B0291
Title: Semiparametric trend analysis for stratified recurrent gap times under weak comparability constraint
Authors: Peng Liu - University of Kent (United Kingdom) [presenting]
Abstract: Recurrent event data are frequently found in many clinical trial studies and medical research, where each subject encounters more than one sequential event. A much-discussed aspect of the recurrent events is the presence or absence of the time trend, where trend refers to a systematic variation among the length of the sequential gap times, which can be used as a measure of disease progression. Under the accelerated failure time (AFT) model, a comparability concept has been previously proposed to estimate the trend among sequential gap times (slope parameter). Each individual gap time has the same distribution subject to the comparability constraint, and thus the estimation can be easily conducted. However, their comparability is a strong constraint. We propose a weak comparability constraint under the previos assumption for the AFT model. The estimator will utilize more data due to weaker constraint, and thus it will result in a more efficient estimate for the slope parameter in the AFT model. Monte Carlo simulation is performed to validate the effectiveness of the new method. The method is applied to the HIV Prevention Trial Network (HPTN) 052 study.
