A pattern-mixture model with non-future dependence and shift in current missing values

4^th International Conference and Exhibition on Biometrics & Biostatistics

November 16-18, 2015 San Antonio, USA

Dayong Li

Actavis Plc., USA

Posters-Accepted Abstracts: J Appl Computat Math

Abstract :

In clinical trials, missing data is often present due to various reasons. A pattern-mixture model was studied for inference of treatment effect at the final time point. Missing values were assumed to be monotonic and only depended on observed data and the current missing value. The conditional distribution of the current missing value differs from that of the observed value by location and scale shifts. The shift parameters measure the departure from the missing at random mechanism. Multiple imputations were used and the usual multiple imputation variance estimator is shown to be valid for the overall mean estimate at the final time point when only location shifts are considered. Application to a real clinical study will be discussed.

Biography :

Email: dayongonline@yahoo.com