Estimating constrained concentration-response functions between air pollution and health

Powell, H., Lee, D. and Bowman, A. (2012) Estimating constrained concentration-response functions between air pollution and health. Environmetrics, 23(3), pp. 228-237. (doi: 10.1002/env.1150)

Full text not currently available from Enlighten.


The health risks associated with short-term exposure to air pollution have been the focus of much recent research, most of which has considered linear Concentration-Response Functions (CRFs) between ambient concentrations of pollution and a health response. A much smaller number of studies have relaxed this assumption of linearity, and allowed the shape of the function to be estimated from the data. However, this increased flexibility has resulted in CRFs being estimated that appear unfeasible, often showing decreases in the risk to health with increasing concentrations. Therefore this paper proposes a Bayesian hierarchical model for estimating constrained concentration-response functions in this context, which is based on monotonic integrated splines. These splines produce non-decreasing CRFs, due to the associated regression parameters being constrained to be non-negative, which we ensure by modelling the latter with a `slab and spike' prior. The efficacy of our approach is assessed via simulation, before being applied to a study of ozone concentrations and respiratory disease in Greater London between 2000 and 2005.

Item Type:Articles
Glasgow Author(s) Enlighten ID:Bowman, Prof Adrian and Lee, Professor Duncan and Powell, Miss Helen
Authors: Powell, H., Lee, D., and Bowman, A.
Subjects:H Social Sciences > HA Statistics
College/School:College of Science and Engineering > School of Mathematics and Statistics
Journal Name:Environmetrics
Publisher:John Wiley and Sons
ISSN (Online):1099-095X
Published Online:10 February 2012

University Staff: Request a correction | Enlighten Editors: Update this record