A Gap Time Approach for the Analysis of Interval-censored Recurrent Event Data

• Activity Number: 413
• Type: Contributed
• Date/Time: Thursday, August 12, 2004 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Dan Sheng*+ and Mimi Kim
• Companies: New York University and Albert Einstein College of Medicine
• Keywords: interval censoring ; recurrent events ; gap time ; robust inference ; multivariate survival analysis

Abstract

Prospective biomedical studies frequently involve the monitoring of recurrent events. When the occurrences of the event can only be determined through periodic assessments or laboratory tests performed at pre-scheduled visits, the event times can sometimes be interval-censored. In evaluating the effect of covariates on the gap times between interval-censored recurrent events, the complicating issue is that the gap time in between two events may be doubly interval-censored. To address this problem, we propose a method based on the discrete analogue of the proportional hazards model for the marginal distribution of each event. A robust estimator for the variance-covariance matrix is utilized to account for the potential within-subject correlation in the gap times between multiple events. An extensive simulation study indicates that the proposed method yields estimates which are less biased than a midpoint imputation approach. The proposed method is applied to data from the Systemic Lupus Erythematosus National Assessment (SELENA) study.