Online Program Home
  My Program

Abstract Details

Activity Number: 503 - SPAAC Poster Competition
Type: Topic Contributed
Date/Time: Wednesday, August 2, 2017 : 10:30 AM to 12:20 PM
Sponsor: Scientific and Public Affairs Advisory Committee
Abstract #324118
Title: Identification of Sepsis Phenotypes Using Clustering Methods
Author(s): Zhongying Xu* and Hernando Gomez and Joyce Chang
Companies: and University of Pittsburgh and University of Pittsburgh
Keywords: Sample Clustering ; Sepsis Phenotypes

Sepsis is a high risk and life-threatening syndrome caused by the body's overwhelming inflammatory response to infection. Traditionally, sepsis has been considered as one syndrome with clinical presentations varying only by severity. However, recent data has challenged this paradigm, and has suggested that sepsis probably encompasses multiple phenotypes. In this study, we focused on exploring diverse patterns of sepsis in a cohort of critically ill patients, in order to better understand different responses and further design targeted treatment on specific phenotypes. The dataset we used includes clinical, laboratory, and demographic information on adult patients admitted to a tertiary care institution with 8 intensive care units (ICU) in 8 years. Several clustering methods were applied including k-means based consensus clustering, tight clustering, and hierarchical clustering. Comparing different clustering methods, we found the number of clusters and the patterns describing these clusters are very similar. In addition, we also found that the resulting clusters were associated with subsequent clinical characteristics showing the clinical value of the identified clusters. 

Authors who are presenting talks have a * after their name.

Back to the full JSM 2017 program

Copyright © American Statistical Association