Health Informatics Journal最新文献

The aim was to evaluate the content of videos titled "How to administer subcutaneous immunoglobulin in immunodeficiency" on YouTube. The search term 'How to administer subcutaneous immunoglobulin in immunodeficiency?' was searched on YouTube™ (https://www.youtube.com) and the first 200 videos were reviewed on December 16, 2023. The majority of the 40 videos included in the study were uploaded by patients (62.5%). It was found that the understandable rate of patients' uploads was significantly lower (4.0%) than other (46.7%) (p = .000). The number of likes and comments per 1000 views were higher in the patient group (p = .000, p = .006, respectively), but the GQS and mDISCERN scores were lower and statistically significant (p = .040, p = .000, respectively). Healthcare professionals and organizations have not shared enough videos on the use of subcutaneous immunoglobulin, and studies on this subject appear insufficient. In addition, a control mechanism is needed for video content on the internet related to health.

Objectives: We aimed to compare the Epic^® social determinants of health (SDOH) "wheel" to validated SDOH questionnaires in the domains of transportation security and financial toxicity to determine its accuracy in risk stratifying patients. Methods: We enrolled patients presenting to orthopaedic clinics at an urban tertiary care center, the University of Chicago Medical Center. Patients completed two validated surveys (the COmprehensive Score for financial Toxicity (COST) questionnaire and Transportation Security Index (TSI) questionnaire) and their Epic equivalents. The sensitivity and specificity of each Epic domain was calculated using validated questionnaires as the gold-standard. Results: 203 patients completed the transportation surveys while 199 completed the financial toxicity surveys. In the domain of financial toxicity, Epic's sensitivity and specificity were 35% 53%, respectively. In the domain of transportation security, Epic's sensitivity and specificity were 53% and 94%, respectively. Conclusions: The Epic SDOH wheel demonstrated poor sensitivity in both domains studied, suggesting limitations in its ability to serve as an effective screening tool.

Background: Nurses' clinical decision-making skills are vital for ensuring safe care and achieving optimal patient outcomes. Similarly, evidence-based practice improves quality of care and standardizes nursing services. Research is needed to examine factors affecting these skills. Objective: This study examined the relationship between informatics competency, attitudes toward evidence-based practice, and clinical decision-making skills among nurses. Method: This descriptive correlational study was conducted in 2024 with 300 nurses from hospitals affiliated with Birjand University of Medical Sciences, Birjand, Iran. Data were collected using questionnaires on demographic information, informatics competency, attitudes toward evidence-based practice, and clinical decision-making skills. Data were analyzed using SPSS-25 software at a significance level of p < 0.05. Results: A significant positive correlation was found between informatics competency (and its components), clinical decision-making skills, and evidence-based practice in the studied nurses. Informatics competency predicted about 26% of the variance in clinical decision-making skills and 20% of the variance in attitudes toward evidence-based practice. Conclusion: Nurse managers should implement targeted interventions to enhance informatics competency and improve attitudes toward evidence-based practice and decision-making skills.

Objective: This review assesses the efficacy of machine learning (ML) models for classification and management of Chronic Lymphocytic Leukaemia (CLL).Methods: Twenty studies published between 2014 and 2023 were reviewed, focusing on supervised ML models to predict patient outcomes or guide treatment decisions. Studies were identified through PubMed, Google Scholar, and IEEExplore, with the final search in March 2023. Inclusion criteria consisted of studies focused on ML applications in CLL. Exclusion criteria included studies lacking sufficient methodology or focused solely on experimental settings without clinical validation. Most studies used small, single-centre datasets, potentially contributing to overfitting and limited applicability to real-world settings.Results: Despite dataset limitations, all reviewed studies reported positive outcomes, with some demonstrating improvements in clinical workflows. Our findings advocate developing ML models using larger, multimodal, and multi-institutional datasets. Improved model interpretability and NLP implementation to harness unstructured clinical data were identified as key areas for advancement. Additionally, innovations like cross-site federated learning and automated redaction could help address data integration and privacy challenges.Conclusion: This review underscores the transformative potential of ML in CLL management. However, addressing limitations, including diverse datasets and enhanced model interpretability, is crucial for fully leveraging ML capabilities in haemato-oncology.

The disastrous era of COVID-19 has altered the perspectives of nearly all nations concerning the health and education sectors. Artificial intelligence is a pressing need that needs to be implemented thoroughly in the medical and educational fields. Imperatively, the diagnosis of Covid-19 has become crucial. In this study, we have designed a classification model based on Convolutional Neural Network (CNN) and transfer learning. The COVID-19 chest X-ray images have been considered for the proposed methodology and are classified as COVID-19 positive and normal cases. The proposed shallow CNN Model achieved an accuracy of 96%, which is computationally very effective as only three Convolutional blocks are required. Then, the Xception architecture-based model is experimented with. The accuracy and loss of the proposed model have been evaluated using Adam and SGD optimizer. With the Adam Optimizer, Xception Net achieved the best classification accuracy of 99.94%. The precision, recall, and f₁-score of 100% are achieved. The proposed model has outperformed the previous studies in the same domain, which highlights the model's state-of-the-art performance. Our study will be helpful for decision-makers and can help further minimize mortality and morbidity by effectively diagnosing the disease.

Introduction: Globally, many patients with severe or uncontrolled asthma receive inadequate treatment, which contributes to significant disease burden, healthcare resource utilisation, and decreased health-related quality of life. There is a need to develop a tool that can support identification of patients with severe or uncontrolled asthma, optimise their management in primary care, and facilitate appropriate referrals to severe asthma specialists. Methods: We describe the development of two novel digital tools to improve asthma management in the United Kingdom and the Netherlands: ReferID⁺ and AsthmaOptimiser. These tools have been designed to assist healthcare professionals to conduct structured asthma reviews, focusing on the most common causes of poor asthma control, and to help identify patients who may benefit from a referral to a specialist. Results: Both the ReferID+ and AsthmaOptimiser tools consist of a panel of asthma assessment questions completed by the healthcare professional through a digital interface during an in-person or virtual clinical visit. ReferID⁺ was developed for use in the UK National Health Service environment and is currently undergoing effectiveness testing in the randomised controlled OASIS study. The ReferID⁺ was adapted into the AsthmaOptimiser to address specific needs in the Netherlands; evaluation and implementation are currently underway in the CAPTURE study. Conclusions: ReferID⁺ and AsthmaOptimiser support comprehensive asthma consultations by providing personalised recommendations with guideline-based strategies to optimise asthma management.

Invited panelists from different countries, who are actively involved in digital medicine, discussed the current state and future prospects of telemedicine at a roundtable during an international conference. The discussion covered various aspects of telemedicine, including the available technologies and the critical need for comprehensive databases, as well as insights on completed projects and their long-term viability. Our expertise in technology, sustainability, and telemedicine initiatives can be valuable, with the understanding that the ideas expressed can be applied to all fields and situations, while ensuring that equity and equality in their application are paramount to avoid exacerbating existing disparities. The overall aim is to leverage experience to support the successful implementation of new telemedicine endeavors across different healthcare sectors, with a focus on wide access to technology, affordability, digital literacy, and cultural and linguistic inclusivity.

Objective: To evaluate large language models (LLMs) in managing mild cognitive impairment (MCI) and supporting nonspecialist healthcare professionals and care partners, comparing English and Chinese responses. Methods: Seventy-two MCI-related questions were submitted to ChatGPT-4o, Gemini, and Kimi. Responses were assessed for accuracy, comprehensibility, specificity, and actionability using a 5-point Likert scale. Statistical analyses included intraclass correlation coefficients and Mann-Whitney U tests. Results: LLMs performed best in the symptoms and diagnosis domain (M = 4.11 ± 0.15). Healthcare professionals' needs were better met than those of care partners, particularly in comprehensibility and actionability (p < .001). English responses were significantly more comprehensible and specific than Chinese responses (p < .001). Conclusion: This study highlights the potential of LLMs like ChatGPT, Gemini, and Kimi in supporting MCI management, especially in diagnosis and providing actionable insights. However, their performance varied across languages and user groups, with English responses generally more effective than Chinese. The findings emphasize the need for culturally and linguistically adapted LLMs to enhance accuracy and usability. Future research should focus on expanding user diversity, improving adaptability, and incorporating region-specific data to optimize LLMs for MCI care.

Dengue fever remains a persistent global health threat, particularly in Southeast Asia, the Pacific, and the Americas. This study aims to improve early detection and prediction of dengue outbreaks by addressing the challenges of data scarcity and complex transmission factors through a hybrid machine learning approach. We developed a methodology that integrates clustering and classification techniques to identify and predict seasonal patterns of dengue risk. Using regional data from Bangladesh, clustering was performed to uncover latent patterns, with optimal clusters selected based on low inertia and high silhouette scores. The supervised machine learning models were then trained on labeled data to classify dengue risk levels using key meteorological and demographic characteristics. Clustering analysis revealed well-defined structures within the data, with a silhouette score of 0.774, indicating robust clustering quality. The classification models demonstrated exceptional performance, achieving more than 99% in accuracy, precision, recall, and F1 score metrics. These models effectively identified high-risk periods and regions with strong seasonal trends in dengue incidence. Overall, this study presents a data-driven framework for the early detection of dengue outbreaks, supporting proactive public health strategies, while also contributing to the identification of dengue patterns and serving as a tool for controlling infectious diseases.

Machine learning techniques offer significant potential for improving the diagnosis of coronary heart disease by enabling earlier detection and timely intervention. This study presents a machine learning-based method utilizing clinical records to evaluate the impact of different data preprocessing sequences on predictive accuracy. Two clinical datasets were examined: one comprising heart failure patient data with 14 clinical features, and the Cleveland Heart Disease Dataset. The investigation compared two preprocessing strategies: standardisation prior to balancing, and balancing prior to scaling. Six machine learning models (XGBoost, GBDT, AdaBoost, Random Forest, KNN, and RaSE) were trained on an 80:20 data split and assessed using accuracy, precision, recall, and F1-score. Hyperparameters were optimized with Bayesian Optimisation. Results showed that both preprocessing designs achieved perfect accuracy on the Cleveland dataset. For the heart failure dataset, balancing before scaling led to improved accuracy (95%) compared with standardising before balancing (93.33%), and yielded higher macro-average and weighted-average F1-scores, signifying better overall classification performance. Among the evaluated models, XGBoost consistently provided the most robust predictions across conditions. These findings highlight the critical influence of preprocessing sequence on model effectiveness in imbalanced clinical data and suggest that balancing before scaling significantly enhances classification accuracy. XGBoost stands out as a reliable model for potential implementation in clinical decision support systems. Overall, this study advances the development of AI-driven tools for digital health applications, contributing meaningful insights to the field of health informatics.