Anne Goldman
Division of Biostatistics
University of Minnesota
Wednesday, October 23, 2002
3:30 PM
PWB 2-470
Minneapolis Campus
Please note: This seminar is being held in a different location from our regularly scheduled talks.
Abstract:
Some clinical research programs attempt to follow their patients for a long
period of time. This is especially true in Cancer Centers and other programs
that treat patients with chronic disease because there is interest in the long-term
consequences of treatment, not just the primary outcome of response, relapse,
and survival. Blood and marrow transplants have the potential for curing the
underlying disease and many patients, who survive the very intensive pre- and
post transplant treatment, go on to lead a relatively normal life and a normal
life-span.
It is impractical and costly to follow an increasingly large population of surviving patients indefinitely. As the proportion of patients that can be followed becomes smaller with the years since transplant statistical analysis becomes more and more unreliable, both in terms of bias and variance of outcome estimates. We are therefore proposing a new strategy to the UMN BMT program that will restrict complete follow-up to a defined period of time. After that, patients will be followed for survival only, through annual searches of the National Death Index.
Data from the BMT research database will be presented with a detailed description of the proposed plan.