|
Activity Number:
|
375
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 2, 2016 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
|
Abstract #319322
|
|
|
Title:
|
Correlating a Right-Censored Risk Factor with a Neuropathology Finding Subject to Missing Values
|
|
Author(s):
|
Richard Kryscio* and Erin L. Abner and Peter T. Nelson and David Fardo and Frederick A. Schmitt
|
|
Companies:
|
University of Kentucky and University of Kentucky and University of Kentucky and University of Kentucky and University of Kentucky
|
|
Keywords:
|
right censored risk ;
missing outcomes ;
exponential, ;
mild cognitive impairment
|
|
Abstract:
|
This presentation will investigate methods for correlating a risk factor subject to right censoring with a neuropathology finding subject to missing values. These comparisons will investigate maximum likelihood estimation and multiple imputation in a contingency model when the right censored risk factor is stratified into relevant categories and will investigate a Cox model for the risk factor in the presence of a missing neuropathology covariate. The special case when the risk factor is exponentially distributed will be discussed. Data on 624 mild cognitive impairment (MCI) cases derived from the Statistical Modeling of Aging and Risk of Transition (SMART) project who died without further impairment and came to autopsy will be used to illustrate the results. The SMART project combines autopsy data from five centers conducting longitudinal studies on dementia. Application will focus on correlating length of time in MCI with the presence of arteriolosclerosis of the brain. The latter is missing in some cases and the former is right censored, since some study participants are impaired at baseline.
|
Authors who are presenting talks have a * after their name.
