Tracking Associations Between Ambient Ozone and Asthma-Related Emergency Department Visits Using Case-Crossover Analysis

Abstract

Traditional environmental public health surveillance consists of separately measuring hazards, exposures, and health outcomes. The Environmental Public Health Tracking (EPHT) Network seeks to accrue additional information by linking hazard or exposure data to health outcomes data. A natural progression is to consider tracking the "link" itself, that is, to track the association between an environmental hazard and a health outcome. The Maine EPHT Program conducted a case-crossover analysis to measure associations between daily estimated ambient ozone and particulate matter (PM2.5) and asthma-related emergency department (ED) visits for 2000-2003. We found an overall association of 7 percent (95% confidence interval, 4-11) excess asthma-related ED visits per 10-ppb increase in ozone averaged over 4 days (lag 0-3). The association was positive in the first 3 years and negative in the last. The excess risk was concentrated among females aged 15 to 34 and males younger than 15. The methodology for tracking associations between ambient air quality and acute morbidity is not generalizable to most other EPHT topic areas, but there are ample reasons to pursue this activity. The analysis can potentially help evaluate the effectiveness of regulatory and intervention programs, as well as inform us about trends, sensitive subpopulations, and changing potency of air constituents.

Evaluating Potential Response-Modifying Factors for Associations between Ozone and Health Outcomes: A Weight-of-Evidence Approach

“…However, some studies have reported positive associations among both children and adults, with no evidence of effect measure modification by age (Ko et al 2007; Mar and Koenig 2009; Paulu and Smith 2008). …”

“…For example, there is evidence for higher relative risk estimates in females compared with males in studies of chronic obstructive pulmonary disease HAs and ED visits (Arbex et al 2009; Wong et al 2009). However, in studies of asthma ED visits, differences between males and females were observed by age: Larger relative risk estimates were reported for males 2–14 years and females 15–34 years of age, with no evidence of any sex differences in those 35–64 years of age (Paulu and Smith 2008). These results are consistent with Thaller et al (2008), who found evidence of decreased lung function in females compared with males 16–27 years of age.…”

A Decade of Environmental Public Health Tracking (2002-2012)

“…Overall, a majority of the articles focused on outdoor air pollution and accounted for 37% of publications in this category. The articles that demonstrated a specific adverse health outcome or health effect associated with outdoor air pollution is as follows: asthma (38%), 14,19,42,43,48,55,66,69 acute myocardial infarction (19%), 14,57,58,69 cancer (10%), 21,34 adverse birth or reproductive outcomes (10%), 68,75 other respiratory diseases (10%), 42,56 and other cardiovascular disease outcomes (10%). 61,68 Nine percent of the articles in the Science and Research category focused on water contaminants.…”

Engaging academia to advance the science and practice of environmental public health tracking

“…Tracking Program partners have evaluated the association between ambient levels of ozone and PM 2.5 and traffic density with short-term asthma-related health outcomes (Babin et al, 2007; Paulu and Smith, 2008; Wilhelm et al, 2008). These analyses identified age- and sex- specific populations at risk as well as differences in the exacerbation of asthma-related outcomes particular to the exposure scenario.…”

“…It's unclear how modeling decisions, data characteristics, outcome sparseness, strength of association, and spatial autocorrelation may affect results of this and similar studies. Similarly, the case-crossover method is used frequently to analyze the effects of air pollution and heat, however, we need to understand the effects of decisions such as referent period selection and estimation of air pollution and meteorological conditions on study results (Paulu and Smith, 2008). Sophisticated modeling approaches for air pollution, such as kriging or land use-based regression modeling, better characterize spatial and temporal variability in air pollution (Jerrett et al, 2005; Wilhelm et al, 2008), however, they require expertise and resources that are not always available within a single institution or agency.…”