|
Activity Number:
|
361
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #303565 |
|
Title:
|
Biomarker Evaluation Using Percentile Value Standardization Approach with Event Time Outcomes
|
|
Author(s):
|
Yuying Jin*+ and Margaret Pepe and Yingye Zheng
|
|
Companies:
|
University of Washington and University of Washington and Fred Hutchinson Cancer Research Center
|
|
Address:
|
, Seattle, WA, ,
|
|
Keywords:
|
ROC curve ; percentile value ; time-dependent outcome ; censoring ; inverse probability weighting ; Kaplan-Meier
|
|
Abstract:
|
The diagnostic accuracy of a biomarker is usually assessed with the ROC curve. Percentile value standardization provides an alternative framework to ROC analysis. We develop methods to integrate both the time-varying nature of the marker and the clinical onset time of the disease based on this framework for biomarker evaluation and comparison. The inverse probability weighting (IPW) and Kaplan-Meier approach are proposed to improve the efficiency estimating PVs. IPW approach corrects for the bias caused by the censoring when analyzing the standardized markers in the affected population. The new framework is more intuitive and familiar to researchers and offers a natural approach for evaluating the accuracy of makers for time-dependent outcome. The new estimates are evaluated by simulated studies and applied to Seattle Heart Failure data. Conclusion and discussion are offered afterwards.
|
