|
Activity Number:
|
153
- Contributed Poster Presentations: Biopharmaceutical Section
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 8, 2022 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
|
Abstract #323258
|
|
|
Title:
|
Statistical Methods for Selective Biomarker Testing
|
|
Author(s):
|
Natalie DelRocco* and Adam Ding and Samuel Wu
|
|
Companies:
|
University of Florida and Northeastern University and University of Florida
|
|
Keywords:
|
Outcome Dependent Sampling;
Linear Regression;
Reverse Regression;
Extreme Sampling;
Clinical Trials;
Biomarkers
|
|
Abstract:
|
Studies investigating the association between clinical outcomes and relevant biomarkers are integral in many current statistical applications. When a large number of clinical outcomes are available, there are benefits to considering which units of observation are sampled for biomarker analysis, i.e. adopting a special case of outcome dependent sampling (ODS) where only those units with the most extreme clinical outcomes are biomarker typed. In this work, we employ a joint Gaussian assumption to derive point and interval estimates for the association between a continuous outcome and a biomarker under the above ODS sampling scheme. We show that this method is unbiased and more efficient than random sampling when assumptions are met. It is easy to implement using standard statistical software. The real-world performance is demonstrated in a chronic pain clinical trial.
|
Authors who are presenting talks have a * after their name.
