|
Activity Number:
|
56
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 8, 2004 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #301249 |
|
Title:
|
Using Tensor Product Splines in Modeling Exposure-time-response Relationships: Application to the Colorado Plateau Uranium Miners Cohort
|
|
Author(s):
|
Kiros Berhane*+ and Michael Hauptmann and Bryan Langholz
|
|
Companies:
|
University of Southern California and National Cancer Institute and University of Southern California
|
|
Address:
|
Dept. of Preventive Medicine, Los Angeles, CA, 90089-9011,
|
|
Keywords:
|
splines ; latency ; nested case control ; dose-response ; excess relative risk
|
|
Abstract:
|
We propose tensor product spline-based flexible modeling techniques for describing exposure-response relationships for protracted time-dependent exposure histories in epidemiologic studies of occupational exposures. The methods use multidimensional nonparametric estimation techniques based on tensor product splines to jointly model age, time-since-exposure (latency) and exposure-response effects. The main advantage of this proposed methodology is its ability to allow for latency functions that vary by exposure levels and, conversely, exposure-response relationships that are influenced by the latency structure. We present a detailed analysis of data from the Colorado Plateau Uranium Miners cohort. We compare our results with previous analyses of the same dataset via other less flexible techniques that focus on flexible modeling of only the latency function or the exposure response relationship, by keeping the other fixed at a predetermined functional form.
|
