|
Activity Number:
|
500
- Beyond Randomized Studies: Nonrandomized, Single Arm, or Special Studies
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 2, 2017 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Medical Devices and Diagnostics
|
|
Abstract #323001
|
View Presentation
|
|
Title:
|
Confidence Interval Estimate of Success Rate on Correlated Data in a Single Arm Study
|
|
Author(s):
|
Kate Wang* and Richard Weng and Yijun Sun
|
|
Companies:
|
Allergan and Allergan and Allergan
|
|
Keywords:
|
Confidence Interval Estimate ;
Success Rate on Correlated Data ;
Single Arm ;
Coverage Probability ;
Bootstrap ;
GEE
|
|
Abstract:
|
The purpose of an observational study was to estimate the success rate of a treatment on a random sample of individuals. Each individual was treated twice, the success of the treatment was assessed after each of the treatment. Success rate was estimated by taking total number of success over the total number of measurement. The confidence interval of the success rate estimate need to take into the consideration of the correlation between the two binary measurement within the individual. Three methods to calculate the confidence interval were evaluated using the coverage probability calculated with Monte Carlo simulation. They all improved accuracy of the coverage than the naïve method that ignore the correlation. The bootstrap and GEE methods have similar coverage probability that are much closer to the confidence level. A simple bivariate normal approximation method has less accurate probability coverage.
|
Authors who are presenting talks have a * after their name.
