|
Activity Number:
|
517
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 10, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #307214 |
|
Title:
|
On Sensitivity Analysis of HIV-Free Survival
|
|
Author(s):
|
Peng Zhang*+ and Stephen W. Lagakos
|
|
Companies:
|
Harvard University and Harvard University
|
|
Address:
|
655 Huntington Ave., Boston, MA, 02115,
|
|
Keywords:
|
composite ending points ; HIV vertical transmission ; Monte Carlo ; panel data ; sensitivity analysis
|
|
Abstract:
|
We consider estimation of the distribution of HIV-free survival in infants of HIV-infected mothers. Because HIV infection is a silent event, HIV infection status is assessed periodically using a diagnostic test, which is imperfect. Thus neither the occurrence of infection nor the time of occurrence can be known. The competing cause of failure--death--is observed continuously. The problem is further complicated by missing data, which includes non-informative missingness due to skipping pre-scheduled visits and possibly informative missingness due to death or lost follow-up. Because of the data structure, aspects of the underlying process are non-identifiable. We apply nonparametric methods to estimate the estimable aspects of the process and supplement these with non-parametric sensitivity analyses to make the inference about the distributions of HIV survival and overall survival time.
|
