General Departmental Seminar Series
Some Problems in Survival and Event History Analysis Arising
from Intermittent Follow-up of Individuals
Jerry Lawless, Ph.D.,
Department of Statistics and Actuarial Science,
University of Waterloo
September 22nd, 2004, 4 - 5 pm in 140 Bardeen Medical Laboratory, 1300 University Ave.
When individuals in a follow-up study can only be observed intermittently, information on events between successive observation times is ascertained retrospectively, and information may often be incomplete. In addition, persons may become lost to follow-up at some unspecified time. This talk will discuss three problems associated with such situations: (i) the assessment and treatment of non-independent losses to follow-up in studies with widely spaced observation times, (ii) the treatment of censoring, delayed reporting of events, and loss to follow-up in observational studies based on clinic data bases, and (iii) inference problems requiring modeling of the observation process. Example from disease follow-up studies and from longitudinal surveys will be considered.
This seminar is jointly sponsored with the Department of Statistics.
Return to seminar list