Dataset Information

Feature selection for the accurate prediction of septic and cardiogenic shock ICU mortality in the acute phase.

ABSTRACT: Circulatory shock is a life-threatening disease that accounts for around one-third of all admissions to intensive care units (ICU). It requires immediate treatment, which is why the development of tools for planning therapeutic interventions is required to deal with shock in the critical care environment. In this study, the ShockOmics European project original database is used to extract attributes capable of predicting mortality due to shock in the ICU. Missing data imputation techniques and machine learning models were used, followed by feature selection from different data subsets. Selected features were later used to build Bayesian Networks, revealing causal relationships between features and ICU outcome. The main result is a subset of predictive features that includes well-known indicators such as the SOFA and APACHE II scores, but also less commonly considered ones related to cardiovascular function assessed through echocardiograpy or shock treatment with pressors. Importantly, certain selected features are shown to be most predictive at certain time-steps. This means that, as shock progresses, different attributes could be prioritized. Clinical traits obtained at 24h. from ICU admission are shown to accurately predict cardiogenic and septic shock mortality, suggesting that relevant life-saving decisions could be made shortly after ICU admission.

SUBMITTER: Aushev A

PROVIDER: S-EPMC6245679 | biostudies-literature | 2018

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Feature selection for the accurate prediction of septic and cardiogenic shock ICU mortality in the acute phase.

Aushev Alexander A Ripoll Vicent Ribas VR Vellido Alfredo A Aletti Federico F Pinto Bernardo Bollen BB Herpain Antoine A Post Emiel Hendrik EH Medina Eduardo Romay ER Ferrer Ricard R Baselli Giuseppe G Bendjelid Karim K

PloS one 20181120 11

Circulatory shock is a life-threatening disease that accounts for around one-third of all admissions to intensive care units (ICU). It requires immediate treatment, which is why the development of tools for planning therapeutic interventions is required to deal with shock in the critical care environment. In this study, the ShockOmics European project original database is used to extract attributes capable of predicting mortality due to shock in the ICU. Missing data imputation techniques and ma ...[more]

PMID: 30457997

Similar Datasets

Project description:Background:Based on promising results over the past 10 years, the method of extracorporeal membrane oxygenation (ECMO) has developed from being used as a 'rescue therapy' to become an accepted treatment option for patients with acute lung failure (ARDS). Subsequently, the indication was extended also to patients suffering from cardiogenic and septic shock. Our aim was to evaluate hospital mortality and associated prognostic variables in patients with lung failure, cardiogenic, and septic shock undergoing ECMO. Furthermore, a cumulative sum (CUSUM) analysis was used to assess the learning curve of ECMO-treatment in our department. Methods:We retrospectively analysed the data of 131 patients undergoing ECMO treatment in the intensive care unit of the Asklepios Hospital of Langen over the time period from April 2011 to July 2016. We categorised the patients into three groups: lung failure (n?=?54); cardiogenic shock (n?=?58); and septic shock (n?=?19). The primary outcome variable was hospital mortality along with identification of prognostic variables on mortality before initiating ECMO using logistic regression. Second outcome variable was the learning curve of our department in patients with ECMO. Results:6-year hospital mortality was 54% in patients with lung failure, 59% in patients with cardiogenic shock, and 58% in patients with septic shock.The CUSUM analysis revealed a typical learning curve with a point of inflection in the year 2014. Patients treated before 2014 had a worse outcome (p?=?0.04 whole cohort; p?=?0.03 for lung failure). Furthermore, less than 20 treatments per year respectively treatment before 2014 were associated negatively with hospital mortality of lung failure patients showing an odds ratio of 4.04, as well as in the entire cohort with an odds ratio of 3.19. Conclusion:For the first time, a steep ECMO-learning curve using the CUSUM tool has been described. Obviously, the experience with ECMO has to be taken into account when defining the role of ECMO in ARDS, cardiogenic, and septic shock.

Dataset Information

Feature selection for the accurate prediction of septic and cardiogenic shock ICU mortality in the acute phase.

Publications

Feature selection for the accurate prediction of septic and cardiogenic shock ICU mortality in the acute phase.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure