|
Activity Number:
|
299
- Estimands and Imputations Methods
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 30, 2019 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biopharmaceutical Section
|
|
Abstract #306738
|
|
|
Title:
|
Identifying Treatment Effects Using Trimmed Means When Data Are Missing Not at Random
|
|
Author(s):
|
Alex Ocampo*
|
|
Companies:
|
Harvard University
|
|
Keywords:
|
Missing Data;
Clinical Trials;
Estimand;
Trimmed Means
|
|
Abstract:
|
Patients often discontinue treatment in a clinical trial because their health condition is not improving. This missing data problem biases the estimate of a medication's efficacy because the missing outcomes of patients are the reason for dropout i.e. missing not at random (MNAR). One solution to combat this problem - the trimmed means approach for missing data - sets missing values as the worst observed outcome and then trims away a fraction of the distribution from each treatment arm before calculating differences in efficacy. Herein, sufficient and necessary conditions are formalized for when this approach can identify the population treatment effect in the presence of MNAR data. Numerical studies show the trimmed means approach's ability to effectively estimate treatment efficacy when data are MNAR and missingness is strongly associated with an unfavorable outcome. If reasons for discontinuation in a trial are known analysts can improve estimates with a combination of multiple imputation and trimmed means when the assumptions of each missing mechanism hold. When the assumptions are justifiable, using trimmed means can help identify treatment effects despite MNAR data.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2019 program
|
