|
Activity Number:
|
417
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Health Policy Statistics
|
| Abstract - #307048 |
|
Title:
|
Spatial Statistical Methods for Small-Area Health Data with Application to the Association of Breast Cancer Incidence and Local Power Plant Emissions
|
|
Author(s):
|
Heather Watson*+ and Judith D. Goldberg and Mengling Liu
|
|
Companies:
|
New York University and New York University School of Medicine and New York University School of Medicine
|
|
Address:
|
324 E. 52nd Street, New York, NY, 10022,
|
|
Keywords:
|
spatial statistics ; small area ; aggregate ; disease clustering
|
|
Abstract:
|
To analyze small area health data, several spatial statistical methods are compared. Data is aggregated to different administrative levels. Small area health data produces unstable rates and the scale of aggregation has an impact on the inferences. Smoothed disease maps using Bayesian methods and spatial clustering methods for localized disease clustering or focused disease clustering near an environmental hazard are compared. Applied to female breast cancer cases reported from 1994-2000 in Rockland County and parts of New York City at the census tract and zip code level, the methods assess breast cancer incidence rates in excess near local power plants. Although the techniques aggregate the local small area differences and the methods of analysis differ, the techniques provide complementary inferences about areas of excess disease in the geographic region.
|
