Опубликована: Ноя. 26, 2024
Язык: Английский
Journal of Clinical Medicine, Год журнала: 2025, Номер 14(5), С. 1605 - 1605
Опубликована: Фев. 27, 2025
Background/Objectives: Artificial intelligence (AI) is transforming healthcare, enabling advances in diagnostics, treatment optimization, and patient care. Yet, its integration raises ethical, regulatory, societal challenges. Key concerns include data privacy risks, algorithmic bias, regulatory gaps that struggle to keep pace with AI advancements. This study aims synthesize a multidisciplinary framework for trustworthy focusing on transparency, accountability, fairness, sustainability, global collaboration. It moves beyond high-level ethical discussions provide actionable strategies implementing clinical contexts. Methods: A structured literature review was conducted using PubMed, Scopus, Web of Science. Studies were selected based relevance ethics, governance, policy prioritizing peer-reviewed articles, analyses, case studies, guidelines from authoritative sources published within the last decade. The conceptual approach integrates perspectives clinicians, ethicists, policymakers, technologists, offering holistic “ecosystem” view AI. No trials or patient-level interventions conducted. Results: analysis identifies key current governance introduces Regulatory Genome—an adaptive oversight aligned trends Sustainable Development Goals. quantifiable trustworthiness metrics, comparative categories applications, bias mitigation strategies. Additionally, it presents interdisciplinary recommendations aligning deployment environmental sustainability goals. emphasizes measurable standards, multi-stakeholder engagement strategies, partnerships ensure future innovations meet practical healthcare needs. Conclusions: Trustworthy requires more than technical advancements—it demands robust safeguards, proactive regulation, continuous By adopting recommended roadmap, stakeholders can foster responsible innovation, improve outcomes, maintain public trust AI-driven healthcare.
Язык: Английский
Процитировано
5
Опубликована: Июнь 11, 2024
With increasing electronic medical data and the development of artificial intelligence, Clinical Decision Support Systems (CDSSs) assist clinicians in diagnosis prescription. Traditional knowledge-based CDSSs follow an accumulated knowledgebase a predefined rule system, which clarifies decision-making process; however, maintenance cost issues exist quality control standardization process. Non-knowledge-based utilize vast amounts algorithms to effectively decide; deep learning black-box problem causes unreliable results. EXplainable Artificial Intelligence (XAI)-based CDSS provides valid rationale explainable It ensures trustworthiness transparency by showing recommendation prediction results process through techniques. However, existing systems have limitations, such as scope utilization lack explanatory power AI models. This study proposes new XAI-based framework address these issues; introduce resources, datasets, models that can be utilized; foundation model support various disease domains. Finally, we propose future directions for technology highlight societal need addressed emphasize potential future.
Язык: Английский
Процитировано
7
Technologies, Год журнала: 2025, Номер 13(2), С. 72 - 72
Опубликована: Фев. 8, 2025
Surgical waiting lists present significant challenges to healthcare systems, particularly in resource-constrained settings where equitable prioritization and efficient resource allocation are critical. We aim address these issues by developing a novel, dynamic, interpretable framework for prioritizing surgical patients. Our methodology integrates machine learning (ML), stochastic simulations, explainable AI (XAI) capture the temporal evolution of dynamic scores, qp(t), while ensuring transparency decision making. Specifically, we employ Light Gradient Boosting Machine (LightGBM) predictive modeling, simulations account variables competitive interactions, SHapley Additive Explanations (SHAPs) interpret model outputs at both global patient-specific levels. hybrid approach demonstrates strong performance using dataset 205 patients from an otorhinolaryngology (ENT) unit high-complexity hospital Chile. The LightGBM achieved mean squared error (MSE) 0.00018 coefficient determination (R2) value 0.96282, underscoring its high accuracy estimating qp(t). Stochastic effectively captured changes, illustrating that Patient 1’s qp(t) increased 0.50 (at t=0) 1.026 t=10) due growth such as severity urgency. SHAP analyses identified (Sever) most influential variable, contributing substantially non-clinical factors, capacity participate family activities (Lfam), exerted moderating influence. Additionally, our achieves reduction times up 26%, demonstrating effectiveness optimizing prioritization. Finally, strategy combines adaptability interpretability, transparent aligns with evolving patient needs constraints.
Язык: Английский
Процитировано
0
Опубликована: Май 8, 2025
(1) Background: The current uses of smartwatch wearable devices have expanded, not only being a part everyday routine life but also playing dynamic role in the early detection many behavioral patterns users. Furthermore, modern era, there is an increasing trend mental disturbances even adolescence, phenomenon that continues into academic life. Taking account situation, objective this systematic literature review emphasizes AI symptom burnout student population. (2) Methods: A was designed based on PRISMA guidelines. general extracted aspect to exploit all related research evidence about effectiveness (3) Results: reviewed studies document importance physiological monitoring and AI-driven predictive models, with collaboration self-reported scales assessing well-being. It reported stress most frequently studied burnout-related symptom. Meanwhile, heart rate (HR) variability (HRV) are commonly used biomarkers can be monitor evaluate detection. (4) Conclusions: Despite promising potential these technologies, several challenges limitations must addressed enhance their reliability.
Язык: Английский
Процитировано
0
Processes, Год журнала: 2024, Номер 12(12), С. 2771 - 2771
Опубликована: Дек. 5, 2024
In the context of smart cities with advanced Internet Things (IoT) systems, ensuring sustainability and safety freshwater resources is pivotal for public health urban resilience. This study introduces EWAIS (Ensemble Learning Explainable AI System), a novel framework designed monitoring assessment water quality. Leveraging strengths Ensemble models Artificial Intelligence (XAI), not only enhances prediction accuracy quality but also provides transparent insights into factors influencing these predictions. integrates multiple models—Extra Trees Classifier (ETC), K-Nearest Neighbors (KNN), AdaBoost Classifier, decision tree (DT), Stacked Ensemble, Voting (VEL)—to classify as drinkable or non-drinkable. The system incorporates techniques handling missing data statistical analysis, robust performance even in complex datasets. To address opacity traditional Machine models, employs XAI methods such SHAP LIME, generating intuitive visual explanations like force plots, summary dependency plots. achieves high predictive performance, VEL model reaching an 0.89 F1-Score 0.85, alongside precision recall scores 0.85 0.86, respectively. These results demonstrate proposed framework’s capability to deliver both accurate predictions actionable decision-makers. By providing interpretable system, supports informed management strategies, contributing well-being populations. has been validated using controlled datasets, IoT implementation suggested enhance city environments.
Язык: Английский
Процитировано
2
Journal of Clinical Medicine, Год журнала: 2024, Номер 13(19), С. 5893 - 5893
Опубликована: Окт. 2, 2024
Background: The accurate segmentation of the appendix with well-defined boundaries is critical for diagnosing conditions such as acute appendicitis. manual identification time-consuming and highly dependent on expertise radiologist. Method: In this study, we propose a fully automated approach to detection using deep learning architecture based U-Net specific training parameters in CT scans. proposed trained an annotated original dataset abdominal scans segment efficiently high performance. addition, extend set, data augmentation techniques are applied created dataset. Results: experimental studies, model implemented hyperparameter optimization performance evaluated key metrics measure diagnostic reliability. achieved slices Dice Similarity Coefficient (DSC), Volumetric Overlap Error (VOE), Average Symmetric Surface Distance (ASSD), Hausdorff 95 (HD95), Precision (PRE) Recall (REC) 85.94%, 23.29%, 1.24 mm, 5.43 86.83% 86.62%, respectively. Moreover, our outperforms other methods by leveraging U-Net’s ability capture spatial context through encoder–decoder structures skip connections, providing correct output. Conclusions: showed reliable segmenting region, some limitations cases where was close structures. These improvements highlight potential significantly improve clinical outcomes detection.
Язык: Английский
Процитировано
1
Опубликована: Ноя. 26, 2024
Язык: Английский
Процитировано
0