Evaluating sepsis watch generalizability through multisite external validation of a sepsis machine learning model

Abstract Sepsis accounts for a substantial portion of global deaths and healthcare costs. The objective of this reproducibility study is to validate Duke Health’s Sepsis Watch ML model, in a new community healthcare setting and assess its performance and clinical utility in early sepsis detection at...

Full description

Saved in:
Bibliographic Details
Main Authors: Bruno Valan, Anusha Prakash, William Ratliff, Michael Gao, Srikanth Muthya, Ajit Thomas, Jennifer L. Eaton, Matt Gardner, Marshall Nichols, Mike Revoir, Dustin Tart, Cara O’Brien, Manesh Patel, Suresh Balu, Mark Sendak
Format: Article
Language:English
Published: Nature Portfolio 2025-06-01
Series:npj Digital Medicine
Online Access:https://doi.org/10.1038/s41746-025-01664-5
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850218535876296704
author Bruno Valan
Anusha Prakash
William Ratliff
Michael Gao
Srikanth Muthya
Ajit Thomas
Jennifer L. Eaton
Matt Gardner
Marshall Nichols
Mike Revoir
Dustin Tart
Cara O’Brien
Manesh Patel
Suresh Balu
Mark Sendak
author_facet Bruno Valan
Anusha Prakash
William Ratliff
Michael Gao
Srikanth Muthya
Ajit Thomas
Jennifer L. Eaton
Matt Gardner
Marshall Nichols
Mike Revoir
Dustin Tart
Cara O’Brien
Manesh Patel
Suresh Balu
Mark Sendak
author_sort Bruno Valan
collection DOAJ
description Abstract Sepsis accounts for a substantial portion of global deaths and healthcare costs. The objective of this reproducibility study is to validate Duke Health’s Sepsis Watch ML model, in a new community healthcare setting and assess its performance and clinical utility in early sepsis detection at Summa Health’s emergency departments. The study analyzed the model’s ability to predict sepsis using a combination of static and dynamic patient data using 205,005 encounters between 2020 and 2021 from 101,584 unique patients. 54.7% (n = 112,223) patients were female and the average age was 50 (IQR [38,71]). The AUROC ranged from 0.906 to 0.960, and the AUPRC ranged from 0.177 to 0.252 across the four sites. Ultimately, the reproducibility of the Sepsis Watch model in a community health system setting confirmed its strong and robust performance and portability across different geographical and demographic contexts with little variation.
format Article
id doaj-art-8201502c15d045d7abbb23ab561b87ac
institution OA Journals
issn 2398-6352
language English
publishDate 2025-06-01
publisher Nature Portfolio
record_format Article
series npj Digital Medicine
spelling doaj-art-8201502c15d045d7abbb23ab561b87ac2025-08-20T02:07:41ZengNature Portfolionpj Digital Medicine2398-63522025-06-018111010.1038/s41746-025-01664-5Evaluating sepsis watch generalizability through multisite external validation of a sepsis machine learning modelBruno Valan0Anusha Prakash1William Ratliff2Michael Gao3Srikanth Muthya4Ajit Thomas5Jennifer L. Eaton6Matt Gardner7Marshall Nichols8Mike Revoir9Dustin Tart10Cara O’Brien11Manesh Patel12Suresh Balu13Mark Sendak14Duke Institute for Health InnovationDuke Institute for Health InnovationDuke Institute for Health InnovationDuke Institute for Health InnovationCohere Med IncCohere Med IncSumma Health Research & InnovationDuke Institute for Health InnovationDuke Institute for Health InnovationDuke Institute for Health InnovationDuke University HospitalDuke University HospitalDepartment of Medicine, Duke University School of MedicineDuke Institute for Health InnovationDuke Institute for Health InnovationAbstract Sepsis accounts for a substantial portion of global deaths and healthcare costs. The objective of this reproducibility study is to validate Duke Health’s Sepsis Watch ML model, in a new community healthcare setting and assess its performance and clinical utility in early sepsis detection at Summa Health’s emergency departments. The study analyzed the model’s ability to predict sepsis using a combination of static and dynamic patient data using 205,005 encounters between 2020 and 2021 from 101,584 unique patients. 54.7% (n = 112,223) patients were female and the average age was 50 (IQR [38,71]). The AUROC ranged from 0.906 to 0.960, and the AUPRC ranged from 0.177 to 0.252 across the four sites. Ultimately, the reproducibility of the Sepsis Watch model in a community health system setting confirmed its strong and robust performance and portability across different geographical and demographic contexts with little variation.https://doi.org/10.1038/s41746-025-01664-5
spellingShingle Bruno Valan
Anusha Prakash
William Ratliff
Michael Gao
Srikanth Muthya
Ajit Thomas
Jennifer L. Eaton
Matt Gardner
Marshall Nichols
Mike Revoir
Dustin Tart
Cara O’Brien
Manesh Patel
Suresh Balu
Mark Sendak
Evaluating sepsis watch generalizability through multisite external validation of a sepsis machine learning model
npj Digital Medicine
title Evaluating sepsis watch generalizability through multisite external validation of a sepsis machine learning model
title_full Evaluating sepsis watch generalizability through multisite external validation of a sepsis machine learning model
title_fullStr Evaluating sepsis watch generalizability through multisite external validation of a sepsis machine learning model
title_full_unstemmed Evaluating sepsis watch generalizability through multisite external validation of a sepsis machine learning model
title_short Evaluating sepsis watch generalizability through multisite external validation of a sepsis machine learning model
title_sort evaluating sepsis watch generalizability through multisite external validation of a sepsis machine learning model
url https://doi.org/10.1038/s41746-025-01664-5
work_keys_str_mv AT brunovalan evaluatingsepsiswatchgeneralizabilitythroughmultisiteexternalvalidationofasepsismachinelearningmodel
AT anushaprakash evaluatingsepsiswatchgeneralizabilitythroughmultisiteexternalvalidationofasepsismachinelearningmodel
AT williamratliff evaluatingsepsiswatchgeneralizabilitythroughmultisiteexternalvalidationofasepsismachinelearningmodel
AT michaelgao evaluatingsepsiswatchgeneralizabilitythroughmultisiteexternalvalidationofasepsismachinelearningmodel
AT srikanthmuthya evaluatingsepsiswatchgeneralizabilitythroughmultisiteexternalvalidationofasepsismachinelearningmodel
AT ajitthomas evaluatingsepsiswatchgeneralizabilitythroughmultisiteexternalvalidationofasepsismachinelearningmodel
AT jenniferleaton evaluatingsepsiswatchgeneralizabilitythroughmultisiteexternalvalidationofasepsismachinelearningmodel
AT mattgardner evaluatingsepsiswatchgeneralizabilitythroughmultisiteexternalvalidationofasepsismachinelearningmodel
AT marshallnichols evaluatingsepsiswatchgeneralizabilitythroughmultisiteexternalvalidationofasepsismachinelearningmodel
AT mikerevoir evaluatingsepsiswatchgeneralizabilitythroughmultisiteexternalvalidationofasepsismachinelearningmodel
AT dustintart evaluatingsepsiswatchgeneralizabilitythroughmultisiteexternalvalidationofasepsismachinelearningmodel
AT caraobrien evaluatingsepsiswatchgeneralizabilitythroughmultisiteexternalvalidationofasepsismachinelearningmodel
AT maneshpatel evaluatingsepsiswatchgeneralizabilitythroughmultisiteexternalvalidationofasepsismachinelearningmodel
AT sureshbalu evaluatingsepsiswatchgeneralizabilitythroughmultisiteexternalvalidationofasepsismachinelearningmodel
AT marksendak evaluatingsepsiswatchgeneralizabilitythroughmultisiteexternalvalidationofasepsismachinelearningmodel