Activity Number:
|
300
- SPEED: Statistical Epidemiology
|
Type:
|
Contributed
|
Date/Time:
|
Tuesday, July 31, 2018 : 8:30 AM to 10:20 PM
|
Sponsor:
|
Section on Statistics in Epidemiology
|
Abstract #329916
|
Presentation
|
Title:
|
Expected Versus Observed Effects on Conditional Probability for Clinical Trial Futility Assessment
|
Author(s):
|
Zhibao Mi* and Kelsey A.L. Alexovitz and Xiaoli Lu and Kousick Biswas and Joseph F Collins
|
Companies:
|
VA CSPCC Perry Point and VA Cooperative Studies Program Coordinating Center and VA Cooperative Studies Program Coordinating Center and VA Cooperative Studies Program Coordinating Center and VA Cooperative Studies Program Coordinating Center
|
Keywords:
|
Trial futility;
Observed effect;
Expected effect;
Conditional probability
|
Abstract:
|
Interim analysis for futility is common in clinical trial practice. A trial in which a treatment is unlikely to be effective can be stopped early at the interim time point to save limited resources and prevent unnecessary exposure to adverse effects. Futility is often assessed by calculating a conditional probability of rejecting the null hypothesis if the study is completed. The probability calculation includes two parts, observed data before the interim analysis point and unobserved data in the future whose distribution can be assumed as the same as the distribution of the observed data or as the distributions specified in the study hypotheses. Consequently, futility decisions could vary because of the different probabilities conditioned on observed effects or expected effects of unobserved data. To assess the effects of different distributions of expected and observed data on conditional probabilities for determining futility, we retrospectively examine ten multicenter clinical trials and simulated data under different assumptions, hoping that the analytical results would help researchers to better assess a trial futility.
|
Authors who are presenting talks have a * after their name.