|
Activity Number:
|
424
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #306812 |
|
Title:
|
Estimating Treatment Effect in Clinical Trials with Disease-Dependant Noncompliance
|
|
Author(s):
|
Kuenhi Tsai*+ and Peter Westfall and Stephan Ogenstad and Miles Dunn
|
|
Companies:
|
Vertex Pharmaceuticals Incorporated and Texas Tech University and Vertex Pharmaceuticals Incorporated and Vertex Pharmaceuticals Incorporated
|
|
Address:
|
130 Waverly Street, Cambridge, MA, 02139,
|
|
Keywords:
|
noncompliance ; clinical trials ; simulations ; covariate adjustment
|
|
Abstract:
|
Estimating treatment effect in clinical trials is often complicated by noncompliance with the nominal prescribed dosage. Incorporated noncompliance as a covariate becomes a trend in estimating treatment effect. However, the impact of noncompliance on treatment effect is dependent upon the therapeutic disease. We illustrate its impact in 3 therapeutic areas of HIV, rheumatoid arthritis, and Alzheimer. Due to the difference, the covariate adjustment of noncompliance in estimating the treatment effect has to be appropriately selected according to the natural history of the disease. Based on published PD models and available references, we simulate the data under various schemes of noncompliance on these 3 therapies. The simulated data are analyzed with and without noncompliance as a covariate in estimating treatment effect. Our results show the significant impact of noncompliance.
|
