Studies in health technology and informatics最新文献

Developing a scalable multi-layer AI adoption model for 6P medicine (Predictive, Preventive, Personalized, Participatory, Precision-oriented, and Public-centered) requires careful consideration of computational infrastructure, data processing and integration, healthcare-specific requirements, security and privacy, performance optimization, and system interoperability. The described multi-layer architecture in this work provides a flexible conceptual model to accommodate the diverse needs of AI implementation in different healthcare domains while maintaining scalability, security, governance, and efficiency.

Medication reconciliation (MR) aims to prevent medication errors during transitions of care, particularly at hospital admission. Despite its importance, MR remains time-consuming, and existing electronic tools lack collaborative features and visual approaches. This study describes the design of a new electronic tool for MR, developed within ABiMed, a clinical decision support system for medication review and polypharmacy management. The tool follows the main steps of the MR process: the best possible medication history (BPMH) elaborated by the pharmacist is compared to the admission medication order (AMO), and discrepancies are semi-automatically identified and classified. Unresolved discrepancies can be directly sent to the prescriber. The tool builds on features included in ABiMed, such as an ontology-based structure allowing for integration in other tools, real-time collaboration between pharmacists and prescribers, visual drug data specific to the patient, and automatic execution of STOPP/START clinical guidelines. These approaches encourage shared responsibility, and support more clinically relevant and useful pharmacist interventions. The tool has yet to undergo clinical evaluation. Future work will assess usability and impact on outcomes such as time spent on MR, prescriber acceptance of interventions, and the tool will further be expanded to include more clinical guidelines.

This paper describes a two-year experience in designing, implementing, and restructuring an artificial intelligence in medicine course for first-year medical students. They had no prior training in computer science, mathematics, or clinical medical disciplines. The practical activities were organized into three categories: seminars (exercises, problems), hands-on practical work (initially, regressions; later, also neural networks), and video demonstrations. First-year evaluations highlighted difficulties in logic and ontologies, as well as a high variability in the quality of individual projects. In the second year, changes focused on applied work: ontology building exercises, direct comparison of simple neural networks with classical regression methods, and an introduction to Prompt Engineering. These adjustments led to a clear increase in performance and consistency of the final results. The paper supports the feasibility of early introduction of AI in medical training and the relevance of an iterative curriculum design, with a focus on conversational skills and guided applicative activity.

Patients' online record access (ORA) enables patients to involve their informal caregivers in care management by sharing health information, either through proxy access functionality or informally. The European Health Data Space mandates that member countries should ensure that patients can assign a proxy to have online access to their health data. In this study, we aimed to explore the current state of proxy ORA in four countries with mature ORA implementations; Sweden, Norway, Finland, and Estonia. We identified three types of proxy ORA; full proxy ORA, no proxy ORA, and controlled proxy ORA. Further guidance on ethically sound and secure proxy ORA functionality that complies with national and EU regulations and policies is warranted to ensure equal rights for citizens across Europe.

The scope of this work is to describe the overall process of assessing the compliance of the main Digital Objects produced in the OneAquaHealth project with the FAIR principles (Findability, Accessibility, Interoperability, and Reusability) via a custom FAIR Data Maturity Model. The model was designed and developed according to the project features, and in according with the One Digital Health Framework. Its goal is also to provide a tool characterized by a solid educational ground, so as to set the foundation of a timely FAIRification process for the next project steps.

This paper introduces a simple approach for assessing whether laypeople or AI-based automations can adequately substitute for expert ratings in the evaluation of AI-powered Q&A systems It employs weighted Cohen's Kappa to assess inter-rater reliability, establishing an expert agreement benchmark and comparing this to individual alternative rater-expert agreements. By visualizing these results in an inter-rater reliability matrix, it is a transparent and structured way to determine the adequacy of non-expert raters. Our findings, based on a real project, suggest that laypeople or AI, in some cases, can match or exceed expert agreement, particularly when risk aversion is a factor. The approach can be adapted to different contexts and rating attributes.

This paper examines a large-scale, collaborative data infrastructure project implemented across three Finnish wellbeing services counties in response to national health reform mandates. Using a qualitative case study approach and the European Interoperability Framework (EIF), this paper analyzes how organizational, political, semantic, and technical interoperability were jointly managed to develop standardized yet distributed data lake solutions. While legislative obligations drove the collaboration, success required close operational coordination, agile methods, and shared technical frameworks. By deploying parallel yet aligned environments, the counties achieved standardized reporting, reduced duplication, and improved cost-efficiency. The case highlights that sustainable digital health reform demands a holistic approach integrating top-down mandates with localized coordination and technical agility, ensuring all EIF dimensions are consistently addressed. Future evaluations should assess the long-term cost-effectiveness, scalability, and governance sustainability of this standardized, distributed model.

Cancer registries collect data about cancer patients, such as information about the tumor histology and progress, but tend to be incomplete in some variables, which complicates further analysis like survival probabilities. Imputation can benefit these analyses. Most imputation methods aim to learn the underlying data distribution of the available data, but often only feature-wise errors are evaluated. In this paper a new approach to evaluate the learned data distribution in case of survival analysis for two state-of-the-art imputation methods is presented. To estimate the survival probability after a cancer diagnosis Kaplan-Meier (KM) curves are used and calculated for survival cohorts. Stratifying the data using the UICC tumor stadium, we aim to evaluate the imputation quality using the comparison of the survival time probability. Two KM curves are calculated for each UICC-stage, while one curve is based on the survival time of the known UICC-stage and the other is computed for the survival times of the imputed UICC-stages. Differences in KM curves will be tested with a log-rank test, a modified Manhattan-Distance and the maximum absolute distance. The best result for all evaluation metrics is achieved for the UICC-stage II, which was imputed with the imputer Miss Forest and aligns well with the qualitative result of the plotted KM curves. Especially for the survival analysis the proposed metrics can help epidemiological researchers to choose an imputation method, which can preserve the trend of the survival probabilities.

Trustworthy health Artificial Intelligence (AI) must respect human rights and ethical standards, while ensuring AI robustness and safety. Despite the availability of general good practices, health AI developers lack a practical guide to address the construction of trustworthy AI (TAI). We introduce a TAI development framework (TAIDEV) as a reference guideline for the creation of TAI health systems. The framework core is a TAI matrix that classifies technical methods addressing the EU guideline for Trustworthy AI requirements (privacy and data governance; diversity, non-discrimination and fairness; transparency; and technical robustness and safety) across the different AI lifecycle stages (data preparation; model development, deployment and use, and model management). TAIDEV is complemented with generic, customizable example code pipelines for the different requirements with state-of-the-art AI techniques using Python. A related checklist is provided to help validate the application of different methods on new problems. The framework is validated using two open datasets, the UCI Heart Disease and the Diabetes 130-US Hospitals, with four code pipelines adapting TAIDEV for each dataset. The TAI framework and its example tutorials are provided as Open Source in the GitHub repository: https://github.com/bdslab-upv/trustworthy-ai. The TAIDEV framework provides health AI developers with an extensible theoretical development guideline with practical examples, aiming to ensure the development of ethical, robust and safe health AI and Clinical Decision Support Systems.

Traditional usability questionnaires include complex terminology and abstract technical phrasing, making them inappropriate to individuals with limited digital (health) literacy. As a result, these people are underrepresented in evaluations of digital health technologies (DHTs) in a home setting, leading to feedback that does not adequately reflect their needs and experiences. Ultimately, this contributes to the development of technologies that may not be fully accessible, usable, or effective for all users. To address this gap, a visually enhanced usability questionnaire was co-produced with people living with dementia and students. The resulting card-deck can be applied to support the conversation and evaluation of DHT for individuals with lower digital (health) literacy. Further research is needed to determine whether this visual tool effectively improves their representation and contributes to more equitable assessments of DHTs in practice.