Advertisement
Published in:
2015 | OriginalPaper | Chapter
Accounting for Temperature when Modeling Population Health Risk Due to Air Pollution
Authors : Wesley S. Burr, Hwashin H. Shin
Published in: Interdisciplinary Topics in Applied Mathematics, Modeling and Computational Science
Publisher: Springer International Publishing
Activate our intelligent search to find suitable subject content or patents.
Select sections of text to find matching patents with Artificial Intelligence. powered by
Select sections of text to find additional relevant content using AI-assisted search. powered by
Abstract
Air Health Indicator (AHI) is a joint Health Canada/Environment Canada initiative. A component in the indicator is an estimate of the time-dependent population health risk due to short-term (acute) effects of air pollution. The standard approach for this risk estimation uses a generalized additive model (GAM) framework, which includes one or more air pollutants and one or more temperature terms as covariates, as well as a smooth function of time. In this risk-modeling framework, the temperature is not the primary focus, but is included to ensure that common structure between the mortality (response), the pollutant(s), and the temperature is not included in the risk attribution.
We examine the smooth function link that is commonly used when including temperature. We show that for a single lag of temperature, the traditional J-, U-, or V-shaped relationship between temperature and mortality is largely a function of low-frequency mortality structure and is thus accounted for by the smooth function of time typically included in risk models. We further compare and contrast the first two primary lags of temperature in the context of these findings, and demonstrate differences in their structure, advocating the inclusion of only the first (lag-0) parametric temperature series in the model.