|
Activity Number:
|
265
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 10, 2004 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #300500 |
|
Title:
|
Evaluating and Correcting Guess Effect in not Perfect Double-blinded Clinical Trials
|
|
Author(s):
|
Jianfeng Cheng*+ and Eva Petkova
|
|
Companies:
|
Columbia University and Columbia University
|
|
Address:
|
1051 Riverside Dr., New York, NY, 10032,
|
|
Keywords:
|
double-blinded clinical trial ; assessing guess effect ; correcting guess effect
|
|
Abstract:
|
Failure to maintain blindness is a common phenomenon in double-blinded clinical trials testing the efficacy of antidepressant medications, which may affect the magnitude of the drug effects and makes the efficiency of psychotropic drugs to unknown degrees untrustworthy. It is most often thought that the bias of the researchers' rating will benefit the active treatment in a placebo controlled experiment or the competing agent in trials where the control is some traditional treatment. We propose a method for assessing the presence of possible rater's bias due to treatment guessing in antidepressant clinical trials and offer a proposed estimator of treatment efficacy by correcting such guess effect. Simulation studies showed the properties of the proposed estimators: unbiasedness and smaller mean square error than the naive ones. Through the examples, we presented how to apply the methods to the clinical trials, even for those without guess data collected. The proposed approach allows for reevaluation of previously conducted clinical trials and provides a method for measuring and correcting the potential bias.
|
