|
Activity Number:
|
243
- Contributed Poster Presentations: Biopharmaceutical Section
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 31, 2017 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
|
Abstract #324219
|
|
|
Title:
|
A Modified LOCF-PMM Method with Multiple Imputation for Handling Missing Data in Longitudinal Studies
|
|
Author(s):
|
Busola Sanusi* and Kenneth Liu and Gregory Golm
|
|
Companies:
|
The University of North Carolina at Chapel Hill and Merck & Co. and Merck & Co., Inc.
|
|
Keywords:
|
Depression ;
Paroxetine ;
Placebo ;
Clinical trial
|
|
Abstract:
|
Last Observation Carried Forward (LOCF) was once commonly used to analyze incomplete longitudinal data because of its simplicity. LOCF imputes a subject's missing data with his last previously observed value. A drawback of LOCF is that it yields biased estimates and it underestimates variability. Mixed models and Pattern Mixture Models (PMM) are methods currently used to handle incomplete longitudinal data. We compare LOCF, a mixed model, and 2 PMMs. The first PMM, Jump-to-Control (JC), is a widely used PMM; the second PMM is our proposed LOCF-PMM. LOCF-PMM performs imputation by adding variability to the LOCF value along with using multiple imputation to provide the correct variance. We apply these methods to a depression study and assess their bias and variance in a simulation study.
|
Authors who are presenting talks have a * after their name.
