|
Activity Number:
|
480
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #308860 |
|
Title:
|
Regression Analysis of a Disease Onset Distribution Using Diagnosis Data
|
|
Author(s):
|
Jessica G. Young*+ and Nicholas P. Jewell and Steven J. Samuels
|
|
Companies:
|
University of California, Berkeley and University of California, Berkeley and University at Albany-SUNY
|
|
Address:
|
711 Everett Street, El Cerrito, CA, 94530,
|
|
Keywords:
|
Current status data ; Proportional hazards ; Uterine fibroids
|
|
Abstract:
|
We consider methods for estimating the effect of a covariate on disease onset when the data consist of right-censored data on diagnosis times and current status data on onset times for individuals not yet diagnosed. Dunson and Baird (2001) approached this problem using maximum likelihood, with the assumption that the ratio of the diagnosis and onset distributions is monotonic non-decreasing. We propose an alternative estimator that is computationally simpler and requires no assumptions on this ratio. A simulation study is performed comparing estimates from these two approaches, as well as that from a current status analysis ignoring diagnosis data. Results show that the Dunson and Baird estimator outperforms ours when the monotonicity assumption holds, but the reverse is true when this assumption fails. A data example is provided where the monotonicity assumption is seen to fail.
|
