Premium
Functional measure of ozone exposure to model short‐term health effects
Author(s) -
Arisido Maeregu Woldeyes
Publication year - 2016
Publication title -
environmetrics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.68
H-Index - 58
eISSN - 1099-095X
pISSN - 1180-4009
DOI - 10.1002/env.2394
Subject(s) - ozone , environmental science , pollutant , term (time) , measure (data warehouse) , regression analysis , environmental health , statistics , meteorology , medicine , mathematics , geography , computer science , chemistry , physics , quantum mechanics , database , organic chemistry
This article explores the issue of measuring daily ozone exposure for epidemiological studies of pollutants and their effects on health. Generally, hourly monitored ozone has been reduced to daily summary measures, which are then assumed to represent daily individual exposure to the pollutant. We instead propose the use of a functional measure of ozone exposure, which means, all hourly ozone measurements from a single day, to model short‐term health effects. Functional data analysis techniques are used to convert hourly monitored ozone from a single day into a function accounting for the daily profile of ozone. The role of this functional exposure measure is to better represent daily individual exposure and to identify the periods of daily ozone exposure that are most harmful to health. This is demonstrated by analyzing the relationship between daily hospital admissions and ozone exposure in the Italian city of Milan using functional regression models. A statistically significant association between hospital admissions and this functional measure of ozone exposure was found in the afternoon hours of the day. We propose that the functional measure of ozone exposure has superior predictive performance compared with that obtained by measuring daily exposure through summary measures. Copyright © 2016 John Wiley & Sons, Ltd.