|
Activity Number:
|
92
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, July 30, 2007 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Health Policy Statistics
|
| Abstract - #308277 |
|
Title:
|
Evaluating Bias Correction in Weighted Proportional Hazards Regression
|
|
Author(s):
|
Qing Pan*+ and Douglas Schaubel
|
|
Companies:
|
University of Michigan and University of Michigan
|
|
Address:
|
2355 Arrowwood Trl, Ann Arbor, MI, 48105,
|
|
Keywords:
|
Confidence bands ; Inverse-selection-probability weights ; Observational studies ; Proportional hazards model ; Selection Bias ; Wald test
|
|
Abstract:
|
With possible biased sample, Pan and Schaubel recently proposed inference procedures for an inverse selection probability weighted Cox model, applicable when selection probabilities are estimated empirically. The proposed weighted estimator is consistent for the target population parameter, while the unweighted estimator converges to a modification of the true value. Similar statements apply to the cumulative hazard estimators. Although parameter estimation is consistent, computation is more intensive. Here, we propose methods for evaluating bias in the unweighted partial likelihood and Breslow-Aalen estimators. Asymptotic properties of the proposed statistics are derived. The finite-sample significance level and power are evaluated through simulation. The proposed methods are applied to national organ failure registry data to evaluate the bias in a post kidney transplant survival model.
|
