Abstract:
|
Many administrative health databases are organized by a calendar time window and doubly-censored data naturally arises when the events are only observed within a range bounded by the entry and the end of a study, possibly together with the births of subjects. In addition, administrative health databases evolve over time and the regulations about their access may change. Such changes pose another layer of complexity to the statistical challenges in the ability to analyze administrative data from different time periods. Motivated by the demand to compare patterns in frequency of pediatric emergency department visits between two groups born in different eras based on information from a health administrative database, this talk will present approaches to analyzing doubly-censored recurrent event ED visit data from different eras. A class of marginal rate models for recurrent events is employed to explore the two groups' difference in age-varying effects of potential risk factors. This is a joint work with Dr. Joan Hu (Simon Fraser University) and Dr. Rhonda Rosychuk (University of Alberta).
|