|
Activity Number:
|
419
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #305657 |
|
Title:
|
Joint Modeling of Missing Data Due to Nonresponse and Death in Longitudinal Epidemiological Studies
|
|
Author(s):
|
Kumar B. Rajan*+ and Sue Leurgans
|
|
Companies:
|
Rush University and Medical Center and Rush University and Medical Center
|
|
Address:
|
1645 W Jackson Blvd, Suite 675, Chicago , IL, 60612,
|
|
Keywords:
|
Missingness due to death ; , Joint missing mechanism ; Weighting procedure ; Selection bias due to death
|
|
Abstract:
|
In longitudinal studies of aging populations, missing data could result from non-response, or from death among participants. Diehr et al., 2005, treated death as a possible outcome to account for selection bias. Conventional missing data methods account for selection bias due to non-response but not death. We present a weighting procedure to model the missingness due to non-response and death jointly. Estimates are consistent and asymptotically normal under plausible conditions. Simulation studies will be presented to show the appropriateness of this weighting procedure. Additionally, the joint weighting procedure will be applied to study risk factors for cognitive decline among the aging population.
|
