|
Activity Number:
|
436
- SPEED: Tests, Trials, Biomarkers, and Other Topics in Biometrics
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2018 : 3:05 PM to 3:50 PM
|
|
Sponsor:
|
Biometrics Section
|
|
Abstract #332659
|
|
|
Title:
|
A Two-Stage Method to Analyze Multivariate Cluster Biomarkers in Prediction on a Single Binary Outcome
|
|
Author(s):
|
Xiaoying Yu* and Wenyaw Chan and Gracie Vargas and Rahul Pal
|
|
Companies:
|
University of Texas Medical Branch at Galveston and University of Texas Health Science Center at Houston and University of Texas Medical Branch at Galveston and University of Texas Medical Branch at Galveston
|
|
Keywords:
|
multivariate cluster data;
binary;
ROC
|
|
Abstract:
|
In an experimental study, the multiple biomarker levels have commonly been repeatedly collected from the same subject, while the primary binary outcome is measured once or fixed. It is often a statistical issue for analyzing this type of data. In the literature, analysts often average the repeated measures for each biomarker before conducting analysis. We proposed a two-stage multivariate mixed effects model. First, a multivariate random coefficient model for biomarkers assuming unstructured variance-covariance G matrix and diagonal R matrix was conducted, then the best linear unbiased predictors from this model were used as predictors in the logistic regression and ROC analysis. Two methods and the incorrect Naïve method assuming the independence among the repeated measures have been applied to a real-word study to compare the areas under curve using different markers. Three methods yielded different estimates and could lead to different conclusion. The proposed method accounts for the heterogeneity for the subjects and correlations among biomarkers while generating the summary measures.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2018 program
|
