Abstract:
|
Recurrent adverse events often occur for some time duration in clinical studies. The number of events along with their duration-time is clinically considered as a measure of the severity of a disease. All the methods available for analyzing recurrent events data estimate either the event-rate (in terms of intensity or hazard) or the total duration-time per subject or rate and total duration side by side. No effort is made so far to combine them for producing a measure of the severity of the disease. Here a non-parametric estimator with asymptotic Normality property and easily derivable variance function is derived, which estimates the severity of a relapsing disease by combining the number of event recurrences (signifying the intensity of the disease severity) with the event duration-time (signifying the extent of the disease severity). The proposed estimator is algebraically very similar to the cumulative cost measure but can be expressed as the cumulative (integrated over time) weighted combination of event-recurrence rate and the event-duration (continuation) rate, where the weights are the varying (over time) proportion of patients, to whom the event recurred and continued.
|