Estimating a Treatment Effect with Repeated Measurements Accounting for Varying Effectiveness Duration
To assess treatment efficacy in clinical trials, certain clinical outcomes are repeatedly measured for same subject over time. They can be regarded as function of time. The difference in their mean functions between the treatment arms usually characterises a treatment effect. Due to the potential existence of subject-specific treatment effectiveness lag and saturation times, erosion of treatment effect in the difference may occur during the observation period of time. Instead of using ad hoc parametric or purely nonparametric time-varying coefficients in statistical modeling, we first propose to model the treatment effectiveness durations, which are the varying time intervals between the lag and saturation times. Then some mean response models are used to include such treatment effectiveness durations. Our methodologies are demonstrated by simulations and an application to the dataset of a landmark HIV/AIDS clinical trial of short-course nevirapine against mother-to-child HIV vertical transmission during labour and delivery.
Biostatistics | Clinical Trials | Longitudinal Data Analysis and Time Series | Multivariate Analysis | Statistical Methodology | Statistical Theory
Chen, Ying Qing; Yang, Jingrong; and Cheng, Su-Chun , "Estimating a Treatment Effect with Repeated Measurements Accounting for Varying Effectiveness Duration" (November 2005). UW Biostatistics Working Paper Series. Working Paper 265.
Biostatistics Commons, Clinical Trials Commons, Longitudinal Data Analysis and Time Series Commons, Multivariate Analysis Commons, Statistical Methodology Commons, Statistical Theory Commons