Multiple statistics for multiple events, with application to repeated infections in the growth factor studies.

Abstract

Clinical studies that involve the recording of two or more distinct and well-defined events on each subject give rise to multiple event data. Treatment comparisons are usually reported in univariate analyses of time to first event or number of events observed. However, this approach may not uncover the 'full story' of the treatment effect; moreover, it may be inefficient because it does not make full use of the available data. There are a number of published statistical methods for analysing multiple event data. Using data from a real life example, this paper compares the results obtained using the 'older ad hoc' methods with those based on the more recent methods that utilize the multiplicity of the data.