An Improved Multi-Output Gaussian Process RNN with Real-Time Validation for Early Sepsis Detection

dc.contributor.author

Futoma, Joseph

dc.contributor.author

Hariharan, Sanjay

dc.contributor.author

Sendak, Mark

dc.contributor.author

Brajer, Nathan

dc.contributor.author

Clement, Meredith

dc.contributor.author

Bedoya, Armando

dc.contributor.author

O'Brien, Cara

dc.contributor.author

Heller, Katherine

dc.date.accessioned

2022-08-26T13:25:48Z

dc.date.available

2022-08-26T13:25:48Z

dc.date.issued

2017-08-19

dc.date.updated

2022-08-26T13:25:48Z

dc.description.abstract

Sepsis is a poorly understood and potentially life-threatening complication that can occur as a result of infection. Early detection and treatment improves patient outcomes, and as such it poses an important challenge in medicine. In this work, we develop a flexible classifier that leverages streaming lab results, vitals, and medications to predict sepsis before it occurs. We model patient clinical time series with multi-output Gaussian processes, maintaining uncertainty about the physiological state of a patient while also imputing missing values. The mean function takes into account the effects of medications administered on the trajectories of the physiological variables. Latent function values from the Gaussian process are then fed into a deep recurrent neural network to classify patient encounters as septic or not, and the overall model is trained end-to-end using back-propagation. We train and validate our model on a large dataset of 18 months of heterogeneous inpatient stays from the Duke University Health System, and develop a new "real-time" validation scheme for simulating the performance of our model as it will actually be used. Our proposed method substantially outperforms clinical baselines, and improves on a previous related model for detecting sepsis. Our model's predictions will be displayed in a real-time analytics dashboard to be used by a sepsis rapid response team to help detect and improve treatment of sepsis.

dc.identifier.uri

https://hdl.handle.net/10161/25613

dc.subject

stat.ML

dc.subject

stat.ML

dc.subject

stat.AP

dc.subject

stat.ME

dc.title

An Improved Multi-Output Gaussian Process RNN with Real-Time Validation for Early Sepsis Detection

dc.type

Journal article

duke.contributor.orcid

Bedoya, Armando|0000-0001-6496-7024

pubs.organisational-group

Duke

pubs.organisational-group

School of Medicine

pubs.organisational-group

Basic Science Departments

pubs.organisational-group

Clinical Science Departments

pubs.organisational-group

Biostatistics & Bioinformatics

pubs.organisational-group

Medicine

pubs.organisational-group

Medicine, Pulmonary, Allergy, and Critical Care Medicine

Files

Original bundle

Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
1708.05894v1.pdf
Size:
696.03 KB
Format:
Adobe Portable Document Format