{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"medRxiv"}],"indexed":{"date-parts":[[2026,2,3]],"date-time":"2026-02-03T20:32:57Z","timestamp":1770150777061,"version":"3.49.0"},"posted":{"date-parts":[[2026,1,30]]},"group-title":"Intensive Care and Critical Care Medicine","reference-count":34,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2026,1,30]],"date-time":"2026-01-30T00:00:00Z","timestamp":1769731200000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2026,1,30]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                <jats:sec>\n                  <jats:title>Purpose<\/jats:title>\n                  <jats:p>Ventilator-associated pneumonia (VAP) remains one of the most serious hospital-acquired infections in the intensive care unit (ICU), with high morbidity and mortality. Early identification of patients at risk for developing VAP could enable timely diagnostics and intervention. However, current clinical tools are limited in their ability to detect early physiologic signals preceding VAP onset. We aimed to build supervised machine learning models to predict short term onset of VAP.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Methods<\/jats:title>\n                  <jats:p>We analyzed electronic health record data from a prospective observational cohort of ICU patients, where VAP was adjudicated using a standardized published protocol by a panel of critical care physicians. Clinical features (including vital signs, ventilator settings, laboratory values, and support devices) were extracted for each patient-ICU-day. We explored unsupervised clustering to characterize feature dynamics associated with VAP onset. We built multiple machine learning models across different prediction windows (3, 5, 7 days before VAP). We examined model performance in two external cohorts, MIMIC-IV and secondary analysis of the AMIKINHAL trial. Results were evaluated with discrimination metrics such as AUROC.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>The internal cohort included 507 patients with BAL-confirmed diagnoses: 261 developed VAP and 246 did not have VAP. Visualization using clustering identified distinct physiologic states enriched for VAP-labeled days. The best-performing model achieved an AUROC of 0.866 in predicting VAP up to seven days before clinical diagnosis. Temporal model probability trajectories showed rising model confidence in the days leading up to VAP. On external validation in MIMIC-IV, the best model achieved an AUROC of 0.817 for forecasting VAP within five days. There was low feature overlap with the AMIKINHAL trial data, leading to poor model performance. Feature analysis revealed that platelet count, positive end-expiratory pressure (PEEP), ventilator duration, and inflammatory markers were key drivers of model predictions.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Conclusions<\/jats:title>\n                  <jats:p>Machine learning models trained on routinely collected ICU data with careful labeling can anticipate VAP onset up to a week in advance with strong predictive performance. Model performance generalized to data from an entirely different hospital system despite differences in practice and labeling patterns, but did not perform well when there was poor feature overlap. Future work should focus on real-time prospective evaluation.<\/jats:p>\n                <\/jats:sec>","DOI":"10.64898\/2026.01.28.26344858","type":"posted-content","created":{"date-parts":[[2026,1,30]],"date-time":"2026-01-30T22:41:31Z","timestamp":1769812891000},"source":"Crossref","is-referenced-by-count":0,"title":["Developing and externally validating machine learning models to forecast short-term risk of ventilator-associated pneumonia"],"prefix":"10.64898","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-9082-4000","authenticated-orcid":false,"given":"Alec K.","family":"Peltekian","sequence":"first","affiliation":[{"name":"Department of Computer Science, Northwestern University McCormick School of Engineering and Applied Science, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0001-5070-5610","authenticated-orcid":false,"given":"Wan-Ting","family":"Liao","sequence":"additional","affiliation":[{"name":"Division of Pulmonary and Critical Care, Northwestern University Feinberg School of Medicine, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9132-514X","authenticated-orcid":false,"given":"Vijeeth","family":"Guggilla","sequence":"additional","affiliation":[{"name":"Division of Biostatistics and Informatics, Northwestern University Feinberg School of Medicine, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3659-4387","authenticated-orcid":false,"given":"Nikolay","family":"Markov","sequence":"additional","affiliation":[{"name":"Division of Pulmonary and Critical Care, Northwestern University Feinberg School of Medicine, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9942-3529","authenticated-orcid":false,"given":"Karolina","family":"Senkow","sequence":"additional","affiliation":[{"name":"Division of Pulmonary and Critical Care, Northwestern University Feinberg School of Medicine, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0009-0490-2339","authenticated-orcid":false,"given":"Zewei","family":"Liao","sequence":"additional","affiliation":[{"name":"Division of Pulmonary and Critical Care, Department of Medicine, University of Chicago, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1679-9473","authenticated-orcid":false,"given":"Marjorie","family":"Kang","sequence":"additional","affiliation":[{"name":"Division of Pulmonary and Critical Care, Northwestern University Feinberg School of Medicine, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4497-8049","authenticated-orcid":false,"given":"Luke","family":"Rasmussen","sequence":"additional","affiliation":[{"name":"Division of Biostatistics and Informatics, Northwestern University Feinberg School of Medicine, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0798-1182","authenticated-orcid":false,"given":"Elsa","family":"Tavernier","sequence":"additional","affiliation":[{"name":"CIC, INSERM 1415, CHRU Tours, Tours, France"},{"name":"Methods in Patients-Centered Outcomes and Health Research SPHERE, INSERM 1246, Tours and Nantes, France"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6221-4467","authenticated-orcid":false,"given":"Stephan","family":"Ehrmann","sequence":"additional","affiliation":[{"name":"M\u00e9decine Intensive R\u00e9animation, INSERM CIC1415, CRICS-TriggerSEP F-CRIN researche network, CHRU de Tours, Tours, France"},{"name":"Centre d\u2019\u00e9tude des pathologies respiratoires, INSERM U1100, Universit\u00e9 de Tours, Tours, France"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8197-5103","authenticated-orcid":false,"given":"Rebecca K.","family":"Clepp","sequence":"additional","affiliation":[{"name":"Division of Pulmonary and Critical Care, Northwestern University Feinberg School of Medicine, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5540-4278","authenticated-orcid":false,"given":"Thomas","family":"Stoeger","sequence":"additional","affiliation":[{"name":"Division of Pulmonary and Critical Care, Northwestern University Feinberg School of Medicine, Chicago, IL, USA"},{"name":"The Potocsnak Longevity Institute, Northwestern University, Chicago, IL, USA"},{"name":"Simpson Querrey Lung Institute for Translational Science, Northwestern University Feinberg School of Medicine, Chicago, IL, USA"},{"name":"NSF-Simons National Institute for Theory and Mathematics in Biology, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7653-3650","authenticated-orcid":false,"given":"Theresa","family":"Walunas","sequence":"additional","affiliation":[{"name":"Division of Biostatistics and Informatics, Northwestern University Feinberg School of Medicine, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8152-6319","authenticated-orcid":false,"given":"Alok","family":"Choudhary","sequence":"additional","affiliation":[{"name":"Department of Computer Science, Northwestern University McCormick School of Engineering and Applied Science, Chicago, IL, USA"},{"name":"Department of Electrical and Computer Engineering, Northwestern University McCormick School of Engineering and Applied Science, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2879-3789","authenticated-orcid":false,"given":"Alexander V.","family":"Misharin","sequence":"additional","affiliation":[{"name":"Division of Pulmonary and Critical Care, Northwestern University Feinberg School of Medicine, Chicago, IL, USA"},{"name":"Simpson Querrey Lung Institute for Translational Science, Northwestern University Feinberg School of Medicine, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5775-8427","authenticated-orcid":false,"given":"Benjamin D.","family":"Singer","sequence":"additional","affiliation":[{"name":"Division of Pulmonary and Critical Care, Northwestern University Feinberg School of Medicine, Chicago, IL, USA"},{"name":"Simpson Querrey Lung Institute for Translational Science, Northwestern University Feinberg School of Medicine, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3114-5208","authenticated-orcid":false,"given":"GR Scott","family":"Budinger","sequence":"additional","affiliation":[{"name":"Division of Pulmonary and Critical Care, Northwestern University Feinberg School of Medicine, Chicago, IL, USA"},{"name":"Simpson Querrey Lung Institute for Translational Science, Northwestern University Feinberg School of Medicine, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8527-4195","authenticated-orcid":false,"given":"Richard G.","family":"Wunderink","sequence":"additional","affiliation":[{"name":"Division of Pulmonary and Critical Care, Northwestern University Feinberg School of Medicine, Chicago, IL, USA"},{"name":"Simpson Querrey Lung Institute for Translational Science, Northwestern University Feinberg School of Medicine, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5576-3943","authenticated-orcid":false,"given":"Catherine A.","family":"Gao","sequence":"additional","affiliation":[{"name":"Division of Pulmonary and Critical Care, Northwestern University Feinberg School of Medicine, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5519-0302","authenticated-orcid":false,"given":"Ankit","family":"Agrawal","sequence":"additional","affiliation":[{"name":"Department of Electrical and Computer Engineering, Northwestern University McCormick School of Engineering and Applied Science, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"name":"The NU SCRIPT Study Investigators","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"54368","reference":[{"key":"2026020300351073000_2026.01.28.26344858v1.1","unstructured":"G. A. Kohbodi , Venkat Rajasurya , and Asif Noor . Ventilator-Associated Pneumonia. Stat-Pearls Publishing, Treasure Island (FL), 2018. URL https:\/\/www.ncbi.nlm.nih.gov\/books\/NBK507711\/. StatPearls [Internet]."},{"issue":"1","key":"2026020300351073000_2026.01.28.26344858v1.2","first-page":"2024","article-title":"Ventilator-associated pneumonia: pathobiological heterogeneity and diagnostic challenges","volume":"15","year":"6447","journal-title":"Nature communications"},{"key":"2026020300351073000_2026.01.28.26344858v1.3","doi-asserted-by":"publisher","DOI":"10.1128\/CMR.00051-05"},{"key":"2026020300351073000_2026.01.28.26344858v1.4","doi-asserted-by":"publisher","DOI":"10.1086\/653062"},{"key":"2026020300351073000_2026.01.28.26344858v1.5","doi-asserted-by":"publisher","DOI":"10.1056\/NEJMoa2305958"},{"key":"2026020300351073000_2026.01.28.26344858v1.6","doi-asserted-by":"crossref","unstructured":"Silvia Minozzi , Silvia Pifferi , Luca Brazzi , Valentina Pecoraro , Giorgia Montrucchio , and Roberto D\u2019Amico . Topical antibiotic prophylaxis to reduce respiratory tract infections and mortality in adults receiving mechanical ventilation. Cochrane Database of Systematic Reviews, 2021.","DOI":"10.1002\/14651858.CD000022.pub4"},{"issue":"5","key":"2026020300351073000_2026.01.28.26344858v1.7","doi-asserted-by":"crossref","first-page":"375","DOI":"10.1016\/S2213-2600(23)00471-X","article-title":"Ceftriaxone to prevent early ventilator-associated pneumonia in patients with acute brain injury: a multicentre, randomised, double-blind, placebo-controlled, assessor-masked superiority trial","volume":"12","year":"2024","journal-title":"The Lancet Respiratory Medicine"},{"key":"2026020300351073000_2026.01.28.26344858v1.8","doi-asserted-by":"crossref","unstructured":"Qi Zhang , Chunjie Li , Helen V. Worthington , Fang Hua , et al. Oral hygiene care for critically ill patients to prevent ventilator-associated pneumonia. Cochrane Database of Systematic Reviews, 2020.","DOI":"10.1002\/14651858.CD008367.pub4"},{"key":"2026020300351073000_2026.01.28.26344858v1.9","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-022-01894-0"},{"key":"2026020300351073000_2026.01.28.26344858v1.10","doi-asserted-by":"crossref","first-page":"579774","DOI":"10.3389\/fdata.2020.579774","article-title":"Machine learning methods to predict acute respiratory failure and acute respiratory distress syndrome","volume":"3","year":"2020","journal-title":"Frontiers in big Data"},{"issue":"1","key":"2026020300351073000_2026.01.28.26344858v1.11","doi-asserted-by":"crossref","first-page":"125","DOI":"10.1038\/s43856-022-00180-x","article-title":"Interpretable machine learning prediction of all-cause mortality","volume":"2","year":"2022","journal-title":"Communications medicine"},{"issue":"23","key":"2026020300351073000_2026.01.28.26344858v1.12","doi-asserted-by":"crossref","first-page":"e26246","DOI":"10.1097\/MD.0000000000026246","article-title":"Predicting ventilator-associated pneumonia with machine learning","volume":"100","year":"2021","journal-title":"Medicine"},{"issue":"Abstracts","key":"2026020300351073000_2026.01.28.26344858v1.13","doi-asserted-by":"crossref","first-page":"A7723","DOI":"10.1164\/ajrccm.2025.211.Abstracts.A7723","article-title":"Machine learning to predict the onset of ventilator-associated pneumonia using electronic health record data: A7723","volume":"211","year":"2025","journal-title":"American Journal of Respiratory and Critical Care Medicine"},{"key":"2026020300351073000_2026.01.28.26344858v1.14","unstructured":"Northwestern University. Successful clinical response in pneumonia therapy. https:\/\/script.northwestern.edu\/, 2025. Accessed September 12, 2025."},{"key":"2026020300351073000_2026.01.28.26344858v1.15","unstructured":"Nikolay Markov , Catherine A Gao , Thomas Stoeger , Anna Pawlowski , Mengjia Kang , Prasanth Nannapaneni , Rogan Grant , Luke Rasmussen , Daniel Schneider , Justin Starren , Richard Wunderink , G R Scott Budinger , Alexander Misharin , Benjamin Singer , and Study Investigators, NU SCRIPT. SCRIPT CarpeDiem dataset: demographics, outcomes, and per-day clinical parameters for critically ill patients with suspected pneumonia, 2023."},{"key":"2026020300351073000_2026.01.28.26344858v1.16","unstructured":"Catherine A Gao , Nikolay S Markov , Thomas Stoeger , Anna E Pawlowski , Mengjia Kang , Prasanth Nannapaneni , Rogan A Grant , Chiagozie Pickens , James M Walter , Jacqueline M Kruser , Luke V Rasmussen , Daniel Schneider , Justin Starren , Helen K Donnelly , Alvaro Donayre , Yuan Luo , G R Scott Budinger , Richard G Wunderink , Alexander V Misharin , and Benjamin D Singer . Machine learning links unresolving secondary pneumonia to mortality in patients with severe pneumonia, including COVID-19. J. Clin. Invest., April 2023."},{"key":"2026020300351073000_2026.01.28.26344858v1.17","doi-asserted-by":"publisher","DOI":"10.1093\/ofid\/ofad336"},{"key":"2026020300351073000_2026.01.28.26344858v1.18","first-page":"0176","article-title":"and Missing Data in Electronic health Records (MINDER) Group","volume":"4","year":"2024","journal-title":"Moving beyond medical statistics: A systematic review on missing data handling in electronic health records. Health Data Science"},{"key":"2026020300351073000_2026.01.28.26344858v1.19","doi-asserted-by":"crossref","first-page":"27552","DOI":"10.1038\/s41598-025-12264-4","article-title":"Developing and validating machine learning models to predict next-day extubation","volume":"15","year":"2025","journal-title":"Scientific Reports"},{"key":"2026020300351073000_2026.01.28.26344858v1.20","doi-asserted-by":"publisher","DOI":"10.1093\/jamia\/ocaa139"},{"key":"2026020300351073000_2026.01.28.26344858v1.21","first-page":"A865","article-title":"Can serial evaluation of sofa score predict developing ventilator-associated pneumonia?","volume":"18","year":"2010","journal-title":"In Annual Meeting of the American Society Anesthesiologists"},{"key":"2026020300351073000_2026.01.28.26344858v1.22","first-page":"1","article-title":"Introduction to the logistic regression model","volume":"2","year":"2000","journal-title":"Applied logistic regression"},{"key":"2026020300351073000_2026.01.28.26344858v1.23","doi-asserted-by":"publisher","DOI":"10.1023\/A:1010933404324"},{"key":"2026020300351073000_2026.01.28.26344858v1.24","doi-asserted-by":"publisher","DOI":"10.1145\/2939672.2939785"},{"key":"2026020300351073000_2026.01.28.26344858v1.25","first-page":"2825","article-title":"Scikit-learn: Machine learning in python","volume":"12","year":"2011","journal-title":"the Journal of machine Learning research"},{"key":"2026020300351073000_2026.01.28.26344858v1.26","doi-asserted-by":"crossref","unstructured":"Takuya Akiba , Shotaro Sano , Toshihiko Yanase , Takeru Ohta , and Masanori Koyama . Optuna: A next-generation hyperparameter optimization framework. In Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining, pages 2623\u20132631, 2019.","DOI":"10.1145\/3292500.3330701"},{"key":"2026020300351073000_2026.01.28.26344858v1.27","doi-asserted-by":"crossref","unstructured":"Bradley Efron and Robert J Tibshirani . An introduction to the bootstrap. Chapman and Hall\/CRC, 1994.","DOI":"10.1201\/9780429246593"},{"key":"2026020300351073000_2026.01.28.26344858v1.28","unstructured":"Gary S Collins , Karel GM Moons , Paula Dhiman , Richard D Riley , Andrew L Beam , Ben Van Calster , Marzyeh Ghassemi , Xiaoxuan Liu , Johannes B Reitsma , Maarten Van Smeden , et al. Tripod+ ai statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods. bmj, 385, 2024."},{"key":"2026020300351073000_2026.01.28.26344858v1.29","unstructured":"Alistair Johnson , Lucas Bulgarelli , Tom Pollard , Steven Horng , Leo Anthony Celi , and Roger Mark . Mimic-iv, 2020. URL https:\/\/physionet.org\/content\/mimiciv\/1.0\/. PhysioNet."},{"issue":"10","key":"2026020300351073000_2026.01.28.26344858v1.30","first-page":"2025","article-title":"An innovative deep learning approach for ventilator-associated pneumonia (vap) prediction in intensive care units\u2014pneumonia risk evaluation and diagnostic intelligence via computational technology (predict)","volume":"14","year":"3380","journal-title":"Journal of Clinical Medicine"},{"issue":"3","key":"2026020300351073000_2026.01.28.26344858v1.31","doi-asserted-by":"crossref","first-page":"100167","DOI":"10.1016\/j.infpip.2021.100167","article-title":"Automated healthcare-associated infection surveillance using an artificial intelligence algorithm","volume":"3","year":"2021","journal-title":"Infection Prevention in Practice"},{"key":"2026020300351073000_2026.01.28.26344858v1.32","doi-asserted-by":"crossref","first-page":"76","DOI":"10.1016\/j.ejim.2023.11.009","article-title":"Han Shi Jocelyn Chew, and Miia Jansson. Early prediction of ventilator-associated pneumonia with machine learning models: A systematic review and meta-analysis of prediction model performance","volume":"121","year":"2024","journal-title":"European Journal of Internal Medicine"},{"issue":"1","key":"2026020300351073000_2026.01.28.26344858v1.33","first-page":"e57026","article-title":"Ventilator-associated pneumonia prediction models based on ai: scoping review","volume":"12","year":"2024","journal-title":"JMIR Medical Informatics"},{"issue":"1","key":"2026020300351073000_2026.01.28.26344858v1.34","doi-asserted-by":"crossref","first-page":"250","DOI":"10.1186\/s12890-022-02031-w","article-title":"Early prediction of ventilator-associated pneumonia in critical care patients: a machine learning model","volume":"22","year":"2022","journal-title":"BMC Pulmonary Medicine"}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.64898\/2026.01.28.26344858","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,2,3]],"date-time":"2026-02-03T08:35:20Z","timestamp":1770107720000},"score":1,"resource":{"primary":{"URL":"http:\/\/medrxiv.org\/lookup\/doi\/10.64898\/2026.01.28.26344858"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,1,30]]},"references-count":34,"URL":"https:\/\/doi.org\/10.64898\/2026.01.28.26344858","relation":{},"subject":[],"published":{"date-parts":[[2026,1,30]]},"subtype":"preprint"}}