|
Activity Number:
|
209
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 30, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #310292 |
|
Title:
|
Assessing the Potential Impact of Missing Data in a Longitudinal Study with a Continuous Outcome
|
|
Author(s):
|
Owen Devine*+ and Jorge Rosenthall and Nyasha Skerrette
|
|
Companies:
|
Centers for Disease Control and Prevention and Centers for Disease Control and Prevention and Emory University
|
|
Address:
|
1600 Clifton Road, Atlanta, GA, 30333,
|
|
Keywords:
|
Missing Data ; Uncertainty ; Sensitivity
|
|
Abstract:
|
Based on Daniels and Hogan (Biometrics, 2000), we employ a simple method to assess the sensitivity of model-based parameter estimates developed using study completers to assumptions on possible outcomes for dropouts. A key issue in this type of assessment is generating reasonable potential values for missing data given available information. We develop uncertainty distributions, based on expert opinion and/or analysis of available data, for parameters underlying the missing data. Monte Carlo simulation is used to generate realizations for both the missing data and the associated model parameter estimates that reflect both uncertainty in the underlying parameters and sampling variation inherent in model estimation. We illustrate the approach using a longitudinal study comparing the effectiveness of two folic acid dosing schemes among Honduran women.
|
