Conference presentation
Analysis of Temporal Patterns of Sepsis-related Cellular and Physiological Responses to Inform Personalized Prediction of Mortality
INFORMS 2019 (Seattle, Washington, United States, 20 Oct 2019 - 23 Oct 2019)
21 Oct 2019
Abstract
Sepsis is a multi-stage health condition with high mortality rate. While sepsis-related cellular and physiological responses have been utilized as independent factors to predict outcomes, temporal dependencies between these factors are understudied. Using electronic health records data from 13,367 unique patients, we model sepsis progression developing novel co-failure networks and association rule algorithms. Key results show that proximity and frequency of cellular and physiological co-failures are sepsis stage-specific and there are significant proximate failures that play a major role in predicting in-hospital mortality in a personalized manner.
Metrics
2 Record Views
Details
- Title
- Analysis of Temporal Patterns of Sepsis-related Cellular and Physiological Responses to Inform Personalized Prediction of Mortality
- Creators
- Muge Capan - Drexel UniversityAli Jazayeri - Drexel UniversityC C Yang - Drexel UniversityMin Chi - North Carolina State UniversityJulie Simmons Ivy - North Carolina State UniversityJeanne Marie Huddleston - Mayo ClinicRyan Arnold - Drexel University
- Conference
- INFORMS 2019 (Seattle, Washington, United States, 20 Oct 2019 - 23 Oct 2019)
- Resource Type
- Conference presentation
- Language
- English
- Academic Unit
- Decision Sciences (and Management Information Systems); Emergency Medicine; Information Science (Informatics)
- Identifiers
- 991021856364004721