Activity Number:
|
422
|
Type:
|
Contributed
|
Date/Time:
|
Wednesday, August 5, 2009 : 8:30 AM to 10:20 AM
|
Sponsor:
|
Section on Statistics and the Environment
|
Abstract - #303631 |
Title:
|
Hierarchical 2-Level Cluster Model to Estimate Regional and National Adverse Effects of NO2 on Public Health
|
Author(s):
|
Hwashin H. Shin*+ and Richard T. Burnett and Dave Stieb and Barry Jessiman
|
Companies:
|
Health Canada and Institute of Population Health, University of Ottawa and Health Canada and Health Canada and Air Health Science Division, Health Canada
|
Address:
|
Environmental Health Science and Research Bureau, Ottawa, ON, K1A 0K9, Canada
|
Keywords:
|
air pollution ; , Bayesian hierarchical model ; 2-level cluster model ; random effects model ; Poisson generalized additive model ; mortality
|
Abstract:
|
We propose a Bayesian hierarchical cluster model to estimate annual mortality risk associated with nitrogen dioxide exposures at city, region and nation levels. The estimator also estimates heterogeneity among cities within region and that among regions. The model is based on time series using Poisson generalized additive model and applied to 24 Canadian cities for 1984--2004. The cities are considered to be random and clustered into 4 geographical regions. The annual national risk estimates from 1 and 2-level cluster models appear to be similar to each other and highly correlated (r=0.99). The annual regional risk estimates are less correlated with one another and some regions show negative correlations for warm season. The heterogeneity among regions is usually larger than that among cities within region (p=0.00), and increasing time trends in both heterogeneities were detected.
|