|
Activity Number:
|
165
|
|
Type:
|
Luncheons
|
|
Date/Time:
|
Monday, August 7, 2006 : 12:30 PM to 1:50 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #306544 |
|
Title:
|
Sampling from Large Cohorts When Covariate Ascertainment Is Expensive
|
|
Author(s):
|
William Barlow*+
|
|
Companies:
|
Cancer Research and Biostatistics
|
|
Address:
|
1730 Minor Ave., Seattle, WA, 98101,
|
|
Keywords:
|
sampling ; cohort designs ; case-control ; case-cohort ; screening
|
|
Abstract:
|
We will discuss sampling strategies when we have a large cohort available with known outcomes. We wish to sample efficiently to reduce costs of covariate ascertainment. The covariates could be biomarkers or risk factors obtainable only by medical record review. Possible choices for sampling include nested case-control, case-cohort, and counter-matching designs. We will concentrate on analyses that use failure time outcomes. The possible choices will be illustrated with reference to evaluating the efficacy of mammography screening on breast cancer mortality when chart review of screening history is required. Participants are welcome to contribute other possible applications.
|
