Frontiers in digital health最新文献

Background: Pancreas transplantation, a crucial treatment for diabetes, is underutilized due to its invasiveness, strict criteria, organ scarcity, and limited centers. This highlights the need for enhanced public education and awareness through digital health platforms.

Methods: We utilized Google's AI-driven, consensus-based model and Claude AI 3.0 Opus by Anthropic to analyze public perceptions of pancreas transplantation. The top 10 websites identified by Google as of April-May 2024 were reviewed, focusing on sentiment, consensus, content readability, and complexity to develop strategies for better public engagement and understanding using digital health technologies.

Results: The top 10 websites, originating from the US and UK, showed a neutral and professional tone, targeting medical professionals and patients. Complex content was updated between 2021 and 2024, with a readability level suitable for high school to early college students. AI-driven analysis revealed strategies to increase public interest and understanding, including incorporating patient stories, simplifying medical jargon, utilizing visual aids, emphasizing quality of life improvements, showcasing research progress, facilitating patient outreach, promoting community engagement, partnering with influencers, and regularly updating content through digital health platforms.

Conclusion: To increase interest in pancreas transplantation in the era of connected health, we recommend integrating real patient experiences, simplifying medical content, using visual explanations, emphasizing post-transplant quality-of-life improvements, highlighting recent research, providing outreach opportunities, encouraging community connections, partnering with influencers, and keeping information current through digital health technologies. These methods aim to make pancreas transplantation more accessible and motivating for a diverse audience, supporting informed decision-making.

Introduction: Artificial intelligence (AI) is being developed for mental healthcare, but patients' perspectives on its use are unknown. This study examined differences in attitudes towards AI being used in mental healthcare by history of mental illness, current mental health status, demographic characteristics, and social determinants of health.

Methods: We conducted a cross-sectional survey of an online sample of 500 adults asking about general perspectives, comfort with AI, specific concerns, explainability and transparency, responsibility and trust, and the importance of relevant bioethical constructs.

Results: Multiple vulnerable subgroups perceive potential harms related to AI being used in mental healthcare, place importance on upholding bioethical constructs, and would blame or reduce trust in multiple parties, including mental healthcare professionals, if harm or conflicting assessments resulted from AI.

Discussion: Future research examining strategies for ethical AI implementation and supporting clinician AI literacy is critical for optimal patient and clinician interactions with AI in mental healthcare.

Introduction: Current preoperative exam guidelines utilize extensive lab tests, including blood tests and urine analysis, which are crucial for assessing surgical readiness. However, logistical challenges, especially for patients traveling long distances for high-quality medical care, create significant delays and burdens. This study aims to address these challenges by applying a previously developed point-of-care (POC) device system to perform accurate and rapid lab tests. This device is designed to assist both healthcare providers in resource-limited settings and patients by offering a low-cost, portable diagnostic tool that enables both in-clinic and at-home testing.

Methods: The system was tested for adaptability and compatibility by transitioning from its original Android platform to an iOS platform. A custom application was developed to maintain the system's capabilities of capturing optimal cell images across different mobile platforms. The system's cell counting algorithm was tailored to process the captured images, featuring a streamlined workflow that includes image processing and automated cell detection using a Hough circle algorithm.

Results: The new system provided good-quality raw images with 26.3 px/ $\mu$ m pixel resolution and 2.19  $\mu$ m spatial resolution, facilitating effective cell recognition and counting. The cell counting algorithm demonstrated high precision (0.8663) and high recall (0.9312), with a correlation ( $R^2=0.89535$ ) between algorithm-generated counts and actual counts.

Discussion: This study highlights the potential of the POC device to streamline preoperative testing, making it more accessible and efficient, particularly for patients in rural areas or those needing to travel for medical care. Future enhancements, including wider field-of-view, adjustable magnification, more advanced and integrated algorithms as well as integration with a microfluidic channel for direct sample analysis, are proposed to expand the device's functionality. The device's portability, ease of use, and rapid processing time position it as a promising alternative to traditional lab tests, ultimately aiming to improve patient care and surgical outcomes.

[This corrects the article DOI: 10.3389/fdgth.2024.1440701.].

This systematic review investigates the application of federated learning in mental health and human activity recognition. A comprehensive search was conducted to identify studies utilizing federated learning for these domains. The included studies were evaluated based on publication year, task, dataset characteristics, federated learning algorithms, and personalization methods. The aim is to provide an overview of the current state-of-the-art, identify research gaps, and inform future research directions in this emerging field.

Background: Methods for accurately predicting the prognosis of patients with recurrent hepatolithiasis (RH) after biliary surgery are lacking. This study aimed to develop a model that dynamically predicts the risk of hepatolithiasis recurrence using a machine-learning (ML) approach based on multiple clinical high-order correlation data.

Materials and methods: Data from patients with RH who underwent surgery at five centres between January 2015 and December 2020 were collected and divided into training and testing sets. Nine predictive models, which we named the Correlation Analysis and Recurrence Evaluation System (CARES), were developed and compared using machine learning (ML) methods to predict the patients' dynamic recurrence risk within 5 post-operative years. We adopted a k-fold cross validation with k = 10 and tested model performance on a separate testing set. The area under the receiver operating characteristic curve was used to evaluate the performance of the models, and the significance and direction of each predictive variable were interpreted and justified based on Shapley Additive Explanations.

Results: Models based on ML methods outperformed those based on traditional regression analysis in predicting the recurrent risk of patients with RH, with Extreme Gradient Boosting (XGBoost) and Light Gradient Boosting Machine (LightGBM) showing the best performance, both yielding an AUC (Area Under the receiver operating characteristic Curve) of∼0.9 or higher at predictions. These models were proved to have even better performance on testing sets than in a 10-fold cross validation, indicating that the model was not overfitted. The SHAP method revealed that immediate stone clearance, final stone clearance, number of previous surgeries, and preoperative CA19-9 index were the most important predictors of recurrence after reoperation in RH patients. An online version of the CARES model was implemented.

Conclusion: The CARES model was firstly developed based on ML methods and further encapsulated into an online version for predicting the recurrence of patients with RH after hepatectomy, which can guide clinical decision-making and personalised postoperative surveillance.

The increasing prevalence of Autonomous Systems (AS) powered by Artificial Intelligence (AI) in society and their expanding role in ensuring safety necessitate the assessment of their trustworthiness. The verification and development community faces the challenge of evaluating the trustworthiness of AI-powered AS in a comprehensive and objective manner. To address this challenge, this study conducts a semi-structured interview with experts to gather their insights and perspectives on the trustworthiness of AI-powered autonomous systems in healthcare. By integrating the expert insights, a comprehensive framework is proposed for assessing the trustworthiness of AI-powered autonomous systems in the domain of healthcare. This framework is designed to contribute to the advancement of trustworthiness assessment practices in the field of AI and autonomous systems, fostering greater confidence in their deployment in healthcare settings.

A key challenge in monitoring, managing, and mitigating global health crises is the need to coordinate clinical decision-making with systems outside of healthcare. In the 21st century, human engagement with Internet-connected ubiquitous devices generates an enormous amount of big data, which can be used to address complex, intersectoral problems via participatory epidemiology and mHealth approaches that can be operationalized with digital citizen science. These big data - which traditionally exist outside of health systems - are underutilized even though their usage can have significant implications for prediction and prevention of communicable and non-communicable diseases. To address critical challenges and gaps in big data utilization across sectors, a Digital Citizen Science Observatory (DiScO) is being developed by the Digital Epidemiology and Population Health Laboratory by scaling up existing digital health infrastructure. DiScO's development is informed by the Smart Framework, which leverages ubiquitous devices for ethical surveillance. The Observatory will be operationalized by implementing a rapidly adaptable, replicable, and scalable progressive web application that repurposes jurisdiction-specific cloud infrastructure to address crises across jurisdictions. The Observatory is designed to be highly adaptable for both rapid data collection as well as rapid responses to emerging and existing crises. Data sovereignty and decentralization of technology are core aspects of the observatory, where citizens can own the data they generate, and researchers and decision-makers can re-purpose digital health infrastructure. The ultimate aim of DiScO is to transform health systems by breaking existing jurisdictional silos in addressing global health crises.

Over the last years, studies using artificial intelligence (AI) for the detection and prediction of diseases have increased and also concentrated more and more on vulnerable groups of individuals, such as infants. The release of ChatGPT demonstrated the potential of large language models (LLMs) and heralded a new era of AI with manifold application possibilities. However, the impact of this new technology on medical research cannot be fully estimated yet. In this work, we therefore aimed to summarise the most recent pre-ChatGPT developments in the field of automated detection and prediction of diseases and disease status in infants, i.e., within the first 12 months of life. For this, we systematically searched the scientific databases PubMed and IEEE Xplore for original articles published within the last five years preceding the release of ChatGPT (2018-2022). The search revealed 927 articles; a final number of 154 articles was included for review. First of all, we examined research activity over time. Then, we analysed the articles from 2022 for medical conditions, data types, tasks, AI approaches, and reported model performance. A clear trend of increasing research activity over time could be observed. The most recently published articles focused on medical conditions of twelve different ICD-11 categories; "certain conditions originating in the perinatal period" was the most frequently addressed disease category. AI models were trained with a variety of data types, among which clinical and demographic information and laboratory data were most frequently exploited. The most frequently performed tasks aimed to detect present diseases, followed by the prediction of diseases and disease status at a later point in development. Deep neural networks turned out as the most popular AI approach, even though traditional methods, such as random forests and support vector machines, still play a role-presumably due to their explainability or better suitability when the amount of data is limited. Finally, the reported performances in many of the reviewed articles suggest that AI has the potential to assist in diagnostic procedures for infants in the near future. LLMs will boost developments in this field in the upcoming years.