Sepsis Prediction: An Attention-Based Interpretable Approach

Baghaei, Kourosh T.; Rahimi, Shahram

doi:10.1109/fuzz-ieee.2019.8858808

Cited by 4 publications

(8 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Following our previous research (Baghaei and Rahimi, 2019), in this study, we implemented a bidirectional GRU recurrent network with an added attention layer in order to capture the contributions of each of the medical parameters (including vital signs and lab measurements) to the final predicted outcome. We also illustrated and discussed this model's usability with qualitative analysis on a few case studies.…”

Section: Discussionmentioning

confidence: 99%

“…Following our previous work (Baghaei and Rahimi, 2019), since the format of the data of our study is time series, we have decided to utilize Recurrent Neural Network (RNN) in order to make predictions. More specifically, we chose Gated Recurrent Units (GRU) in our architecture (Bahdanau et al, 2014).…”

Section: Methodsmentioning

confidence: 99%

“…The model in this research project is similar to that of our previous work (Baghaei and Rahimi, 2019). We built our model upon the proposed model in (Choi et al, 2016).…”

Section: Architecture Of the Modelmentioning

confidence: 99%

“…In this study, as a sequel to our previous research (Baghaei and Rahimi, 2019), we have attempted to address the aforementioned issue by implementing a deep attention-based recurrent neural network so that a medical doctor can understand the factors that have contributed most to the diagnosis of sepsis in a given patient. The contributions of this research can be summarized as follows:…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

An Interpretable Approach to Prediction of Sepsis Using Multivariate Time Series

Baghaei

Rahimi

Bozorgzad

2022

Preprint

Self Cite

View full text Add to dashboard Cite

Sepsis is the major cause of death in ICUs and amounts to a large portion of medical expenses. Earlier prediction of sepsis can assist in reducing inpatient mortality and healthcare costs. Despite predictive capabilities of the state-of-the-art approaches, deep models still lack the transparency that provides some sort of reasoning for their predictions. In this work, we studied the predictive performance of an attention-based recurrent neural network and its ability to provide interpretability for the relative contribution of each medical factor to the final prediction. We also qualitatively assess the explainability of the proposed model. This approach sheds light on the opacity issue of deep predictive models. Moreover, it enables the clinical staff to make more reliable decisions based on the status of the patient in ICU thus decreasing the sepsis inpatient mortality.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

“…The model in this research project is similar to that of our previous work (Baghaei and Rahimi, 2019). We built our model upon the proposed model in (Choi et al, 2016).…”

Section: Architecture Of the Modelmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

An Interpretable Approach to Prediction of Sepsis Using Multivariate Time Series

Baghaei

Rahimi

Bozorgzad

2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Longitudinal studies have also appeared [61][62][63] , together with methods integrating alternative data sources such as omics 64 and others. In the end of the 2010s, the computational intelligence revolution entered the playground too, and deep learning approaches flooded the specialized journals [65][66][67][68][69][70][71][72][73] , also considering the interpretability issue 74,75 . Fleuren et al 76 published comprehensive review of the different aspects.…”

mentioning

confidence: 99%

Survival prediction of patients with sepsis from age, sex, and septic episode number alone

Chicco

Jurman

2020

Sci Rep

View full text Add to dashboard Cite

Sepsis is a life-threatening condition caused by an exaggerated reaction of the body to an infection, that leads to organ failure or even death. Since sepsis can kill a patient even in just one hour, survival prediction is an urgent priority among the medical community: even if laboratory tests and hospital analyses can provide insightful information about the patient, in fact, they might not come in time to allow medical doctors to recognize an immediate death risk and treat it properly. In this context, machine learning can be useful to predict survival of patients within minutes, especially when applied to few medical features easily retrievable. In this study, we show that it is possible to achieve this goal by applying computational intelligence algorithms to three features of patients with sepsis, recorded at hospital admission: sex, age, and septic episode number. We applied several data mining methods to a cohort of 110,204 admissions of patients, and obtained high prediction scores both on this complete dataset (top precision-recall area under the curve PR AUC = 0.966) and on its subset related to the recent Sepsis-3 definition (top PR AUC = 0.860). Additionally, we tested our models on an external validation cohort of 137 patients, and achieved good results in this case too (top PR AUC = 0.863), confirming the generalizability of our approach. Our results can have a huge impact on clinical settings, allowing physicians to forecast the survival of patients by sex, age, and septic episode number alone.

show abstract

Adversarial Activity Detection Using Keystroke Acoustics

Fallahi

Phoha

2021

Computer Security – ESORICS 2021

View full text Add to dashboard Cite

Sepsis Prediction: An Attention-Based Interpretable Approach

Cited by 4 publications

References 18 publications

An Interpretable Approach to Prediction of Sepsis Using Multivariate Time Series

An Interpretable Approach to Prediction of Sepsis Using Multivariate Time Series

Survival prediction of patients with sepsis from age, sex, and septic episode number alone

Adversarial Activity Detection Using Keystroke Acoustics

Contact Info

Product

Resources

About