Online Program Home
My Program

Abstract Details

Activity Number: 78 - Bayesian Generalized Linear Models for Medicine
Type: Contributed
Date/Time: Sunday, July 29, 2018 : 4:00 PM to 5:50 PM
Sponsor: Section on Bayesian Statistical Science
Abstract #330639 Presentation
Title: Characterization of Critically Ill Patients Using the Profile Counts of Laboratory Tests and Medications
Author(s): Eduardo PhD Antonio Trujillo Rivera* and Qing PhD Zeng and James Bost and Anita MD Patel and Hiroki PhD Morizono and Dongkyu PhD Kim and James MD Chamberlain and Murray MBA MD Pollack
Companies: George Washington University, Children's National Health System and George Washington University, Department of Veterans Affairs and Children's National Medical Center and Children's National Health System, George Washington University and Center for Genetic Medicine Research, Children's Research Institute and Children's Research Institute, Children's National Health System and Children's National Health System, George Washington University and Children's National Health System, George Washington University
Keywords: Health Record; Dirichlet-Multinomial; Bayes; Random Effects
Abstract:

In this work we characterize pediatric critically ill patients using their clinical record. We use the profiles of laboratory test requested and medications dispensed; specifically we use the counts. The sample of encounters used in this study include patients that do not have ICU admissions but were located in routine care units and patients that had ICU admissions with different number of ICU re-admissions. We study the vectors of counts using a Bayesian regression model with random effects and response Dirichlet-Multinomial; we adjust for demographic information and hospital. During the presentation, we describe the process in which the cohort was chosen and the fraction of their record that is considered for this study, we describe the model; finally, we present the results of the characterization.


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

Back to the full JSM 2018 program