|
Activity Number:
|
322
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #304188 |
|
Title:
|
Proportional Odds Model for Dose-Finding Clinical Trial Designs with Ordinal Toxicity Grading
|
|
Author(s):
|
Emily M. Van Meter*+ and Elizabeth Garrett-Mayer and Dipankar Bandyopadhyay
|
|
Companies:
|
Medical University of South Carolina and Medical University of South Carolina and Medical University of South Carolina
|
|
Address:
|
1244 Aruba Circle, Charleston, SC, 29412,
|
|
Keywords:
|
Continual Reassessment Method ; Clinical Trials ; Phase I ; Proportional Odds
|
|
Abstract:
|
Currently many phase I clinical trial designs, including the continual reassessment method (CRM), dichotomize toxicity outcomes based on pre-specified dose-limiting toxicity criteria. This loss of information is particularly inefficient due to the small sample sizes in phase I trials. This study aims to extend the CRM to include ordinal toxicity outcomes using the proportional odds model and to compare the results with the traditional CRM. We present a sensitivity analysis of the new design comparing various prior distributions, sample sizes, and cohort sizes. A simulation study shows that the proposed proportional odds CRM performs as well or better than the standard CRM on all criteria compared, and notably with more accuracy to estimate the MTD. These findings suggest that it is beneficial to incorporate ordinal toxicities into phase I trial designs.
|
