Healthcare Informatics Research最新文献

Objectives: Diabetic foot ulcer (DFU) is a critical complication of diabetes that can lead to severe outcomes such as infection, amputation, and increased mortality if left untreated. Early detection and continuous monitoring are essential but remain challenging, especially in resource-limited settings such as India. This study developed and validated a deep learning algorithm to classify diabetic foot images into severity grades based on the International Working Group on the Diabetic Foot classification: grade 0 (healthy), grade 1 (mild), grade 2 (moderate), and grade 3 (severe).

Methods: A dataset of 407 clinical images was collected from open-source platforms and clinics in South India and expanded to 612 images through data augmentation. The dataset was divided into training (70%), validation (15%), and testing (15%) subsets. Multiple machine learning models were tested, including MobileNet_V2, EfficientNet-b0, DenseNet121, ResNet_50, VGG16, and ViT_b_16.

Results: Among the evaluated models, MobileNet_V2 demonstrated the highest validation accuracy (82%) and achieved an F1-score of 79% on the test set. Although the model showed strong training accuracy, minor overfitting was observed, particularly in distinguishing adjacent severity grades. To address this, dropout, batch normalization, and early stopping were employed. Overall, the model generalized well, showing high accuracy in detecting healthy cases and acceptable performance across ulcer severity grades.

Conclusions: This study underscores the potential of machine learning-based tools to support frontline healthcare workers and facilitate patient self-monitoring in low-resource environments. Future work will focus on refining the model and integrating it into user-friendly applications.

Objectives: The effects of maternal coffee consumption during pregnancy on childhood allergic diseases (ADs) remain insufficiently established. This study aimed to investigate the association between maternal coffee consumption during pregnancy and the risk of ADs in offspring up to 36 months of age.

Methods: We analyzed data from 3,252 mother-child pairs enrolled in the Korean Children's Environmental health Study (Ko-CHENS). Maternal coffee and caffeine intake were assessed using a food frequency questionnaire. Childhood ADs were identified based on caregiver reports of physician diagnoses. Cox proportional hazards models were applied to estimate hazard ratios (HRs) and 95% confidence intervals (CIs), with adjustment for potential confounding factors.

Results: Overall, two-thirds (67.5%) of children were reported to have at least one AD, with cumulative incidences at 36 months of age of 47.8% for atopic dermatitis, 23.9% for food allergy, 30.2% for allergic rhinitis, and 2.4% for asthma. Compared with no coffee intake, maternal coffee consumption of <1 serving/day was associated with a reduced risk of atopic dermatitis (HR = 0.89, 95% CI 0.81-0.99, p = 0.045) and food allergy (HR = 0.86, 95% CI 0.74-1.00, p = 0.061). Higher intake (≥1 serving/day) was significantly associated with a lower risk of food allergy (HR = 0.61, 95% CI 0.42-0.88, p = 0.009). No significant associations were observed for asthma or allergic rhinitis.

Conclusions: Mild maternal coffee intake during pregnancy may be associated with a reduced risk of specific ADs in early childhood.

Objectives: This study aimed to develop an optimized ensemble learning model to improve the prediction of hypertension complicated by coronary heart disease (CHD) through advanced feature selection and classifier fusion, thereby enhancing both accuracy and stability in risk assessment.

Methods: We constructed an ensemble-based predictive model using voting fusion to enhance early detection of hypertension complicated by CHD. The dataset comprised 2,487 patients with essential hypertension (EH) complicated by CHD and 3,904 non-CHD controls. Following data preprocessing procedures, including data cleaning and univariate and multivariate feature selection, an 18-dimensional feature set was derived. Five machine learning algorithms (logistic regression, random forest, XGBoost, CatBoost, and CART) were trained independently and subsequently integrated through a voting ensemble to optimize predictive performance.

Results: The voting fusion model outperformed all individual classifiers, achieving an area under the curve of 0.906 and an accuracy of 0.888 in predicting EH complicated by CHD.

Conclusions: The proposed ensemble model improves classification accuracy and robustness, offering a clinically useful tool for early risk stratification of hypertension-associated CHD. Although the model demonstrates strong predictive performance using cross-sectional data, its reliance on single-timepoint measurements and selected control populations necessitates further validation. Pending additional studies, this framework may serve as a supplementary decision-support tool within clinical informatics systems.

Objectives: This study aimed to develop and evaluate the functionality of a smart system-based prototype, "BABAT TB," a medication box designed to assist tuberculosis (TB) patients in adhering to their treatment schedules.

Methods: The development of the BABAT TB prototype followed the Design Science Research Methodology framework, encompassing the stages of problem identification and motivation, defining the objectives for a solution, and system design and development. Problem identification and motivation were established through semi-structured interviews with TB program officers and document analysis. The prototype integrates two main functional components: a drug quantity monitoring module and a reminder/alarm system for medication schedules, both monitored in real time. Serial communication through a SIM register is used to transmit real-time drug quantity data to the associated application. The system is powered by two 4,000 mAh lithium batteries, providing up to 2 months of use without recharging.

Results: The prototype consists of three core hardware components: the input control circuit, the timer circuit, and the drug amount detection circuit. All modules were successfully assembled and powered. The timer was configured according to medical prescriptions, and the alarm activated at the scheduled times, effectively reminding patients to take their medication.

Conclusions: The BABAT TB prototype effectively measures medication quantities and provides timely alerts, thereby supporting adherence to TB treatment. In addition, it can transmit data related to drug quantities, consultation schedules, and prototype identity cards (IDs) to a database.

Objectives: This study aims to increase the effectiveness of cervical cancer treatment by developing a survival prediction model using an innovative ensemble machine learning approach, namely the selective stacking technique.

Methods: Patient data obtained from the Faculty of Medicine, Chiang Mai University, Thailand, were utilized to validate the real-world applicability of the proposed approach. The selective stacking model employed a two-stage machine learning framework in which outputs from base machine learning models were systematically combined through meta-level learning. Importantly, the performance of the proposed model was compared with that reported in previous studies that relied on individual machine learning algorithms as baselines. To provide deeper insight into the predictive mechanisms of the model, local interpretable model-agnostic explanations were applied to assess feature importance and identify the most influential factors contributing to model predictions.

Results: The classification model developed using the selective stacking technique demonstrated a marked improvement in prediction accuracy, achieving an accuracy of 91.41%. The regression model also showed robust performance, with a root mean square error of 18.92 and an r value of 0.669. Feature importance analysis indicated that side effect status involving surrounding organs emerged as the most influential factor in survival prediction.

Conclusions: The selective stacking model exhibited superior predictive performance compared with the base models, suggesting that this approach offers a promising strategy for cervical cancer survival prediction and may support the development of more personalized treatment planning.

Objectives: This study aimed to investigate the experience, knowledge, and educational needs regarding standard healthcare terminologies-specifically Systematized Nomenclature of Medicine Clinical Terms (SNOMED CT), Logical Observation Identifiers Names and Codes (LOINC), and the International Classification of Diseases/Korean Classification of Diseases (ICD/KCD)-among professionals in the clinical, industrial, and academic sectors in Korea.

Methods: A descriptive survey was conducted using an online questionnaire distributed between November 21 and December 5, 2023. The questionnaire included items on participants' experiences with, self-reported knowledge of, and educational needs for standard terminologies. A total of 325 responses were analyzed.

Results: Participants reported the highest levels of experience and knowledge with ICD/KCD, whereas knowledge of SNOMED CT and LOINC was relatively low. Statistically significant differences in knowledge were observed across professional groups (p < 0.05), with terminology experts reporting higher levels than others. Educational needs were greatest for ICD/KCD and SNOMED CT, particularly in data collection and use case application. The most frequently cited barriers to adopting standard terminologies were a lack of training programs, the cost and time required for training, and the structural complexity of the terminologies.

Conclusions: These findings underscore the importance of customized and systematic educational strategies to promote the use of standard terminologies. Policy-level support, standardized training materials, and the preparation of qualified trainers are essential to enhance semantic interoperability and data utilization in Korean healthcare.

Objectives: This study aimed to explore nurses' perceptions of, and utilization plans for, companion robots to support the physical and mental health of patients with acute stroke. In addition, the study sought to provide foundational data for the development of companion robots tailored to acute stroke patients. It also investigated obstructive factors and potential solutions to difficulties encountered when applying companion robots in the care of patients with acute stroke.

Methods: Using the Delphi technique, this study surveyed 14 nurses working in the neurology ward and stroke intensive care unit of a tertiary hospital in Seoul across three survey rounds.

Results: After completion of the three Delphi survey rounds, Cronbach's α was 0.78, and stability values were all below 0.5; therefore, no additional rounds were conducted. A total of 54 items were finally selected, including 10 items related to educational aspects for nurses and patients, 12 items addressing impacts on nurses and patients, 19 items describing companion robot functions required for stroke patients, and 13 items identifying the most appropriate design elements.

Conclusions: Companion robots are expected to contribute to the physical and emotional care of patients with acute stroke admitted to tertiary hospitals by functioning as a nursing intervention, while also reducing nurses' workload, improving the quality of nursing care, and supporting patient safety management. In addition, efforts should be made to ensure the harmonious control and utilization of newly developed robots and to strengthen robot-related job competencies among nurses.

Objectives: This study aims to explore the key factors that enhance engagement in Facebook health support groups among Australian culturally and linguistically diverse (CALD) communities.

Methods: A cross-sectional online survey was conducted using convenience sampling. A total of 1,145 CALD participants residing in New South Wales, Australia, were initially recruited. From this sample, 150 participants who self-reported regular engagement with Facebook health support groups were included in the final analysis. A pilot test (n = 30) demonstrated strong internal consistency (Cronbach's alpha >0.70). Data collection involved a structured questionnaire employing a 7-point Likert scale to assess factors such as motivation, trust, perceived support (received and given), social connectedness, and sense of virtual community.

Results: Motivation and trust significantly influenced both support dynamics and the perceived sense of virtual community. The sense of virtual community, in turn, strongly predicted engagement in Facebook health support groups. Interestingly, social connectedness alone was not a significant predictor of engagement.

Conclusions: Fostering a strong sense of virtual community appears to be a critical factor in encouraging sustained engagement in digital health platforms among CALD populations.