[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85772-en":3,"doc-seo-85772-105":29,"detail-sidebar-cat-0-en-105":82},{"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":13,"seo_description":14,"update_tm":27,"read_time":28},85772,2336464648322,"Aria","https://ap-avatar.wpscdn.com/avatar/2200025388227c56fec?_k=1778556882303663488",7,"Healthcare","Multimodal Routing for Interpretable, Robust, and Auditable Clinical Prediction","EHR data are inherently multimodal, and leveraging multiple modalities can improve predictive performance. However, deep fusion approaches obscure how each modality contributes, limiting interpretability of multimodal reasoning in clinical settings. A multimodal routing framework is proposed for clinical prediction across structured longitudinal variables, clinical notes, and chest X-rays. The model builds discrete unimodal, directional bimodal, and trimodal routes, enabling asymmetric cross-modal effects. Inference-time route masking audits reasoning by simulating missing modalities and reweighting routes without retraining, assessing performance changes and robustness across patient groups.","Multimodal Routing for Interpretable, Robust, and Auditable  \nClinical Prediction  \nNikkie Hooman  \n[nikkieh@smu.edu](nikkieh@smu.edu)[ ](nikkieh@smu.edu)Southern Methodist University Dallas, USA  \nEric C. Larson∗  \n[eclarson@smu.edu](eclarson@smu.edu)[ ](eclarson@smu.edu)Southern Methodist University Dallas, USA  \nZhongjie Wu∗ [zhongjiew@smu.edu](zhongjiew@smu.edu)[ ](zhongjiew@smu.edu)Southern Methodist University Dallas, USA  \nMehak Gupta∗ [mehakg@smu.edu](mehakg@smu.edu)[ ](mehakg@smu.edu)Southern Methodist University Dallas, USA  \narXiv :2607 .09982v 1 [ cs .LG] 10 Jul 2026  \nAbstract  \nEHR data are inherently multimodal, and leveraging multiple modalities can improve predictive performance. However, most existing approaches rely on deep fusion, which obscures how individual modalities contribute to predictions and limits the interpretability of multimodal reasoning. We propose an explicit multimodal routing framework for clinical prediction that enables interpretable, robust, and auditable reasoning across three EHR modalities: structured longitudinal variables (􀀡), clinical notes (􀀣 ), and chest X-rays (􀀞 ) . Our model constructs discrete unimodal, directional bimodal, and trimodal routes to capture both individual modality signals and asymmetric cross-modal interactions. To audit multimodal reasoning and assess robustness, we introduce inference-time route masking, which simulates missing modalities and reweights the remaining routes without retraining. We analyze changes in performance and routing weights under these scenarios to understand model decision-making. We evaluate our framework on multi-label phenotype prediction (􀀠 = 25) and binary ICU mortality prediction using trimodal patient stays from MIMIC-IV, revealing systematic differences in modality reliance across clinical condition groups. Overall, our framework offers a transparent, auditable, and practical approach to multimodal clinical prediction, providing interpretability, robustness, and insights into how different data sources drive model decisions.  \nCCS Concepts  \n• Computing methodologies → Neural networks; Machine learning; Learning paradigms; Multimodal learning; Model interpretability; • Applied computing → Health informatics.  \nKeywords  \nMultimodal learning, EHR, routing module, missing modalities, ICU outcome prediction  \nPermission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for third-party components of this work must be honored. For all other uses, contact the owner/author(s) .  \nCHASE, City, Country  \n© 2026 Copyright held by the owner/author(s) .  \nACM ISBN 978-1-4503-XXXX-X/26/XX [https://doi.org/XXXXXXX.XXXXXXX](https://doi.org/XXXXXXX.XXXXXXX)  \nACM Reference Format:  \nNikkie Hooman, Zhongjie Wu, Eric C. Larson, and Mehak Gupta. 2026. Multimodal Routing for Interpretable, Robust, and Auditable Clinical Prediction. In Proceedings of Smart Connected Health (CHASE) . ACM, New York, NY, USA, 12 pages. [https://doi.org/XXXXXXX.XXXXXXX](https://doi.org/XXXXXXX.XXXXXXX)  \n1 Introduction  \nClinicians in Intensive Care Units (ICUs) routinely make highstakes decisions by integrating heterogeneous information, including longitudinal vital signs and laboratory measurements, clinical notes, and medical imaging. These diverse data sources are captured within Electronic Health Records (EHRs), creating a strong motivation for multimodal AI (MAI) systems that jointly model structured variables, free text, and images. Compared to unimodal approaches, multimodal models can leverage complementary clinical evidence, improve predictive performance, and better tolerate noisy or incomplete inputs [13, 23] . By supporting structured multimodal reasoning that reflects how clinicians combine evidence across sources, multimodal sy","cbCaigyPxYa5DRi5","https://ap.wps.com/l/cbCaigyPxYa5DRi5","pdf",4576063,1,12,"English","en",105,"# Introduction\n## Multimodal AI for ICU decision support\n## Limitations of deep fusion\n## Proposed multimodal routing framework\n# Evaluation and tasks\n## Dataset and clinical prediction objectives","[{\"question\":\"What tasks and dataset are used to evaluate the framework?\",\"answer\":\"The framework is evaluated on binary ICU mortality prediction and multi-label phenotype prediction across 25 conditions using trimodal patient stays from MIMIC-IV.\"}]",1784206166,30,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":77,"head_meta":79,"extra_data":81,"updated_unix":27},"multimodal-routing-for-interpretable-robust-and-auditable-clinical-prediction","",{"@graph":35,"@context":76},[36,53,67],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/healthcare/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/multimodal-routing-for-interpretable-robust-and-auditable-clinical-prediction/85772/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":61,"encodingFormat":60,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-16",true,{"@type":64,"interactionType":65,"userInteractionCount":4},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70],{"name":71,"@type":72,"acceptedAnswer":73},"What tasks and dataset are used to evaluate the framework?","Question",{"text":74,"@type":75},"The framework is evaluated on binary ICU mortality prediction and multi-label phenotype prediction across 25 conditions using trimodal patient stays from MIMIC-IV.","Answer","https://schema.org",{"og:url":51,"og:type":78,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":80,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":83},[84,88,92,96,101,106,109,113,118,121,125],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":85,"show_sort_weight":86,"slug":87},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":89,"show_sort_weight":90,"slug":91},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":93,"show_sort_weight":94,"slug":95},"Exam",70,"exam",{"id":97,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},5,"Comic",60,"comic",{"id":102,"doc_module":4,"doc_module_name":45,"category_name":103,"show_sort_weight":104,"slug":105},6,"Technology",50,"technology",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":107,"slug":108},40,"healthcare",{"id":110,"doc_module":4,"doc_module_name":45,"category_name":111,"show_sort_weight":28,"slug":112},8,"Research & Report","research-report",{"id":114,"doc_module":4,"doc_module_name":45,"category_name":115,"show_sort_weight":116,"slug":117},9,"Religion & Spirituality",20,"religion-spirituality",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":119,"show_sort_weight":116,"slug":120},"World Cup","world-cup",{"id":122,"doc_module":4,"doc_module_name":45,"category_name":123,"show_sort_weight":122,"slug":124},10,"Lifestyle","lifestyle",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":127,"show_sort_weight":97,"slug":128},19,"General","general"]