{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T18:05:20Z","timestamp":1790877920167,"version":"4.1.0"},"reference-count":51,"publisher":"Public Library of Science (PLoS)","issue":"10","license":[{"start":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T00:00:00Z","timestamp":1790812800000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["www.plosdigitalhealth.org"],"crossmark-restriction":false},"short-container-title":["PLOS Digit Health"],"abstract":"<jats:p>Intensive care unit (ICU) monitoring systems face a critical gap: translating continuous physiological data into interpretable patterns that support clinical assessment. Existing approaches rely on static thresholds or severity scores that fail to capture dynamic disease progression patterns and provide limited insight into why patients change over time. We developed STREAM (State Trajectory Representation and Evolution-Aware Monitoring), which models each patient as a point in a multidimensional physiological space and applies geometric analysis of routine ICU data to track physiological instability. STREAM analyzes 26 routinely collected clinical measurements using optimal transport theory to discover data-derived physiological states without predefined categories and maps each patient to their nearest state. We evaluated STREAM using the multicenter eICU Collaborative Research Database (N\u2009=\u2009158,294) for development and MIMIC-IV (N\u2009=\u200984,517) for external validation. STREAM identified five reproducible data-derived physiological states with distinct clinical signatures. Patients who spent less than 10% of their ICU stay within their expected state (state outliers) had ICU mortality of 37.6%, 16-fold higher than those who remained within their assigned states (2.3%). Mortality prediction achieved an area under the receiver operating characteristic curve of 0.863 at 8 hours and 0.903 at 72 hours, with excellent calibration (expected calibration error of 0.002). External validation on MIMIC-IV maintained robust performance (0.798 and 0.857, respectively), with state outliers exhibiting 10-fold higher mortality (33.5% vs. 3.2%). Feature importance analysis identified which laboratory values and vital signs are associated with movement toward higher-risk states, providing interpretable clinical explanations. STREAM provides transparent monitoring across data-derived physiological states, linking state dynamics with outcome prediction. Strong discrimination, calibration, and reproducibility across multicenter datasets support the method\u2019s potential for prospective evaluation.<\/jats:p>","DOI":"10.1371\/journal.pdig.0001753","type":"journal-article","created":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T17:36:12Z","timestamp":1790876172000},"page":"e0001753","update-policy":"https:\/\/doi.org\/10.1371\/journal.pdig.corrections_policy","source":"Crossref","is-referenced-by-count":0,"title":["STREAM: A data-driven framework for physiological state monitoring in ICU patients"],"prefix":"10.1371","volume":"5","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-2739-0518","authenticated-orcid":true,"given":"Ali","family":"Namvar","sequence":"first","affiliation":[{"name":"University of Michigan","place":["Ann Arbor, Michigan, United States of America"],"department":["Department of Radiology"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sundaresh","family":"Ram","sequence":"additional","affiliation":[{"name":"University of Michigan","place":["Ann Arbor, Michigan, United States of America"],"department":["Department of Radiology"]},{"name":"School of Electrical and Computer Engineering, Emory University and Georgia Institute of Technology","place":["Atlanta, Georgia, United States of America"],"department":["Department of Radiology and Imaging Sciences, Wallace H. Coulter Department of Biomedical Engineering"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Wassim W.","family":"Labaki","sequence":"additional","affiliation":[{"name":"University of Michigan","place":["Ann Arbor, Michigan, United States of America"],"department":["Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Stefanie","family":"Galban","sequence":"additional","affiliation":[{"name":"University of Michigan","place":["Ann Arbor, Michigan, United States of America"],"department":["Department of Radiology"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Njira L.","family":"Lugogo","sequence":"additional","affiliation":[{"name":"University of Michigan","place":["Ann Arbor, Michigan, United States of America"],"department":["Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5596-7487","authenticated-orcid":true,"given":"Craig J.","family":"Galban","sequence":"additional","affiliation":[{"name":"University of Michigan","place":["Ann Arbor, Michigan, United States of America"],"department":["Department of Radiology"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"340","published-online":{"date-parts":[[2026,10,1]]},"reference":[{"issue":"1","key":"pdig.0001753.ref001","doi-asserted-by":"crossref","first-page":"15","DOI":"10.1186\/s13054-022-04290-9","article-title":"The Sequential Organ Failure Assessment (SOFA) Score: Has the time come for an update?","volume":"27","author":"R Moreno","year":"2023","journal-title":"Crit Care"},{"issue":"3","key":"pdig.0001753.ref002","doi-asserted-by":"crossref","first-page":"355","DOI":"10.4103\/joacp.joacp_380_21","article-title":"Comparison of APACHE II and APACHE IV score as predictors of mortality in patients with septic shock in intensive care unit: A prospective observational study","volume":"39","author":"SD Bloria","year":"2023","journal-title":"J Anaesthesiol Clin Pharmacol"},{"key":"pdig.0001753.ref003","doi-asserted-by":"crossref","DOI":"10.3389\/fmed.2021.744907","article-title":"Dynamic APACHE II score to predict the outcome of intensive care unit patients","volume":"8","author":"Y Tian","year":"2022","journal-title":"Frontiers in Medicine"},{"issue":"1","key":"pdig.0001753.ref004","doi-asserted-by":"crossref","first-page":"290","DOI":"10.1186\/s13054-025-05532-2","article-title":"Implementing artificial intelligence in critical care medicine: a consensus of 22","volume":"29","author":"M Cecconi","year":"2025","journal-title":"Critical Care"},{"issue":"3","key":"pdig.0001753.ref005","doi-asserted-by":"crossref","DOI":"10.2196\/28781","article-title":"State of the art of machine learning-enabled clinical decision support in intensive care units: literature review","volume":"10","author":"N Hong","year":"2022","journal-title":"JMIR Med Inform"},{"key":"pdig.0001753.ref006","doi-asserted-by":"crossref","first-page":"752558","DOI":"10.3389\/frai.2021.752558","article-title":"SHAP and LIME: An evaluation of discriminative power in credit risk","volume":"4","author":"A Gramegna","year":"2021","journal-title":"Front Artif Intell"},{"issue":"3","key":"pdig.0001753.ref007","article-title":"Leveraging SHAP and LIME for enhanced explainability in AI-driven diagnostic systems","volume":"2","author":"AK Kalusivalingam","year":"2021","journal-title":"Int J AI and ML"},{"key":"pdig.0001753.ref008","doi-asserted-by":"crossref","DOI":"10.1145\/3097983.3098088","article-title":"Dipole: Diagnosis prediction in healthcare via attention-based bidirectional recurrent neural networks","volume-title":"Proceedings of the 23rd ACM SIGKDD international conference on knowledge discovery and data mining","author":"F Ma","year":"2017"},{"issue":"7260","key":"pdig.0001753.ref009","doi-asserted-by":"crossref","first-page":"53","DOI":"10.1038\/nature08227","article-title":"Early-warning signals for critical transitions","volume":"461","author":"M Scheffer","year":"2009","journal-title":"Nature"},{"issue":"4","key":"pdig.0001753.ref010","doi-asserted-by":"crossref","first-page":"195","DOI":"10.1111\/nicc.12199","article-title":"Critical illness trajectory for patients, families and nurses - a literature review","volume":"21","author":"P Page","year":"2016","journal-title":"Nurs Crit Care"},{"issue":"4","key":"pdig.0001753.ref011","doi-asserted-by":"crossref","first-page":"43","DOI":"10.1109\/MSP.2017.2695801","article-title":"Optimal mass transport: Signal processing and machine-learning applications","volume":"34","author":"S Kolouri","year":"2017","journal-title":"IEEE Signal Process Mag"},{"issue":"6","key":"pdig.0001753.ref012","doi-asserted-by":"crossref","first-page":"945","DOI":"10.1093\/bja\/aeu187","article-title":"Metabolic response to the stress of critical illness","volume":"113","author":"J-C Preiser","year":"2014","journal-title":"Br J Anaesth"},{"key":"pdig.0001753.ref013","first-page":"819528","article-title":"Metabolic acidosis and strong ion gap in critically ill patients with acute kidney injury","volume":"2014","author":"C-M Zheng","year":"2014","journal-title":"Biomed Res Int"},{"issue":"8","key":"pdig.0001753.ref014","doi-asserted-by":"crossref","first-page":"1765","DOI":"10.1097\/00003246-200208000-00015","article-title":"Thrombocytopenia in patients in the medical intensive care unit: bleeding prevalence, transfusion requirements, and outcome","volume":"30","author":"R Strauss","year":"2002","journal-title":"Crit Care Med"},{"issue":"5","key":"pdig.0001753.ref015","doi-asserted-by":"crossref","first-page":"502","DOI":"10.1007\/BF03028318","article-title":"Multiple organ dysfunction syndrome: a narrative review","volume":"48","author":"D Johnson","year":"2001","journal-title":"Can J Anaesth"},{"key":"pdig.0001753.ref016","unstructured":"Wang J, Moore R, Xie Y, Kamaleswaran R. Improving sepsis prediction model generalization with optimal transport. In: Proceedings of the 2nd machine learning for health symposium, 2022. 474\u201388."},{"issue":"4","key":"pdig.0001753.ref017","first-page":"1","article-title":"A hidden absorbing semi-Markov model for informatively censored temporal data: Learning and inference","volume":"19","author":"AM Alaa","year":"2018","journal-title":"J Mach Learn Res"},{"issue":"8","key":"pdig.0001753.ref018","doi-asserted-by":"crossref","first-page":"611","DOI":"10.1016\/S2213-2600(14)70097-9","article-title":"Subphenotypes in acute respiratory distress syndrome: Latent class analysis of data from two randomised controlled trials","volume":"2","author":"CS Calfee","year":"2014","journal-title":"Lancet Respir Med"},{"issue":"12","key":"pdig.0001753.ref019","doi-asserted-by":"crossref","first-page":"2094","DOI":"10.1007\/s00134-024-07669-0","article-title":"Distinct immune profiles and clinical outcomes in sepsis subphenotypes based on temperature trajectories","volume":"50","author":"SV Bhavani","year":"2024","journal-title":"Intensive Care Med"},{"issue":"11","key":"pdig.0001753.ref020","doi-asserted-by":"crossref","DOI":"10.1371\/journal.pone.0241558","article-title":"Commonly used clinical chemistry tests as mortality predictors: Results from two large cohort studies","volume":"15","author":"L Lind","year":"2020","journal-title":"PLoS One"},{"key":"pdig.0001753.ref021","doi-asserted-by":"crossref","first-page":"103528","DOI":"10.1016\/j.jbi.2020.103528","article-title":"Demonstrating the consequences of learning missingness patterns in early warning systems for preventative health care: A novel simulation and solution","volume":"110","author":"CE Gillies","year":"2020","journal-title":"J Biomed Inform"},{"issue":"6","key":"pdig.0001753.ref022","doi-asserted-by":"crossref","first-page":"649","DOI":"10.1164\/rccm.201406-1022OC","article-title":"Multicenter development and validation of a risk stratification tool for ward patients","volume":"190","author":"MM Churpek","year":"2014","journal-title":"Am J Respir Crit Care Med"},{"issue":"1","key":"pdig.0001753.ref023","doi-asserted-by":"crossref","first-page":"44","DOI":"10.1164\/rccm.202007-2700OC","article-title":"Detecting deteriorating patients in the hospital: Development and validation of a novel scoring system","volume":"204","author":"MAF Pimentel","year":"2021","journal-title":"Am J Respir Crit Care Med"},{"issue":"1","key":"pdig.0001753.ref024","doi-asserted-by":"crossref","first-page":"76","DOI":"10.1186\/s13054-024-04866-7","article-title":"Real-time machine learning model to predict short-term mortality in critically ill patients: development and international validation","volume":"28","author":"L Lim","year":"2024","journal-title":"Crit Care"},{"key":"pdig.0001753.ref025","doi-asserted-by":"crossref","unstructured":"Chen S, Fan J, Pishgar E, Alaei K, Placencia G, Pishgar M. Machine learning-based prediction of icu mortality in sepsis-associated acute kidney injury patients using mimic-iv database with validation from eicu database. 2025. https:\/\/doi.org\/arXiv:250217978","DOI":"10.1101\/2025.02.24.25322816"},{"issue":"1","key":"pdig.0001753.ref026","doi-asserted-by":"crossref","first-page":"27887","DOI":"10.1038\/s41598-025-13299-3","article-title":"Explainable machine learning for predicting ICU mortality in myocardial infarction patients using pseudo-dynamic data","volume":"15","author":"M Mesinovic","year":"2025","journal-title":"Sci Rep"},{"issue":"7","key":"pdig.0001753.ref027","doi-asserted-by":"crossref","first-page":"1129","DOI":"10.1513\/AnnalsATS.202006-698OC","article-title":"Evaluating a widely implemented proprietary deterioration index model among hospitalized patients with COVID-19","volume":"18","author":"K Singh","year":"2021","journal-title":"Ann Am Thorac Soc"},{"issue":"1","key":"pdig.0001753.ref028","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1038\/s41597-022-01899-x","article-title":"MIMIC-IV, a freely accessible electronic health record dataset","volume":"10","author":"AEW Johnson","year":"2023","journal-title":"Sci Data"},{"key":"pdig.0001753.ref029","doi-asserted-by":"crossref","first-page":"180178","DOI":"10.1038\/sdata.2018.178","article-title":"The eICU Collaborative Research Database, a freely available multi-center database for critical care research","volume":"5","author":"TJ Pollard","year":"2018","journal-title":"Sci Data"},{"issue":"3","key":"pdig.0001753.ref030","article-title":"mice: Multivariate imputation by chained equations inR","volume":"45","author":"van Buuren S","year":"2011","journal-title":"J Stat Soft"},{"key":"pdig.0001753.ref031","doi-asserted-by":"crossref","DOI":"10.1007\/978-3-319-10247-4","volume-title":"Data preprocessing in data mining","author":"S Garc\u00eda","year":"2015"},{"issue":"4","key":"pdig.0001753.ref032","doi-asserted-by":"crossref","first-page":"784","DOI":"10.1137\/1118101","article-title":"Calculation of the Wasserstein distance between probability distributions on the line","volume":"18","author":"SS Vallender","year":"1974","journal-title":"Theory Probab Appl"},{"key":"pdig.0001753.ref033","article-title":"Joint distribution optimal transportation for domain adaptation","volume":"30","author":"N Courty","year":"2017","journal-title":"Adv Neural Inform Process Syst"},{"issue":"4","key":"pdig.0001753.ref034","doi-asserted-by":"crossref","first-page":"417","DOI":"10.1007\/s10208-011-9093-5","article-title":"Gromov\u2013Wasserstein distances and the metric approach to object matching","volume":"11","author":"F M\u00e9moli","year":"2011","journal-title":"Found Comput Math"},{"key":"pdig.0001753.ref035","doi-asserted-by":"crossref","unstructured":"Satopaa V, Albrecht J, Irwin D, Raghavan B. Finding a \u201cKneedle\u201d in a Haystack: Detecting knee points in system behavior. In: 2011 31st International conference on distributed computing systems workshops. 2011. 166\u201371. https:\/\/doi.org\/10.1109\/icdcsw.2011.20","DOI":"10.1109\/ICDCSW.2011.20"},{"key":"pdig.0001753.ref036","doi-asserted-by":"crossref","first-page":"827","DOI":"10.1007\/978-1-4899-7488-4_196","article-title":"Gaussian mixture models","volume-title":"Encyclopedia of biometrics","author":"D Reynolds","year":"2015"},{"key":"pdig.0001753.ref037","doi-asserted-by":"crossref","unstructured":"Darken C, Moody J. Fast adaptive k-means clustering: some empirical results. In: 1990 IJCNN International Joint Conference on Neural Networks. 1990. 233\u20138 vol.2. https:\/\/doi.org\/10.1109\/ijcnn.1990.137720","DOI":"10.1109\/IJCNN.1990.137720"},{"key":"pdig.0001753.ref038","doi-asserted-by":"crossref","first-page":"107299","DOI":"10.1016\/j.sigpro.2019.107299","article-title":"Selective review of offline change point detection methods","volume":"167","author":"C Truong","year":"2020","journal-title":"Signal Process"},{"issue":"6","key":"pdig.0001753.ref039","first-page":"514","article-title":"Estimating the dimension of a linear-model","volume":"17","author":"J Andel","year":"1981","journal-title":"Kybernetika"},{"issue":"458","key":"pdig.0001753.ref040","doi-asserted-by":"crossref","first-page":"611","DOI":"10.1198\/016214502760047131","article-title":"Model-based clustering, discriminant analysis, and density estimation","volume":"97","author":"C Fraley","year":"2002","journal-title":"J Am Stat Assoc"},{"issue":"1","key":"pdig.0001753.ref041","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1016\/S0169-7439(99)00047-7","article-title":"The Mahalanobis distance","volume":"50","author":"R De Maesschalck","year":"2000","journal-title":"Chemomet Intelligent Lab Syst"},{"issue":"3","key":"pdig.0001753.ref042","doi-asserted-by":"crossref","first-page":"647","DOI":"10.1007\/s10115-013-0679-x","article-title":"Explaining prediction models and individual predictions with feature contributions","volume":"41","author":"E \u0160trumbelj","year":"2013","journal-title":"Knowl Inf Syst"},{"key":"pdig.0001753.ref043","doi-asserted-by":"crossref","unstructured":"Chen T, Guestrin C. Xgboost: A scalable tree boosting system. In: Proceedings of the 22nd ACM SIGKDD International conference on knowledge discovery and data mining. 2016.","DOI":"10.1145\/2939672.2939785"},{"key":"pdig.0001753.ref044","doi-asserted-by":"crossref","unstructured":"Zadrozny B, Elkan C. Transforming classifier scores into accurate multiclass probability estimates. In: Proceedings of the eighth ACM SIGKDD international conference on Knowledge discovery and data mining, 2002. 694\u20139. https:\/\/doi.org\/10.1145\/775047.775151","DOI":"10.1145\/775047.775151"},{"key":"pdig.0001753.ref045","first-page":"2901","article-title":"Obtaining well calibrated probabilities using bayesian binning","volume":"2015","author":"MP Naeini","year":"2015","journal-title":"Proc AAAI Conf Artif Intell"},{"key":"pdig.0001753.ref046","doi-asserted-by":"crossref","DOI":"10.1002\/9781118548387","volume-title":"Applied logistic regression","author":"DW Hosmer","year":"2013"},{"issue":"3","key":"pdig.0001753.ref047","doi-asserted-by":"crossref","first-page":"837","DOI":"10.2307\/2531595","article-title":"Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach","volume":"44","author":"ER DeLong","year":"1988","journal-title":"Biometrics"},{"issue":"2","key":"pdig.0001753.ref048","doi-asserted-by":"crossref","first-page":"157","DOI":"10.1002\/sim.2929","article-title":"Evaluating the added predictive ability of a new marker: From area under the ROC curve to reclassification and beyond","volume":"27","author":"MJ Pencina","year":"2008","journal-title":"Stat Med"},{"issue":"1","key":"pdig.0001753.ref049","doi-asserted-by":"crossref","first-page":"11","DOI":"10.1002\/sim.4085","article-title":"Extensions of net reclassification improvement calculations to measure usefulness of new biomarkers","volume":"30","author":"MJ Pencina","year":"2011","journal-title":"Stat Med"},{"key":"pdig.0001753.ref050","doi-asserted-by":"crossref","first-page":"569","DOI":"10.1007\/978-1-4612-4380-9_41","article-title":"Bootstrap methods: another look at the jackknife.","volume-title":"Kotz S, Johnson NL, editors. Breakthroughs in statistics: Methodology and distribution","author":"B Efron","year":"1992"},{"issue":"12","key":"pdig.0001753.ref051","doi-asserted-by":"crossref","first-page":"1643","DOI":"10.1002\/joc.1602","article-title":"Statistical and dynamical downscaling of the Seine basin climate for hydro-meteorological studies","volume":"27","author":"J Bo\u00e9","year":"2007","journal-title":"Int J Climatol"}],"container-title":["PLOS Digital Health"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pdig.0001753","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T17:36:21Z","timestamp":1790876181000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pdig.0001753"}},"subtitle":[],"editor":[{"given":"Jie-Zhi","family":"Cheng","sequence":"first","affiliation":[{"name":"Shanghai United Imaging Intelligence Co Ltd","place":["CHINA"]}],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2026,10,1]]},"references-count":51,"journal-issue":{"issue":"10","published-online":{"date-parts":[[2026,10,1]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pdig.0001753","relation":{},"ISSN":["2767-3170"],"issn-type":[{"value":"2767-3170","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10,1]]}}}