Online Program Home
  My Program

Abstract Details

Activity Number: 484 - Recent Statistical Advances in Biomarker Studies in the Era of Precision Medicine
Type: Invited
Date/Time: Wednesday, August 2, 2017 : 10:30 AM to 12:20 PM
Sponsor: ENAR
Abstract #322242
Title: Comparison of Approaches for Incorporating New Information into Existing Risk Prediction Models
Author(s): Ruth Pfeiffer * and Sonja Grill and Donna Ankerst and Mitchell Gail and Nilanjan Chatterjee
Companies: National Cancer Institute, NIH, HHS and Technical University of Munich and Technical University of Munich and National Cancer Institute, NIH, HHS and Johns Hopkins University
Keywords: Calibration ; Discrimination ; Model updating ; Independence Bayes ; Risk prediction
Abstract:

We compare the calibration and variability of risk prediction models that were estimated using various approaches for combining information on new predictors (''markers'') with parameter information available for other variables from an earlier model, that was estimated from a large data source. We assess the performance of risk prediction models updated based on likelihood ratio (LR) approaches that incorporate dependence between new and old risk factors and approaches that assume independence (''naive Bayes'' methods). We study the impact of estimating the LR by a) fitting a single model to cases and non-cases when the distribution of the new markers is in the exponential family or b) fitting separate models to cases and controls. We also evaluate a new constrained maximum likelihood method. We study updating the risk prediction model when the new data arise from a cohort and extend methods to accommodate updating using case-control data. Based on results from simulations using real and synthetic data, we recommend the LR method fit using a single model or constrained maximum likelihood.


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

Back to the full JSM 2017 program

 
 
Copyright © American Statistical Association