Proceedings of the 2025 9th International Conference on Medical and Health Informatics. International Conference on Medical and Health Informatics (9th : 2025 : Kyoto, Japan)最新文献

{"title":"Risk Prediction and Interpretation for Fall Events Using Explainable AI and Large Language Models.","authors":"Jake Luo, Masoud Khani, Jazzmyne Adams, Qiang Lu, Kristian O'Connor, David R Friedland","doi":"10.1145/3761712.3761774","DOIUrl":"10.1145/3761712.3761774","url":null,"abstract":"<p><p>Falls represent a significant public health concern, particularly among older adults, with approximately 35.6 million cases annually in the United States alone. Traditional machine learning-based fall risk prediction methods often lack precision and interpretability, limiting their effectiveness in clinical settings. This study aims to develop and validate an explainable machine learning approach for fall risk prediction that combines predictive accuracy with interpretable results and leverages large language models for enhanced communication, enabling healthcare providers to make informed decisions about fall prevention strategies. We developed an integrated machine learning pipeline utilizing XGBoost classifier to process key health indicators such as age, and diagnosis history, with the model's interpretability enhanced through SHAP (SHapley Additive exPlanations) values. Results were transformed into natural language narratives using LangChain and large language models, which automated the generation of personalized reports combining both risk assessments and feature explanations, with patients classified into low (0-40%), medium (40-70%), and high (70-100%) risk groups based on predicted probabilities. The XG-Boost classifier emerged as the best-performing model, achieving 71% accuracy, 69% precision, 76% recall, an F1-score of 72%, and an ROC AUC of 0.71 on the testing dataset. The SHAP analysis provided transparent insights into feature importance, highlighting critical contributors to fall risk prediction. The integration of large language models through LangChain successfully transformed complex model outputs into comprehensible narratives, generating detailed explanations that included both high-level risk summaries and specific feature contributions for healthcare providers and patients. Our integrated approach demonstrates the feasibility of combining high-accuracy fall risk prediction with explainable AI techniques and large language models, with the system's ability to provide interpretable results and generate clear, personalized risk communications representing a significant advancement in developing practical AI-driven tools for clinical fall risk assessment, potentially improving the implementation of preventive interventions in healthcare settings.</p>","PeriodicalId":521379,"journal":{"name":"Proceedings of the 2025 9th International Conference on Medical and Health Informatics. International Conference on Medical and Health Informatics (9th : 2025 : Kyoto, Japan)","volume":"2025 ","pages":"270-277"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13112544/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147795586","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multi-Level Explainable AI for ECG-Based Atrial Fibrillation Detection: Exploring LIME, SHAP, and Grad-CAM for Clinical Interpretation.","authors":"Jake Luo, Amirsajjad Taleban, Patrick Noffke, Rodney Sparapani","doi":"10.1145/3761712.3761721","DOIUrl":"10.1145/3761712.3761721","url":null,"abstract":"<p><p>Atrial fibrillation (AFib) detection through ECG interpretation remains a critical yet complex task in clinical settings. While deep learning models have shown promise in automated ECG analysis, their \"black box\" nature limits clinical adoption. This study presents a multi-level explainable AI (XAI) framework for AFib detection, combining LIME, SHAP, and Grad-CAM approaches to provide clinically meaningful interpretations. Using the PhysioNet AFib Dataset with 5,830 ECG samples, we developed a ResNet-based model that achieved 91.3% precision and 88.7% recall for AFib detection, and 98.4% precision and 98.8% recall for normal rhythm classification. While traditional LIME and SHAP analyses provided feature importance values, their point-wise granularity proved too fine-grained for clinical interpretation. We addressed this limitation by introducing a novel RR-interval aggregation method that aligns XAI outputs with clinical diagnostic patterns. Further examination using Grad-CAM successfully localized three key morphological features: P-wave absence, pronounced T-waves, and irregular RR-intervals. This comprehensive XAI framework bridges the gap between AI predictions and clinical interpretability, offering both beat-level and temporal pattern analysis that aligns with established diagnostic criteria for AFib. Our results demonstrate that integrating multiple XAI approaches can provide clinically relevant explanations while maintaining high diagnostic accuracy.</p>","PeriodicalId":521379,"journal":{"name":"Proceedings of the 2025 9th International Conference on Medical and Health Informatics. International Conference on Medical and Health Informatics (9th : 2025 : Kyoto, Japan)","volume":"2025 ","pages":"52-59"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13112545/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147795548","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}