{"id":"https://openalex.org/W7204200717","doi":"https://doi.org/10.64898/2026.08.22.26361090","title":"Offline Reinforcement Learning for Out-of-Distribution ICU Sepsis Decision Support","display_name":"Offline Reinforcement Learning for Out-of-Distribution ICU Sepsis Decision Support","publication_year":2026,"publication_date":"2026-08-25","ids":{"openalex":"https://openalex.org/W7204200717","doi":"https://doi.org/10.64898/2026.08.22.26361090"},"language":"en","primary_location":{"id":"doi:10.64898/2026.08.22.26361090","doi":"https://doi.org/10.64898/2026.08.22.26361090","is_oa":true,"landing_page_url":"https://doi.org/10.64898/2026.08.22.26361090","pdf_url":"https://www.medrxiv.org/content/medrxiv/early/2026/08/25/2026.08.22.26361090.full.pdf","source":{"id":"https://openalex.org/S3005729997","display_name":"medRxiv","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["cwts-core"],"host_organization":null,"host_organization_name":null,"host_organization_lineage":["https://openalex.org/P4310315909"],"host_organization_lineage_names":["Cold Spring Harbor Laboratory Press"],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"medRxiv","raw_type":"posted-content"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://www.medrxiv.org/content/medrxiv/early/2026/08/25/2026.08.22.26361090.full.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5044991836","display_name":"Emad Arasteh","orcid":"https://orcid.org/0009-0001-8072-9322","observed_orcids":["https://orcid.org/0009-0001-8072-9322"]},"institutions":[{"id":"https://openalex.org/I83019370","display_name":"Eindhoven University of Technology","ror":"https://ror.org/02c2kyt77","country_code":"NL","type":"education","lineage":["https://openalex.org/I83019370"]}],"countries":["NL"],"is_corresponding":false,"raw_author_name":"Emad Arasteh","raw_affiliation_strings":["Department of Mathematics and Computer Science, Eindhoven University of Technology, Eindhoven, the Netherlands"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Mathematics and Computer Science, Eindhoven University of Technology, Eindhoven, the Netherlands","institution_ids":["https://openalex.org/I83019370"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5024703221","display_name":"Maryam S. Mirian","orcid":"https://orcid.org/0000-0001-5411-9902","observed_orcids":["https://orcid.org/0000-0001-5411-9902"]},"institutions":[{"id":"https://openalex.org/I141945490","display_name":"University of British Columbia","ror":"https://ror.org/03rmrcq20","country_code":"CA","type":"education","lineage":["https://openalex.org/I141945490"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Maryam S. Mirian","raw_affiliation_strings":["Pacific Parkinson\u2019s Research Centre, University of British Columbia, Vancouver, Canada"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Pacific Parkinson\u2019s Research Centre, University of British Columbia, Vancouver, Canada","institution_ids":["https://openalex.org/I141945490"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5089317938","display_name":"Maryam Tavakol","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I83019370","display_name":"Eindhoven University of Technology","ror":"https://ror.org/02c2kyt77","country_code":"NL","type":"education","lineage":["https://openalex.org/I83019370"]}],"countries":["NL"],"is_corresponding":true,"raw_author_name":"Maryam Tavakol","raw_affiliation_strings":["Department of Mathematics and Computer Science, Eindhoven University of Technology, Eindhoven, the Netherlands"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Mathematics and Computer Science, Eindhoven University of Technology, Eindhoven, the Netherlands","institution_ids":["https://openalex.org/I83019370"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"corresponding_author_ids":["https://openalex.org/A5089317938"],"corresponding_institution_ids":["https://openalex.org/I83019370"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T13702","display_name":"Machine Learning in Healthcare","score":0.9067000150680542,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T13702","display_name":"Machine Learning in Healthcare","score":0.9067000150680542,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10218","display_name":"Sepsis Diagnosis and Treatment","score":0.07460000365972519,"subfield":{"id":"https://openalex.org/subfields/2706","display_name":"Critical Care and Intensive Care Medicine"},"field":{"id":"https://openalex.org/fields/27","display_name":"Medicine"},"domain":{"id":"https://openalex.org/domains/4","display_name":"Health Sciences"}},{"id":"https://openalex.org/T10462","display_name":"Reinforcement Learning in Robotics","score":0.008500000461935997,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/offline-reinforcement-learning","display_name":"offline reinforcement learning","score":0.9990000128746033},{"id":"https://openalex.org/keywords/out-of-distribution-generalization","display_name":"out-of-distribution generalization","score":0.8740000128746033},{"id":"https://openalex.org/keywords/clinical-decision-support","display_name":"clinical decision support","score":0.8420000076293945},{"id":"https://openalex.org/keywords/off-policy-evaluation","display_name":"off-policy evaluation","score":0.9330000281333923},{"id":"https://openalex.org/keywords/learned-dynamics","display_name":"learned dynamics","score":0.7250000238418579},{"id":"https://openalex.org/keywords/mimic-iii","display_name":"MIMIC-III","score":0.9279999732971191},{"id":"https://openalex.org/keywords/survival-prediction","display_name":"survival prediction","score":0.4320000112056732}],"concepts":[{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.713100016117096},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.6485999822616577},{"id":"https://openalex.org/C67203356","wikidata":"https://www.wikidata.org/wiki/Q1321905","display_name":"Reinforcement","level":2,"score":0.5030999779701233},{"id":"https://openalex.org/C173801870","wikidata":"https://www.wikidata.org/wiki/Q201413","display_name":"Heuristic","level":2,"score":0.4643000066280365},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.44339999556541443},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4099999964237213},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3882000148296356},{"id":"https://openalex.org/C71924100","wikidata":"https://www.wikidata.org/wiki/Q11190","display_name":"Medicine","level":0,"score":0.3747999966144562},{"id":"https://openalex.org/C2777267654","wikidata":"https://www.wikidata.org/wiki/Q3519023","display_name":"Test (biology)","level":2,"score":0.36489999294281006},{"id":"https://openalex.org/C2780490138","wikidata":"https://www.wikidata.org/wiki/Q7079636","display_name":"Offline learning","level":3,"score":0.35589998960494995},{"id":"https://openalex.org/C2778384902","wikidata":"https://www.wikidata.org/wiki/Q183134","display_name":"Sepsis","level":2,"score":0.34540000557899475},{"id":"https://openalex.org/C774472","wikidata":"https://www.wikidata.org/wiki/Q6760393","display_name":"Margin (machine learning)","level":2,"score":0.3440999984741211},{"id":"https://openalex.org/C2780102126","wikidata":"https://www.wikidata.org/wiki/Q10928179","display_name":"Online and offline","level":2,"score":0.3422999978065491},{"id":"https://openalex.org/C535046627","wikidata":"https://www.wikidata.org/wiki/Q30612","display_name":"Clinical trial","level":2,"score":0.3253999948501587},{"id":"https://openalex.org/C2776502983","wikidata":"https://www.wikidata.org/wiki/Q690182","display_name":"Contrast (vision)","level":2,"score":0.3253999948501587},{"id":"https://openalex.org/C107327155","wikidata":"https://www.wikidata.org/wiki/Q330268","display_name":"Decision support system","level":2,"score":0.319599986076355},{"id":"https://openalex.org/C188816634","wikidata":"https://www.wikidata.org/wiki/Q2113324","display_name":"Prospective cohort study","level":2,"score":0.31709998846054077},{"id":"https://openalex.org/C45804977","wikidata":"https://www.wikidata.org/wiki/Q7239673","display_name":"Predictive modelling","level":2,"score":0.3021000027656555},{"id":"https://openalex.org/C10515644","wikidata":"https://www.wikidata.org/wiki/Q543310","display_name":"Survival analysis","level":2,"score":0.2953000068664551},{"id":"https://openalex.org/C2989179672","wikidata":"https://www.wikidata.org/wiki/Q6806500","display_name":"Clinical decision making","level":2,"score":0.28839999437332153},{"id":"https://openalex.org/C177713679","wikidata":"https://www.wikidata.org/wiki/Q679690","display_name":"Intensive care medicine","level":1,"score":0.28610000014305115},{"id":"https://openalex.org/C168563851","wikidata":"https://www.wikidata.org/wiki/Q1436668","display_name":"Randomized controlled trial","level":2,"score":0.2718999981880188},{"id":"https://openalex.org/C2779664074","wikidata":"https://www.wikidata.org/wiki/Q3518405","display_name":"Terminal (telecommunication)","level":2,"score":0.26600000262260437},{"id":"https://openalex.org/C2991859549","wikidata":"https://www.wikidata.org/wiki/Q679690","display_name":"Critically ill","level":2,"score":0.2639999985694885},{"id":"https://openalex.org/C89611455","wikidata":"https://www.wikidata.org/wiki/Q6804646","display_name":"Mechanism (biology)","level":2,"score":0.2563000023365021}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.64898/2026.08.22.26361090","doi":"https://doi.org/10.64898/2026.08.22.26361090","is_oa":true,"landing_page_url":"https://doi.org/10.64898/2026.08.22.26361090","pdf_url":"https://www.medrxiv.org/content/medrxiv/early/2026/08/25/2026.08.22.26361090.full.pdf","source":{"id":"https://openalex.org/S3005729997","display_name":"medRxiv","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["cwts-core"],"host_organization":null,"host_organization_name":null,"host_organization_lineage":["https://openalex.org/P4310315909"],"host_organization_lineage_names":["Cold Spring Harbor Laboratory Press"],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"medRxiv","raw_type":"posted-content"}],"best_oa_location":{"id":"doi:10.64898/2026.08.22.26361090","doi":"https://doi.org/10.64898/2026.08.22.26361090","is_oa":true,"landing_page_url":"https://doi.org/10.64898/2026.08.22.26361090","pdf_url":"https://www.medrxiv.org/content/medrxiv/early/2026/08/25/2026.08.22.26361090.full.pdf","source":{"id":"https://openalex.org/S3005729997","display_name":"medRxiv","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["cwts-core"],"host_organization":null,"host_organization_name":null,"host_organization_lineage":["https://openalex.org/P4310315909"],"host_organization_lineage_names":["Cold Spring Harbor Laboratory Press"],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"medRxiv","raw_type":"posted-content"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/3","display_name":"Good health and well-being","score":0.8199999928474426}],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7204200717.pdf","grobid_xml":"https://content.openalex.org/works/W7204200717.grobid-xml"},"referenced_works_count":3,"referenced_works":["https://openalex.org/W2508921413","https://openalex.org/W2896893468","https://openalex.org/W2904730732"],"related_works":["https://openalex.org/W7160869870","https://openalex.org/W7155356233","https://openalex.org/W4407693924","https://openalex.org/W7203660277","https://openalex.org/W7169873026","https://openalex.org/W7126253308","https://openalex.org/W4224881322","https://openalex.org/W7138893057","https://openalex.org/W7154188127","https://openalex.org/W3010701194"],"abstract_inverted_index":{"Abstract":[0],"Offline":[1],"reinforcement":[2],"learning":[3,10],"(RL)":[4],"provides":[5],"a":[6,78,141,152,164,206],"promising":[7],"framework":[8],"for":[9,195,211],"and":[11,124,131,221],"evaluating":[12],"treatment":[13],"policies":[14,41,76,185,214],"from":[15,67,96,104],"logged":[16,192],"clinical":[17,101,129,193],"data,":[18],"particularly":[19],"in":[20,113,173,215],"sequential":[21],"decision-making":[22],"settings":[23],"where":[24],"prospective":[25,220],"exploration":[26],"would":[27],"be":[28],"unsafe.":[29],"In":[30,52,178],"ICU":[31,216],"sepsis":[32],"management,":[33],"however,":[34],"it":[35],"remains":[36],"unclear":[37],"whether":[38,74,168],"offline":[39,58,75,111,184,212],"RL":[40,59,213],"retain":[42,77],"stable":[43,142],"behavior":[44],"under":[45,146,183],"increasingly":[46],"severe":[47],"out-of-distribution":[48],"(OOD)":[49],"patient":[50],"cohorts.":[51],"this":[53,138,179],"paper,":[54],"we":[55],"evaluate":[56],"standard":[57],"methods":[60],"on":[61],"three":[62],"severity-enriched":[63],"OOD":[64],"test":[65,210],"mixtures":[66],"the":[68,84,92,109],"MIMIC-III":[69],"benchmark":[70],"dataset":[71],"to":[72,98,106],"determine":[73],"stable,":[79],"action-sensitive":[80],"decision-support":[81,144],"signal.":[82],"Under":[83],"shared":[85],"learned-dynamics":[86,203],"off-policy":[87],"evaluation":[88],"(OPE)":[89],"protocol,":[90],"as":[91,205,224],"severe-OOD":[93],"ratio":[94],"increases":[95],"25%":[97],"75%,":[99],"observed":[100,128],"survival":[102,119,130,134],"declines":[103],"67%":[105],"49%,":[107],"while":[108,218],"best":[110],"method":[112],"each":[114],"mixture":[115],"receives":[116],"model-predicted":[117,132],"terminal":[118,133],"values":[120,189],"of":[121,167],"87%,":[122],"86%,":[123],"85%,":[125],"respectively.":[126],"Because":[127],"are":[135],"different":[136],"quantities,":[137],"contrast":[139],"suggests":[140],"model-based":[143],"signal":[145],"severity":[147],"shift.":[148],"We":[149],"further":[150],"present":[151],"secondary":[153],"physiological":[154,160,170,197],"stabilization":[155,161],"analysis":[156],"using":[157],"an":[158],"episode-level":[159],"score":[162],"(EPSS),":[163],"heuristic":[165],"summary":[166],"selected":[169],"variables":[171],"move":[172],"favorable":[174],"directions":[175],"during":[176],"follow-up.":[177],"analysis,":[180],"model-generated":[181],"rollouts":[182],"receive":[186],"higher":[187],"EPSS":[188],"than":[190],"matched":[191],"trajectories":[194],"several":[196],"components.":[198],"Together,":[199],"these":[200],"results":[201],"support":[202],"OPE":[204],"useful":[207],"severity-OOD":[208],"stress":[209],"sepsis,":[217],"leaving":[219],"causal":[222],"validation":[223],"necessary":[225],"next":[226],"steps.":[227]},"counts_by_year":[],"updated_date":"2026-11-10T11:39:43.021295","created_date":"2026-08-26T00:00:00"}