Healthcare Informatics Research最新文献

Objectives: Anemia is characterized by a reduction in red blood cells, leading to insufficient levels of hemoglobin, the molecule responsible for carrying oxygen. The current standard method for diagnosing anemia involves analyzing blood samples, a process that is time-consuming and can cause discomfort to participants. This study offers a comprehensive analysis of non-invasive anemia detection using conjunctiva images processed through various machine learning and deep learning models. The focus is on the palpebral conjunctiva, which is highly vascular and unaffected by melanin content.

Methods: Conjunctiva images from both anemic and non-anemic participants were captured using a smartphone. A total of 764 conjunctiva images were augmented to 4,315 images using the deep convolutional generative adversarial network model to prevent overfitting and enhance model robustness. These processed and augmented images were then utilized to train and test multiple models, including statistical regression, machine learning algorithms, and deep learning frameworks.

Results: The stacking ensemble framework, which includes the models VGG16, ResNet-50, and InceptionV3, achieved a high area under the curve score of 0.97. This score demonstrates the framework's exceptional capability in detecting anemia through a noninvasive approach.

Conclusions: This study introduces a noninvasive method for detecting anemia using conjunctiva images obtained with a smartphone and processed using advanced deep learning techniques.

Objectives: Noncommunicable diseases (NCDs) pose a significant burden, especially in low- and middle-income countries such as the Philippines. To tackle this issue, the Department of Health launched the Philippine Package of Essential Non-Communicable Disease Interventions (PhilPEN), which includes the use of the Noncommunicable Disease Risk Assessment Form. However, healthcare workers have encountered difficulties due to the form's complexity and the lengthy process required. This study aimed to create a mobile medical app for community health workers by adapting the PhilPEN Noncommunicable Disease Risk Assessment Form using the extreme prototyping framework. The focus was on simplifying data collection and improving the usability of health technology solutions.

Methods: The study employed a qualitative research methodology, which included key informant interviews, linguistic validation, and cognitive debriefing. The extreme prototyping framework was utilized for app development, comprising static prototype, dynamic prototype, and service implementation phases. The app was developed with HTML5, CSS3, JavaScript, and Apache Cordova, adhering to World Health Organization (WHO) guidelines and PhilHealth Circular.

Results: The development process involved three prototype cycles, each consisting of multiple mini-cycles of feedback, system design, coding, and testing. Version 1.xx was aligned with WHO guidelines, Version 2.xx integrated the Department of Health NCD Risk Assessment Form, and Version 3.xx adapted to the updated form with expanded requirements.

Conclusions: The extreme prototyping framework was effectively applied in the development of a medical mobile app, facilitating the integration of health science and information technology. Future research should continue to validate the effectiveness of this approach and identify specific nuances related to health science applications.

Objectives: The objective of this study was to develop the weightage identified network of keywords (WINK) technique for selecting and utilizing keywords to perform systematic reviews more efficiently. This technique aims to improve the thoroughness and precision of evidence synthesis by employing a more rigorous approach to keyword selection.

Methods: The WINK methodology involves generating network visualization charts to analyze the interconnections among keywords within a specific domain. This process integrates both computational analysis and subject expert insights to enhance the accuracy and relevance of the findings. In the example considered, the networking strength between the contexts of environmental pollutants with endocrine function as Q1 and systemic health with oral health-related terms as Q2 was examined, and keywords with limited networking strength were excluded. Utilizing the Medical Subject Headings (MeSH) terms identified from the WINK technique, a search string was built and compared to an initial search with fewer keywords.

Results: The application of the WINK technique in building the search string yielded 69.81% and 26.23% more articles for Q1 and Q2, respectively, compared to conventional approaches. This significant increase demonstrates the technique's effectiveness in identifying relevant studies and ensuring comprehensive evidence synthesis.

Conclusions: By prioritizing keywords with higher weightage and utilizing network visualization charts, the WINK technique ensures comprehensive evidence synthesis and enhances accuracy in systematic reviews. Its effectiveness in identifying relevant studies marks a significant advancement in systematic review methodology, offering a more robust and efficient approach to keyword selection.

Objectives: The spread of misinformation through the internet can lead to dangerous behavioral changes and erode trust in reliable sources, especially during public health crises like coronavirus disease 2019 (COVID-19). To combat this issue, innovative strategies that leverage information technology are essential. This study focused on developing and evaluating a mobile application (app), COVID-Iran, aimed at countering COVID-19 misinformation by delivering accurate, reliable, and credible information.

Methods: The development of the app involved a multi-step, user-centered approach that integrated qualitative expert consultations with quantitative survey research to pinpoint and validate key features. The app was initially prototyped using Enterprise Architect software and subsequently developed using Android Studio and MySQL. We conducted a usability evaluation using the System Usability Scale (SUS), where participants engaged in various tasks related to information seeking, self-assessment, and health management. Data were analyzed using descriptive statistics in SPSS version 19.

Results: The findings revealed a high usability level (SUS score of 81.35), with participants reporting ease of use and learnability. The app effectively countered misinformation by providing access to trusted sources and evidence-based counterarguments. User feedback emphasized the app's strengths in clarity, accuracy, trustworthiness, and its comprehensive approach. Plans for future improvements include the integration of artificial intelligence to deliver personalized content.

Conclusions: Despite limitations such as the small sample size and potential self-selection bias, this study highlights the significant potential of mHealth apps to provide reliable health information and combat misinformation.

Objectives: This study was conducted to develop a mobile application for digital health coaching to support self-management in older adults with multiple chronic conditions. Additionally, the usability of this application was evaluated.

Methods: The HAHA2022 mobile application was developed through a multidisciplinary team approach, incorporating digital health coaching strategies targeting community-dwelling older adults with multiple chronic conditions. Usability was assessed with the Korean version of the Mobile Application Rating Scale. The usability tests involved eight expert panel members and 10 older adults (mean age, 74 ± 3 years; 90% women) from one senior welfare center.

Results: HAHA2022 is an Android-based mobile application that is also integrated into wearable devices to track physical activity. It features an age-friendly design and includes five main menus: Home, Action Plan, Education, Health Log, and Community. The average overall usability test scores-covering engagement, functionality, aesthetics, and information-were 4.27 of 5 for the expert panel and 4.53 of 5 for the older adults.

Conclusions: The HAHA2022 application was developed to improve self-management among communitydwelling older adults with multiple chronic conditions. Usability tests indicate that the application is highly acceptable and feasible for use by this population. Consequently, HAHA2022 is anticipated to be widely implemented. Nonetheless, further research is required to confirm its effectiveness through digital health intervention.

Objectives: This study aimed to assess the feasibility of implementing OpenNotes in Korea. It involved developing and evaluating the Open-CHA service, which provides clinical summary information to patients diagnosed with craniofacial deformities and their caregivers following outpatient visits.

Methods: The study included 109 patients diagnosed with craniofacial deformities, along with their caregivers. The Open-CHA service was developed by referencing OpenNotes and involved a user needs assessment, a pilot test, and an evaluation of its effectiveness. Data were analyzed using descriptive statistics and the paired t-test.

Results: Short message service templates for the Open-CHA service based on a user needs assessment conducted with patients, caregivers, and healthcare professionals. These templates were refined and improved following a pilot test. After the implementation of the Open-CHA service, most participants evaluated OpenNotes positively. Additionally, there were observed increases in health knowledge and efficacy in patient-physician interactions. A statistically significant improvement in mobile health literacy was also confirmed.

Conclusions: The implementation of the Open-CHA service significantly enhanced mobile health literacy among patients with craniofacial deformities and their caregivers, indicating positive outcomes for the potential adoption of OpenNotes in Korea. This suggests that introducing OpenNotes into the Korean healthcare system is appropriate.

Objectives: Health students' ability to utilize technology effectively is vital for improving the quality of future healthcare services. Relevant digital health education must be comprehensively integrated into training programs, continuing professional development activities, and school curricula to keep them current. This study investigated the most effective digital health approaches to enhance health students' cognitive, affective, and psychomotor skills, thereby preparing them for the workforce.

Methods: A literature review was conducted by searching for articles from 2013 to 2023 in PubMed, Science Direct, ERIC, and Scopus. The search used the PICO model, focusing on experimental studies and digital learning.

Results: The review identified 26 studies, categorizing digital education methods into platform-based (46.2%), tools-based (30.7%), and training-based approaches (23.1%). Participants included health students (57.7%), healthcare professionals (34.6%), and a combination of both (7.7%). The content materials primarily targeted curriculum objectives (65.4%) and clinical applications (34.6%). The outcomes, classified according to Bloom's taxonomy, were divided into cognitive (84.6%), affective (76.9%), and psychomotor (46.1%) domains.

Conclusions: Digital health education benefits from a variety of approaches. A platformbased approach is recommended for delivering theoretical and methodological materials, a tools-based approach for simulations, and a training-based approach for practical skills to enhance the cognitive domain. Both platform-based and trainingbased approaches are advised to improve the affective and psychomotor dimensions of learning. This study underscores the importance of an integrated digital learning system in health educational institutions to prepare students for evolving health systems and to improve learning outcomes and skill transfer.

Objectives: Disorders of amino acid metabolism fall under the category of inborn errors of metabolism that can be managed with a protein-restricted diet. However, adherence to such a diet often poses challenges, leading to low treatment engagement. Consequently, there is a pressing need for new resources to aid in dietary self-monitoring. The goal is to develop and implement "AminoApp," an application tailored for dietary self-monitoring in patients with inborn errors of metabolism who are on a low-protein diet.

Methods: The design and development of the application adhered to the user-centered design method. This approach emphasizes active participation and collaboration between users and designers/researchers throughout all stages of product development, including requirement gathering, prototype development, and evaluation. Usability was evaluated using the System Usability Scale, which has been validated in Portuguese.

Results: The application's features include a food diary, a food consultation area, exam records, a recipe calculator, and reports on diet composition and metabolic control. The usability test included four patients on a low-protein diet, three caregivers, and three healthcare professionals. The average usability score was 84.9, with averages of 77.5 for patients, 85.8 for caregivers, and 91.6 for professionals, indicating that the application is user-friendly.

Conclusions: AminoApp is the first application developed in Brazil designed to assist in managing inborn errors of metabolism that require a protein-restricted diet. It was found to be easy to use, and the initial results are promising. Further research is necessary to evaluate the impact of the application on metabolic control and treatment adherence.

Objectives: This study evaluated the efficacy of integrating a retrieval-augmented generation (RAG) model and a large language model (LLM) to improve the accuracy of drug name mapping across international vocabularies.

Methods: Drug ingredient names were translated into English using the Japanese Accepted Names for Pharmaceuticals. Drug concepts were extracted from the standard vocabulary of OHDSI, and the accuracy of mappings between translated terms and RxNorm was assessed by vector similarity, using the BioBERT-generated embedded vectors as the baseline. Subsequently, we developed LLMs with RAG that distinguished the final candidates from the baseline. We assessed the efficacy of the LLM with RAG in candidate selection by comparing it with conventional methods based on vector similarity.

Results: The evaluation metrics demonstrated the superior performance of the combined LLM + RAG over traditional vector similarity methods. Notably, the hit rates of the Mixtral 8x7b and GPT-3.5 models exceeded 90%, significantly outperforming the baseline rate of 64% across stratified groups of PO drugs, injections, and all interventions. Furthermore, the r-precision metric, which measures the alignment between model judgment and human evaluation, revealed a notable improvement in LLM performance, ranging from 41% to 50% compared to the baseline of 23%.

Conclusions: Integrating an RAG and an LLM outperformed conventional string comparison and embedding vector similarity techniques, offering a more refined approach to global drug information mapping.