|
Activity Number:
|
230
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistical Consulting
|
| Abstract - #304936 |
|
Title:
|
Predicting Mortality in Infants
|
|
Author(s):
|
Mark Seiss*+
|
|
Companies:
|
Virginia Polytechnic Institute and State University
|
|
Address:
|
403 Hutcheson Hall, Blacksburg, VA, 24061,
|
|
Keywords:
|
Infant Mortality ; Logistic Regression Mixed Model ; Hierarchical Probit Model ; Jackknife
|
|
Abstract:
|
57 babies with congenital diaphragmatic hernia (CDH) before birth had a measurement of their lung size taken called lung to head ratio (LHR). Ideally, this measurement is taken between 20 and 29 weeks of gestation. Doctors would like to use the LHR measurement to predict the probability of death or the use of a heart lung bypass procedure called ECMO that usually leads to death. Ultimately, doctors would use the predicted probability to determine if experimental fetal intervention is necessary as well as location of delivery and allocation of hospital resources. This paper will describe three approaches that were used to predict this probability of death or ECMO using the clinical outcomes of the 57 babies, some with multiple measurements: a logistic regression mixed model, a two step model that first predicts lung size at full term (39 weeks) and then the probability of death or...
|
