A study of grouped failure time data and the misapplication of recurrent event modeling |
| |
Authors: | Melissa C Braschel Gerarda A Darlington Neil Klar |
| |
Institution: | 1. Department of Mathematics and Statistics, University of Guelph, Guelph, Ontario, Canada;2. Department of Epidemiology and Biostatistics, Western University, London, Ontario, Canada |
| |
Abstract: | Clinical trials often assess whether or not subjects have a disease at predetermined follow-up times. When the response of interest is a recurrent event, a subject may respond at multiple follow-up times over the course of the study. Alternatively, when the response of interest is an irreversible event, a subject is typically only observed until the time at which the response is first detected. However, some recent studies have recorded subjects responses at follow-up times after an irreversible event is initially observed. This study compares how existing models perform when failure time data are treated as recurrent events. |
| |
Keywords: | Generalized estimating equations Generalized linear mixed models Hazard ratio Logistic regression Odds ratio Survival analysis |
|
|