Assessing transient carryover effects in recurrent event processes, with application to chronic health conditions

Abstract

In some settings involving recurrent events, the occurrence of one event may produce a temporary increase in the event intensity; we refer to this phenomenon as a transient carryover effect. This paper provides models and tests for carryover effect. Motivation for our work comes from events associated with chronic health conditions, and we consider two studies involving asthma attacks in children in some detail. We consider how carryover effects can be modeled and assessed, and note some difficulties in the context of heterogeneous groups of individuals. We give a simple intuitive test for no carryover effect and examine its properties. In addition, we demonstrate the need for detailed modeling in trying to deconstruct the dynamics of recurrent events.