Statistica Sinica 24 (2014), 313-333

SEMIPARAMETRIC ACCELERATED FAILURE TIME

MODEL FOR LENGTH-BIASED DATA

WITH APPLICATION TO DEMENTIA STUDY

Jing Ning$^{1}$, Jing Qin$^{2}$ and Yu Shen$^{1}$

$^{1}$The University of Texas M. D. Anderson Cancer Center

$^{2}$National Institute of Allergy and Infectious Diseases

Abstract: A semiparametric accelerated failure time (AFT) model is proposed to evaluate the effects of risk factors on the unbiased failure times for the target population given the observed length-biased data. The analysis of length-biased data is complicated by informative right censoring due to the biased sampling mechanism, and consequently the techniques for conventional survival analysis are not applicable. We propose estimating equation methods for estimation and show the asymptotic properties of the proposed estimators. The small sample performance of the estimating methods are investigated and compared with that of existing methods under various underlying distributions and censoring mechanisms. We apply the proposed model and estimating methods to a prevalent cohort study, the Canadian Study of Health and Aging (CSHA), to evaluate the survival duration according to diagnosis of subtype of dementia.

Key words and phrases: Accelerated failure time model, dementia, dependent censoring, estimating equation, length-biased sampling, prevalent cohort.