Abstract:
|
Studies have shown that air quality is associated with population health. Health and air data are collected by two independent data systems, the National Health Interview Survey (NHIS) and the Air Quality System (AQS), respectively. To overcome the limited spatial-time coverage of AQS monitor data, an extensive model-predicted universe of spatial-time air measurements was created. These approaches were developed by the Environmental Protection Agency (EPA) and adopted for use in public health by CDC's Environmental Public Health Tracking Program, National Center for Environmental Health (NCEH). From this universe, air measurements for PM2.5 and ozone were linked at the census tract level to the NHIS sample over years 2001 to 2010 for those areas within the contiguous-US. As this linkage is complete, air/health association analyses on the NHIS can be performed using suitable design-based or model-based methods in contrast to analyzing a partial air/heath linkage with the original AQS air data. This study is somewhat exploratory with the needs of a typical NHIS data user in mind. The study attempts to determine some of the operating characteristics of the linked data and presents suggestions for design-based and model-based analyses. In particular, some basic spatial-time associations are explored. Some thoughts on next steps are also discussed at the end of the paper.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.