Variance-corrected recurrent models in medical studies

In many medical and epidemiological studies, each person can experience several events one after the other sequentially under some circumstances, known as recurrence events. Broad analytical objectives of recurrent events consist of describing event process in an individual, dispersion of events from one person to another, and determination of relations between constant or time-dependent predictors with event time. So the main purpose of this article is to present and interpret advanced statistical models to analyze such events which have been proposed in recent years.
In this article, variance-corrected models including AG, conditional PWP (PWP-TT and PWP-GT) and marginal WLW models for analyzing recurrent events are presented and compared with each other.
When the number of events for each individual are small and the risk of a reoccurrence of an event varies between different events, usually conditional PWP models are used, while marginal WLW and the time-dependent AG models are applied to determine frequency of events.
Medical and epidemiology studies should consider various factors such as the number of events, relationship between events, the variability of the effect of factors from one event to another, the biological process, and the correlation structure of data to select the valid recurrent models.