[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-120700-en":3,"doc-seo-120700-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},120700,687197100911,"Himbo","https://ap-avatar.wpscdn.com/avatar/a000239b6f1da00475?x-image-process=image/resize,m_fixed,w_180,h_180&k=1785132997149421697",7,"Healthcare","Hospital Length of Stay Prediction Based on Multi-modal Data towards Trustworthy Human-AI Collaboration in Radiomics - explainable machine learning and survival analysis","Hospital length-of-stay (LoS) prediction from a single X-ray image is addressed by benchmarking machine-learning survival models on a newly created multimodal dataset of 1235 images paired with human-annotated textual radiology reports. Although black-box approaches tend to achieve better average performance than interpretable baselines such as Cox proportional hazards, they remain difficult to understand. Time-dependent model explanations are introduced to enable trustworthy human-AI decision making. Using both human-labeled and algorithm-extracted radiomics features delivers actionable clinical insights for physicians and supports generalizability. Reproducibility is ensured by releasing open-source code and the tlos dataset.","arXiv :2303 .09817v1 [ cs .CV] 17 Mar 2023  \nHospital Length of Stay Prediction Based on Multi-modal Data towards Trustworthy Human-AI Collaboration in Radiomics  \nHubert Baniecki 1 ;2[0000􀀀0001􀀀6661􀀀5364], Bartlomiej Sobieski2 ,  \nPrzemyslaw Bombi􀀓nski2 ;3 , Patryk Szatkowski2 ;3 , and  \nPrzemyslaw Biecek 1 ;2[0000􀀀0001􀀀8423􀀀1823]  \n1 MI2 .AI, University of Warsaw, Warsaw, Poland  \n[h.baniecki@uw.edu.pl](h.baniecki@uw.edu.pl)  \n2 MI2 .AI, Warsaw University of Technology, Warsaw, Poland  \n[przemyslaw.biecek@pw.edu.pl](przemyslaw.biecek@pw.edu.pl)  \n3 Medical University of Warsaw, Warsaw, Poland  \nAbstract. To what extent can the patient's length of stay in a hospital be predicted using only an X-ray image? We answer this question by comparing the performance of machine learning survival models on a novel multi-modal dataset created from 1235 images with textual radiology reports annotated by humans. Although black-box models predict better on average than interpretable ones, like Cox proportional hazards, they are not inherently understandable. To overcome this trust issue, we introduce time-dependent model explanations into the human-AI decision making process. Explaining models built on both: human-annotated and algorithm-extracted radiomics features provides valuable insights for physicians working in a hospital. We believe the presented approach to be general and widely applicable to other time-to-event medical use cases. For reproducibility, we open-source code and the tlos dataset at [https://github.com/mi2datalab/xlungs-trustworthy-los-prediction](https://github.com/mi2datalab/xlungs-trustworthy-los-prediction).  \nKeywords: explainable AI · survival analysis · healthcare · radiology · interpretable machine learning  \n1 Introduction  \nPredicting patients' hospital length of stay (LoS) is a challenging task supporting the day-to-day decisions of medical doctors and nurses [6] . For example, accurate LoS prediction can increase hospital service e􀀎ciency, cutting costs and improving patient care. Historically, white-box statistical learning methods were used to estimate the anticipated LoS [2] . These provide clear reasoning behind the prediction, which is especially important in medical applications requiring stakeholders to comprehend \\Why?\" [11] . Nowadays, advancements in machine and deep learning for healthcare provide valuable improvements in the performance of predicting LoS [8, 17 , 18] . The natural drawback of using not inherently interpretable black-box models is their complex nature [1, 11] . Indeed, a recent  \n2 H. Baniecki et al.  \nsystematic review on the exact topic of hospital LoS prediction concludes with a concrete statement that there are no studies on the explainability of black-box models predicting LoS [14], a matter of high importance for diverse stakeholders involved in this healthcare process. Therefore in this paper, we demonstrate the applicability of explainable machine learning methods [1,7] in the LoS prediction task as an enabler towards trustworthy human-AI collaboration.  \nContribution. We summarize our contributions as follows. In Section 2, we introduce a novel task of hospital LoS prediction based on multi-modal X-ray data and benchmark on it machine learning survival models. To achieve this, we create the tlos dataset by manually annotating 1235 X-ray textual radiology reports from one of the Polish hospitals resulting in 17 interpretable features. Moreover, we include the state-of-the-art radiomics features extracted from images and critically evaluate their predictive performance. In Section 3, we put recent advancements in time-dependent explainable machine learning to practice. In that, we explain the best-performing models to gain insights into the importance of features and their e􀀋ects on LoS prediction. Analysing complementary explanations leads to an improved human understanding of AI, e.g. allows discovering bias in a model, increasing trust. We conclude with a discussion on t","cbCain6vPJpiyGCE","https://ap.wps.com/l/cbCain6vPJpiyGCE","pdf",2031712,1,10,"English","en",105,"# Introduction\n## Contribution\n## Related work","[{\"question\":\"How does the study predict hospital length of stay (LoS)?\",\"answer\":\"It benchmarks machine-learning survival models using a multimodal dataset created from X-ray images and human-annotated textual radiology reports.\"},{\"question\":\"Why are model explanations needed in this task?\",\"answer\":\"Black-box models can predict better but are not inherently understandable, creating trust issues for clinical stakeholders.\"},{\"question\":\"What kind of interpretability approach is introduced?\",\"answer\":\"The paper introduces time-dependent model explanations, including explanations for models using both human-annotated and algorithm-extracted radiomics features.\"}]","Hospital Length of Stay Prediction Based on Multi-modal Data towards Trustworthy Human-AI Collaboration in Radiomics - explainable machine learning and survival analysis | PDF",1785731610,25,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"hospital-length-of-stay-prediction-based-on-multi-modal-data-towards-trustworthy-human-ai-collaboration-in-radiomics-explainable-machine-learning-and-survival-analysis","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/healthcare/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/hospital-length-of-stay-prediction-based-on-multi-modal-data-towards-trustworthy-human-ai-collaboration-in-radiomics-explainable-machine-learning-and-survival-analysis/120700/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"How does the study predict hospital length of stay (LoS)?","Question",{"text":75,"@type":76},"It benchmarks machine-learning survival models using a multimodal dataset created from X-ray images and human-annotated textual radiology reports.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Why are model explanations needed in this task?",{"text":80,"@type":76},"Black-box models can predict better but are not inherently understandable, creating trust issues for clinical stakeholders.",{"name":82,"@type":73,"acceptedAnswer":83},"What kind of interpretability approach is introduced?",{"text":84,"@type":76},"The paper introduces time-dependent model explanations, including explanations for models using both human-annotated and algorithm-extracted radiomics features.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,118,123,128,131,134],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":116,"slug":117},40,"healthcare",{"id":119,"doc_module":4,"doc_module_name":46,"category_name":120,"show_sort_weight":121,"slug":122},8,"Research & Report",30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":21,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":21,"slug":133},"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]