|
Activity Number:
|
251
- SPEED: Biopharmaceutical Methods and Application I, Part 2
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 29, 2019 : 2:00 PM to 2:45 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
|
Abstract #307605
|
|
|
Title:
|
Reducing Misclassification Effect on Dynamic Treatment Regimen (DTR) of Sequential Multiple Assignment Randomized Trial Designs (SMART)
|
|
Author(s):
|
Jun He* and Roy T Sabo and Donna McClish
|
|
Companies:
|
Virginia Commonwealth University and Virginia Commonwealth University and VCU
|
|
Keywords:
|
SMART;
clinical trial;
DTR;
Misclassification;
Mixture Model;
K neearest neighbor
|
|
Abstract:
|
SMART designs tailor individual treatment by re-randomizing patients to subsequent therapies based on their response to initial treatment. We had shown that misclassification of patient responses to initial treatment can lead to inappropriate treatment assignment which can introduce bias to estimates mean, variance, as well as affect statistical inference for DTR. Normal mixture-model (NM) and k-nearest-neighbor (KNN) strategies were investigated to attempt to reduce bias and improve inference at final stage outcome. The early stage probabilities of being correctly classified for each patient in both methods were calculated, and then the outcome of patient observations were weighted by the function of these probabilities. Simulations were used to compare the performance of these approaches. The results showed that both methods moderately reduced bias, but the tradeoff was increased type I error rate and little effect on power.
|
Authors who are presenting talks have a * after their name.
