encounters involved AI-generated content. Common misinformation themes included supplement misuse (27.1%) and concerns about medication safety (25.7%). Only 34.5% of pharmacists felt confident in identifying or correcting misinformation. Age, years of experience, and frequency of exposure were significantly associated with higher confidence and with perceived patient-behaviour impact. More than half (59.5%) supported formal training in misinformation management.

Conclusions: Digital misinformation is now a routine component of community pharmacy practice in Saudi Arabia and has tangible implications for patient behaviour and medication safety. Although confidence in managing misinformation was low, digitally active community pharmacists expressed strong readiness for targeted training. These findings underscore the need for structured educational initiatives and coordinated policy responses to strengthen digital health literacy and misinformation management within pharmacy practice.

Objectives: Heart failure (HF) patients admitted to intensive care units are prone to early readmission, which leads to adverse outcomes and increased healthcare costs. Existing prediction models often suffer from data heterogeneity, class imbalance, and limited interpretability. This study aimed to develop an interpretable ensemble learning framework to predict 30-day ICU readmission in adult patients with HF and to compare its performance with conventional single-classifier approaches.

Methods: This retrospective study analyzed 5414 adult HF patients from the MIMIC-III database. Using clinical and demographic variables collected within the first 24 h of the index ICU admission, the study aimed to predict 30-day ICU readmission (return to ICU). A two-stage ensemble model was developed using stratified sampling and grid-search optimization, with top learners integrated via a soft-voting mechanism. Additionally, SHapley Additive exPlanation (SHAP) analysis was employed to ensure model interpretability and quantify variable contributions to the predictions.

Results: The KNN-imputed Voting (3 Models) ensemble emerged as the optimal framework, achieving an accuracy of 0.8413, F1-score of 0.8195, and AUROC of 0.6718. Despite moderate AUROC, the model achieved strong recall and reliable calibration, making it suitable for risk stratification in post-ICU care transitions. The SHAP analysis identified Glucose, hemodynamic parameters (e.g., blood pressure, heart rate), and inflammatory indicators as key predictors, aligning with established clinical understanding of stress hyperglycemia and hemodynamic instability in HF.

Conclusion: This interpretable ensemble framework predicts 30-day ICU readmission in HF patients with robust performance, effectively balancing sensitivity and discrimination. It supports electronic health record-based risk stratification and timely intervention. Future work should focus on external validation across diverse populations to ensure generalizability.

Objective: The thyroid gland plays a vital role in human body functions, including metabolism, and any dysfunction of the organ may exacerbate health risks. Early recovery from these conditions depends on the accurate diagnosis of accurate type of thyroid gland disorder. This study aims to develop an automated system for assisting physicians in clinical diagnosis of thyroid gland disorders.

Method: The transfer learning capability of the deep neural network model DenseNet-201 was leveraged, and the model was tailored by altering its fully connected layer and the classification layer for the classification of thyroid gland conditions into seven categories, namely cold nodule, hot nodule, multi-nodular goiter, nodular goiter, thyroiditis, toxic diffuse goiter and normal.

Results: The class-wise and overall performance of the method was evaluated by computing quantitative metrics like accuracy, specificity, precision, sensitivity, F1-score, and area under the curve (AUC). The degree of similarity between the true labels of thyroid disorders annotated by experts and those predicted by the method was gauged using the kappa coefficient. For the five-fold cross-validated experimental results, we obtained an accuracy of 91.48 ± 2.79%, specificity of 98.58 ± 0.47%, precision of 91.57 ± 2.76%, sensitivity of 91.48 ± 2.77%, F1-score of 91.38 ± 2.82%, and AUC of 0.988 ± 0.006. Additionally, the degree of similarity in the diagnostic capability of the proposed method and medical experts was measured by computing the kappa coefficient as 0.9148.

Conclusion: The experimental results of the proposed method were compared with contemporary methods and illustrate relatively better performance in terms of accuracy, sensitivity, precision, and F1-score. The value of the kappa coefficient, 0.9148, also depicts that the proposed method has the potential for applicability in clinical diagnosis to assist physicians in assessing the accurate type of thyroid disorders.

Objective: Smart clinical trials integrating artificial intelligence (AI), wearable and Internet of Things (IoT) sensors, and data-driven platforms have the potential to make clinical research more efficient, inclusive, and patient-centered. However, regulatory permissiveness does not always translate into actual adoption in the real-world. This study aims to examine how national policy environments and operational factors interact to shape the uptake of digital trial technologies in Korea and the United Kingdom between 2015 and 2025.

Methods: We analyzed 1172 interventional trials registered on ClinicalTrials.gov using a multi-label classification pipeline to identify the use of AI, wearable/IoT technologies, clinical data integration, and digital platforms. Adoption patterns were linked to a policy friction index that quantified seven categories of regulatory barriers in each country. Cross-country comparisons were conducted to assess alignment between policy permissiveness and observed technology adoption.

Results: Despite relatively high policy openness, the United Kingdom demonstrated persistently low adoption of AI, wearable/IoT, and digital platform technologies, reflecting implementation barriers such as validation burden, governance requirements, and workflow integration challenges. In contrast, Korea exhibited strong uptake of clinical data integration technologies despite higher regulatory friction, driven by institutional data infrastructures and hospital-centric ecosystems. Overall, adoption patterns diverged systematically from policy expectations in both countries.

Conclusions: These findings suggest that digital transformation in clinical research requires more than permissive policy frameworks; it depends on effective alignment among regulation, infrastructure, and implementation science. By introducing a reproducible framework that links regulatory friction to observed technology adoption, this study provides actionable insights for accelerating safe, interoperable, and scalable smart clinical trial deployment within evolving digital health ecosystems.

Objectives: The primary goal is to address the challenges in brain tumor segmentation (BraTS), such as limited accuracy and high computational costs, by developing a more precise and efficient segmentation technique. The study aims to improve the diagnosis and treatment planning of brain tumors by enabling clinicians to accurately localize and assess tumor regions from multimodal magnetic resonance imaging (MRI) scans.

Methods: We proposed self-attention U-Net (SAU-Net), a novel model that integrated self-attention mechanisms with the U-Net convolutional architecture. This design allowed the model to preserve spatial context while selectively concentrating on pertinent features, thereby enhancing tumor (ET) boundary delineation and overall segmentation accuracy. Extensive experiments were conducted on the BraTS 2018 and BraTS 2020 datasets using thorough cross-validation and testing protocols. The performance of SAU-Net was evaluated and compared against other attention-based U-Net models, including adaptive attention U-Net, multi-head attention U-Net, and group query attention U-Net.

Results: On the BraTS 2018 dataset, SAU-Net achieved Dice scores of 98.16% (whole tumor (WT)), 98.87% (tumor core (TC)), and 98.23% (ET), with an average Dice score of 98.23%. For the BraTS 2020 dataset, the model recorded Dice scores of 98.99% (WT), 98.70% (TC), and 99.18% (ET), with an average Dice score of 98.62%. In addition to superior segmentation performance, the model demonstrated reduced computational complexity in both training and prediction times, along with optimized memory usage.

Conclusion: SAU-Net is a highly effective and computationally efficient model for BraTS. Its superior performance, as evidenced by the high Dice scores on two benchmark datasets, combined with its reduced computational requirements, underscores its potential for practical and impactful clinical applications.

Background: Dashboards are tools that support decision-making by leveraging data, delivering precise and prompt information, and significantly aiding organizations in meeting their informational needs while improving data-driven decision-making.

Objective: This study aims to synthesize the evidence on applications, benefits, design features, and challenges of healthcare dashboards through an umbrella review.

Method: Four databases were searched systematically, including PubMed, Web of Science, Embase, and Scopus, within the time frame of 2010 to 2024, with the final article search completed on 30 December 2024. The retrieved reviews were first appraised by applying the JBI Critical Appraisal Checklist and the Risk of Bias in Systematic Reviews (ROBIS) tool. Following quality assessment and data extraction, the content was organized for synthesis into four analytical categories: applications, benefits, design features, and challenges. Data were further organized according to user groups and the two primary dashboard types (Clinical and Quality) in the tables to facilitate interpretation.

Results: A total of 41 studies were included for data synthesis, with 73% rated as high quality. The majority of articles were from the United States and focused on the public health context. Across the included studies, 12 distinct applications, 9 key benefits, 6 design features, and 8 challenges related to healthcare dashboards were identified. Findings were structured to reflect differences and overlaps between clinical and quality dashboards, highlighting their specific roles in patient care versus organizational performance.

Conclusion: Healthcare dashboards serve diverse applications across clinical and public health settings, offering benefits such as enhanced decision-making, improved efficiency, and increased adherence to evidence-based practices. However, the successful implementation of these systems depends on addressing persistent challenges related to data quality, integration, usability, and user engagement. Classifying dashboards into clinical and quality types provides a clearer framework for context-sensitive design, tailored implementation strategies, and improving their impact in healthcare systems.

Background: Access to timely and reliable maternal health information is a critical determinant of maternal and neonatal outcomes. Mobile health interventions, particularly WhatsApp, are increasingly used to overcome geographic, financial, and sociocultural barriers. However, evidence on their effectiveness in antenatal care (ANC) remains scattered.

Objectives: To synthesize evidence on the effectiveness of WhatsApp-based educational interventions in improving knowledge, behaviors, psychological well-being, clinical outcomes, and service uptake among pregnant women.

Data sources: Six databases (PubMed, Medline, SCOPUS, Web of Science, Cochrane Library, and Google Scholar) and reference lists were searched up to March 2025.

Study eligibility criteria: Eligible studies were randomized controlled trials (RCTs) or quasi-experimental designs evaluating WhatsApp-based interventions in pregnant women, compared with routine ANC or other educational modalities.

Study appraisal and synthesis methods: Screening and data extraction were performed independently by two reviewers using Rayyan. Given heterogeneity in interventions and outcomes, findings were synthesized narratively.

Results: Of 4374 records identified, 21 studies (12 RCTs, 9 quasi-experimental) were included. Most studies (76%) enrolled ≤50 participants per WhatsApp arm and were conducted in upper-middle and high-income countries. WhatsApp-based interventions improved maternal knowledge (e.g., anemia prevention, breastfeeding), health behaviors (self-care, smoking cessation, and dietary practices), psychological outcomes (reduced anxiety, depression, tokophobia and improved maternal-fetal attachment), clinical outcomes (hemoglobin, blood pressure and glycemic control), and ANC attendance and satisfaction. However, WhatsApp was less effective than motivational interviewing, phone calls, or face-to-face education in some domains.

Limitations: Heterogeneity in interventions and outcome measures precluded meta-analysis. Most studies were small, for a short term, and concentrated in higher-income settings.

Conclusions: WhatsApp-based education is a feasible, low-cost adjunct to routine ANC that can improve knowledge, health behaviors, psychological well-being and satisfaction. Future research should prioritize large-scale, multicenter RCTs in low-resource settings, with standardized outcome measures and long-term follow-up, to establish effectiveness and scalability.

Objectives: In the world, skin cancer is a significant health concern, and early diagnosis of this cancer plays a key role in improving patient outcomes. The early detection of this cancer reduces the death rate, but due to the complexity of the diagnosis, incorrect detection and prediction are provided by the experts. Therefore, it is essential to propose a computer-aided diagnostic system based on deep learning and explainable Artificial Intelligence (XAI) techniques that can be used as a second opinion in clinics and help physicians more accurately detect and predict this type of cancer.

Methods: This work presents the proposed deep learning architecture consisting of two modules-skin lesion segmentation and lesion type classification. The proposed architecture is interpreted using XAI techniques to better evaluate the black-box model. In the skin lesion segmentation phase, we implemented DeepLab V3 architecture for semantic segmentation. The ResNet-18 model was used as the backbone, and later hyperparameters were optimized using Bayesian Optimization (BO). In the classification phase, we design a FusedNet architecture called Inverted self-attention with Vision Transformer (ISAwViT). The proposed fused network combines an inverted self-attention residual architecture with a vision transformer. The proposed fused network extracted feature information more deeply than performing an accurate prediction in a later stage. The design model is trained, and later in the testing phase, extracted features are classified using Softmax and several other classifiers.

Results: The lesion segmentation and classification experiment was conducted on the HAM10000 dataset. The accuracy achieved by the HAM10000 dataset was 95.16% for lesion segmentation and 97.5% for lesion classification.

Conclusion: Compared with recent techniques, the proposed model is more effective and efficient. In addition, the interpretation of the proposed model was performed using LIME and Grad-CAM, which show how the fused model makes correct classifications.