Health Informatics Journal最新文献

Objective: This study aimed to investigate the accuracy, reliability, and readability of A-Eye Consult, ChatGPT-4.0, Google Gemini and Copilot AI large language models (LLMs) in responding to patient questions about endophthalmitis. Methods: The LLMs' responses to 25 questions about endophthalmitis, frequently asked by patients, were evaluated by two ophthalmologists using a five-point Likert scale, with scores ranging from 1-5. The DISCERN scale assessed the reliability of the LLMs' responses, whereas the Flesch Reading Ease (FRE) and Flesch-Kincaid Grade Level (FKGL) indices assessed readability and text complexity, respectively. Results: A-Eye Consult and ChatGPT-4.0 outperformed Google Gemini and Copilot in providing comprehensive and precise responses. The Likert score significantly differed across all four LLMs (p < .001), with A-Eye Consult scoring significantly higher than Google Gemini and Copilot (p < .001). Conclusions: A-Eye Consult and ChatGPT-4.0 responses, while more complex than those of other LLMs, provided more reliable and accurate information.

Objectives: This study explores the governance intricacies in establishing a large-scale information infrastructure for integrated care within the National Health Service of England. By focusing on the initial 5 years of a regional interoperability program, we examine the challenges and dynamics of creating a unified system across organizational boundaries.

Methods: A longitudinal multi-methods approach was utilized, collecting data through interviews with health and social care personnel involved in project governance, meeting observations, and document analysis. This provided an in-depth understanding of the processes and challenges of developing the regional interoperability program.

Results: This study reveals a dynamic negotiation process and evolving governance structures that must continuously adapt to integrate diverse systems. Key findings include challenges in reconciling multiplicity and heterogeneity, the importance of social learning and stakeholder adaptation, and the role of ongoing negotiation and collaborative group construction in implementing interorganizational infrastructures.

Conclusion: This study underscores the necessity of continuous adaptation and negotiation among stakeholders in establishing effective governance of integrated care infrastructures. Social learning and active participation play pivotal roles in overcoming the challenges associated with system integration. Strategic insights from this research can guide stakeholders in developing sociotechnical solutions for digital interorganizational infrastructures in integrated care.

Background: Organisational interoperability (OIoP) of the Healthcare Information System (HIS) is crucial for the success of HIS, however little is known about the impact of institutional factors. Objective: This cross-sectional study aimed to investigate the direct effect of institutional factors on OIoP for HIS in public sector hospitals in Malaysia. Methods: A conceptual OIoP framework was developed using the Personal Health Systems Interoperability and Refined eHealth European Interoperability frameworks. A self-administered questionnaire survey was used to solicit data from 300 healthcare professionals. Data were assessed through an Exploratory Factor Analysis followed by a Confirmatory Factor Analysis. Results: Structured equation modelling revealed Security and Privacy Compliance, and Stakeholder Engagement and Awareness to significantly and positively affect OIoP (R² = 0.380). Conclusion: Healthcare organisations should prioritise clear and effective policies and regulations and enough budget and resources for the suggested framework.

Objective: Health recommendation systems suggest behavioral modifications to improve quality of life. However, current approaches do not facilitate the generation or examination of such recommendations considering the multifeature longitudinal evolution of behaviors. This paper proposes the use of a deep learning transformer-based model that allows the analysis of recommendations for behavior changes. Methods: We adapted a prediction approach, namely Behavior Sequence Transformer (BST), which analyzes temporal human routines and patterns, generating inductive outcomes. The evaluation relied on a case study that employed the behavioral history and profile of the English Longitudinal Study of Ageing (ELSA) participants (n = 2682), predicting their psychological mood (normal, pre-depressed, depressed) according to input recommendations for behavioral changes. Root mean squared error (RMSE) and learning curves were used to track the recommendation accuracy evolution and possible overfitting problems. Results: Experiments demonstrated lower RMSE values for the multifeature model (0.28/0.03) when compared to its single-feature versions (marital status, 0.59/0.001), (high pressure, 0.357/0.04), (diabetes, 0.36/0.01), (sleep quality, 0.57/0.02), (level of physical activity, 0.57/0.01). Conclusions: The results demonstrate the architecture's capability to analyze multifeatured longitudinal data, supporting the generation of suggestions for concurrent modifications across multiple input features. Moreover, these suggestions align with findings in specialized literature.

An increasing number of patients turn to YouTube for medical information, driving the growth of research on medical video content, including in dermatology. The objective was to analyze the content of YouTube videos discussing acne, psoriasis, or anti-aging skincare in Arabic. This infodemiological study analyzed the most viewed videos on these topics. A usefulness score was created to compare "useful" and "not useful" videos, along with assessments of global quality and reliability. Among the 98 most viewed videos, 75 were analyzed. Non-professionals produced 53.33%. Median scores for quality (2/5), reliability (1/5), and usefulness (4/19) were low. Most videos (78.67%) were "not useful," while 21.33% were "useful," with the latter showing significantly higher quality and reliability. In conclusion, most videos present shortcomings both in terms of quality and reliability. Videos from professional sources are far fewer in number and less popular.

Objective: Mitigation of racism in artificial intelligence (AI) is needed to improve health outcomes, yet no consensus exists on how this might be achieved. Methods: At an international conference in 2022, experts gathered to discuss strategies for reducing bias in healthcare AI. Results: This paper delineates these strategies along with their corresponding strengths and weaknesses and reviews the existing literature on these strategies. Conclusions: Five major themes resulted: reducing dataset bias, accurate modeling of existing data, transparency of artificial intelligence, regulation of artificial intelligence and the people who develop it, and bringing stakeholders to the table.

Objectives: This article reports on the implementation and evaluation of an established technology-enabled collaborative learning programme (Project ECHO) at an independent UK hospice in the North of England over a 6-year period. Methods: An independent audit of collated, anonymised data from the programme is used to report attendance patterns and session evaluations. Results: The results show a gradual increase in attendances, programmes, sessions and hours of education, coupled with consistently positive evaluation reports. Conclusion: This supports existing evidence that Project ECHO is an effective method of delivering remote healthcare education, demonstrating impact on the first three levels of Moore's education framework; participation, satisfaction and learning. Future expansion in terms of geography and topics covered is proposed, alongside enhanced evaluation methods to demonstrate impact at the higher levels of Moore's framework.

Massively parallel sequencing helps create new knowledge on genes, variants and their association with disease phenotype. This important technological advancement simultaneously makes clinical decision making, using genomic information for cancer patients, more complex. Currently, identifying actionable pathogenic variants with diagnostic, prognostic, or predictive impact requires substantial manual effort. Objective: The purpose is to design a solution for clinical diagnostics of lymphoma, specifically for systematic variant filtering and interpretation. Methods: A scoping review and demonstrations from specialists serve as a basis for a blueprint of a solution for massively parallel sequencing-based genetic diagnostics. Results: The solution uses machine learning methods to facilitate decision making in the diagnostic process. A validation round of interviews with specialists consolidated the blueprint and anchored it across all relevant expert disciplines. The scoping review identified four components of variant filtering solutions: algorithms and Artificial Intelligence (AI) applications, software, bioinformatics pipelines and variant filtering strategies. The blueprint describes the input, the AI model and the interface for dynamic browsing. Conclusion: An AI-augmented system is designed for predicting pathogenic variants. While such a system can be used to classify identified variants, diagnosticians should still evaluate the classification's accuracy, make corrections when necessary, and ultimately decide which variants are truly pathogenic.

Objectives: Rapid epidemiological investigations are fundamental to prevent the spread of infectious diseases such as coronavirus disease 2019. An epidemiological investigation presents significant challenges for both epidemiologists and infected individuals. It requires creating an environment that enables people to independently manage infectious diseases and voluntarily participate in epidemiological investigations. Methods: We developed the KODARI application, an epidemiological investigation support system that users can voluntarily use. We developed the questionnaires based on literature reviews. We evaluated the application through an online survey from December 2 to 14, 2022. Results: The application automatically or manually collect epidemiological investigation information. The application improved data accuracy through accurate information collection. It voluntarily can transmit self-management information to epidemiologist terminals or users in real time. We collected 248 users from an online survey. Most users had high ratings and willingness to use. They have willingness to manage infectious patients was substantial. The application was evaluated as helpful for epidemiological investigations and could shorten the time required for epidemiological investigations by more than 30 min. Conclusion: The application proposes a model based on people's voluntary participation. We demonstrated that the application could enhance epidemiological investigations and diminish the duration of existing epidemiological investigation processes.

Objective: Electronic Health Records (EHRs) are now an integral part of health systems in middle and high-income countries despite recognized deficits in the digital competencies of Healthcare Professionals (HCPs). Therefore, we undertook a scoping review of factors influencing compliance with EHR data privacy policies. Methods: Seven databases revealed 27 relevant studies, covering a range of countries, professional groups, and research methods. The diverse nature of these factors meant that 18 separate theoretical frameworks representing technology-acceptance to behavioral psychology were used to interpret these. Results: The predominant factors influencing compliance with EHR data privacy policies included confidence and competence to comply, perceived ease of use, facilitatory environmental factors, perceived usefulness, fear that non-compliance would be detected and/or punished and the expectations of others. Conclusion: Human factors such as attitudes, social pressure, confidence, and perceived usefulness are as important as technical factors and must be addressed to improve compliance.