Computing network-based features from physiological time series

Application to sepsis detection

Sabato Santaniello, Stephen J. Granite, Sridevi V. Sarma, Raimond Winslow

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Sepsis is a systemic deleterious host response to infection. It is a major healthcare problem that affects millions of patients every year in the intensive care units (ICUs) worldwide. Despite the fact that ICU patients are heavily instrumented with physiological sensors, early sepsis detection remains challenging, perhaps because clinicians identify sepsis by using static scores derived from bed-side measurements individually, i.e., without systematically accounting for potential interactions between these signals and their dynamics. In this study, we apply network-based data analysis to take into account interactions between bed-side physiological time series (PTS) data collected in ICU patients, and we investigate features to distinguish between sepsis and non-sepsis conditions. We treated each PTS source as a node on a graph and we retrieved the graph connectivity matrix over time by tracking the correlation between each pair of sources' signals over consecutive time windows. Then, for each connectivity matrix, we computed the eigenvalue decomposition. We found that, even though raw PTS measurements may have indistinguishable distributions in non-sepsis and early sepsis states, the median /I of the eigenvalues computed from the same data is statistically different (p <0.001) in the two states and the evolution of /I may reflect the disease progression. Although preliminary, these findings suggest that network-based features computed from continuous PTS data may be useful for early sepsis detection.

Original languageEnglish (US)
Title of host publication2014 36th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2014
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages3825-3826
Number of pages2
ISBN (Print)9781424479290
DOIs
StatePublished - Nov 2 2014
Event2014 36th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2014 - Chicago, United States
Duration: Aug 26 2014Aug 30 2014

Other

Other2014 36th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2014
CountryUnited States
CityChicago
Period8/26/148/30/14

Fingerprint

Intensive care units
Time series
Sepsis
Intensive Care Units
Decomposition
Disease Progression
Sensors
Delivery of Health Care
Infection

ASJC Scopus subject areas

  • Health Informatics
  • Computer Science Applications
  • Biomedical Engineering

Cite this

Santaniello, S., Granite, S. J., Sarma, S. V., & Winslow, R. (2014). Computing network-based features from physiological time series: Application to sepsis detection. In 2014 36th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2014 (pp. 3825-3826). [6944457] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/EMBC.2014.6944457

Computing network-based features from physiological time series : Application to sepsis detection. / Santaniello, Sabato; Granite, Stephen J.; Sarma, Sridevi V.; Winslow, Raimond.

2014 36th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2014. Institute of Electrical and Electronics Engineers Inc., 2014. p. 3825-3826 6944457.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Santaniello, S, Granite, SJ, Sarma, SV & Winslow, R 2014, Computing network-based features from physiological time series: Application to sepsis detection. in 2014 36th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2014., 6944457, Institute of Electrical and Electronics Engineers Inc., pp. 3825-3826, 2014 36th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2014, Chicago, United States, 8/26/14. https://doi.org/10.1109/EMBC.2014.6944457
Santaniello S, Granite SJ, Sarma SV, Winslow R. Computing network-based features from physiological time series: Application to sepsis detection. In 2014 36th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2014. Institute of Electrical and Electronics Engineers Inc. 2014. p. 3825-3826. 6944457 https://doi.org/10.1109/EMBC.2014.6944457
Santaniello, Sabato ; Granite, Stephen J. ; Sarma, Sridevi V. ; Winslow, Raimond. / Computing network-based features from physiological time series : Application to sepsis detection. 2014 36th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2014. Institute of Electrical and Electronics Engineers Inc., 2014. pp. 3825-3826
@inproceedings{1c84d5fc0b53473bb77d3d386964d286,
title = "Computing network-based features from physiological time series: Application to sepsis detection",
abstract = "Sepsis is a systemic deleterious host response to infection. It is a major healthcare problem that affects millions of patients every year in the intensive care units (ICUs) worldwide. Despite the fact that ICU patients are heavily instrumented with physiological sensors, early sepsis detection remains challenging, perhaps because clinicians identify sepsis by using static scores derived from bed-side measurements individually, i.e., without systematically accounting for potential interactions between these signals and their dynamics. In this study, we apply network-based data analysis to take into account interactions between bed-side physiological time series (PTS) data collected in ICU patients, and we investigate features to distinguish between sepsis and non-sepsis conditions. We treated each PTS source as a node on a graph and we retrieved the graph connectivity matrix over time by tracking the correlation between each pair of sources' signals over consecutive time windows. Then, for each connectivity matrix, we computed the eigenvalue decomposition. We found that, even though raw PTS measurements may have indistinguishable distributions in non-sepsis and early sepsis states, the median /I of the eigenvalues computed from the same data is statistically different (p <0.001) in the two states and the evolution of /I may reflect the disease progression. Although preliminary, these findings suggest that network-based features computed from continuous PTS data may be useful for early sepsis detection.",
author = "Sabato Santaniello and Granite, {Stephen J.} and Sarma, {Sridevi V.} and Raimond Winslow",
year = "2014",
month = "11",
day = "2",
doi = "10.1109/EMBC.2014.6944457",
language = "English (US)",
isbn = "9781424479290",
pages = "3825--3826",
booktitle = "2014 36th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2014",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Computing network-based features from physiological time series

T2 - Application to sepsis detection

AU - Santaniello, Sabato

AU - Granite, Stephen J.

AU - Sarma, Sridevi V.

AU - Winslow, Raimond

PY - 2014/11/2

Y1 - 2014/11/2

N2 - Sepsis is a systemic deleterious host response to infection. It is a major healthcare problem that affects millions of patients every year in the intensive care units (ICUs) worldwide. Despite the fact that ICU patients are heavily instrumented with physiological sensors, early sepsis detection remains challenging, perhaps because clinicians identify sepsis by using static scores derived from bed-side measurements individually, i.e., without systematically accounting for potential interactions between these signals and their dynamics. In this study, we apply network-based data analysis to take into account interactions between bed-side physiological time series (PTS) data collected in ICU patients, and we investigate features to distinguish between sepsis and non-sepsis conditions. We treated each PTS source as a node on a graph and we retrieved the graph connectivity matrix over time by tracking the correlation between each pair of sources' signals over consecutive time windows. Then, for each connectivity matrix, we computed the eigenvalue decomposition. We found that, even though raw PTS measurements may have indistinguishable distributions in non-sepsis and early sepsis states, the median /I of the eigenvalues computed from the same data is statistically different (p <0.001) in the two states and the evolution of /I may reflect the disease progression. Although preliminary, these findings suggest that network-based features computed from continuous PTS data may be useful for early sepsis detection.

AB - Sepsis is a systemic deleterious host response to infection. It is a major healthcare problem that affects millions of patients every year in the intensive care units (ICUs) worldwide. Despite the fact that ICU patients are heavily instrumented with physiological sensors, early sepsis detection remains challenging, perhaps because clinicians identify sepsis by using static scores derived from bed-side measurements individually, i.e., without systematically accounting for potential interactions between these signals and their dynamics. In this study, we apply network-based data analysis to take into account interactions between bed-side physiological time series (PTS) data collected in ICU patients, and we investigate features to distinguish between sepsis and non-sepsis conditions. We treated each PTS source as a node on a graph and we retrieved the graph connectivity matrix over time by tracking the correlation between each pair of sources' signals over consecutive time windows. Then, for each connectivity matrix, we computed the eigenvalue decomposition. We found that, even though raw PTS measurements may have indistinguishable distributions in non-sepsis and early sepsis states, the median /I of the eigenvalues computed from the same data is statistically different (p <0.001) in the two states and the evolution of /I may reflect the disease progression. Although preliminary, these findings suggest that network-based features computed from continuous PTS data may be useful for early sepsis detection.

UR - http://www.scopus.com/inward/record.url?scp=84929463690&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84929463690&partnerID=8YFLogxK

U2 - 10.1109/EMBC.2014.6944457

DO - 10.1109/EMBC.2014.6944457

M3 - Conference contribution

SN - 9781424479290

SP - 3825

EP - 3826

BT - 2014 36th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2014

PB - Institute of Electrical and Electronics Engineers Inc.

ER -