2002 Progress Report: Air Quality and Reported Asthma Incidence in IllinoisEPA Grant Number: R829402C003
Subproject: this is subproject number 003 , established and managed by the Center Director under grant R829402
(EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
Center: Center for Integrating Statistical and Environmental Science
Center Director: Stein, Michael
Title: Air Quality and Reported Asthma Incidence in Illinois
Investigators: Frederick, John , Draghicescu, Dana , Dukic, Vanja , Eshel, Gidon , Naureckas, Edward , Rathouz, Paul , Zubrow, Alexis
Current Investigators: Frederick, John , Draghicescu, Dana , Dukic, Vanja , Eshel, Gidon , Im, Haekyung , Naureckas, Edward , Rathouz, Paul
Institution: University of Chicago
EPA Project Officer: Packard, Benjamin H
Project Period: March 12, 2002 through March 11, 2007
Project Period Covered by this Report: March 12, 2002 through March 11, 2003
RFA: Environmental Statistics Center (2001) RFA Text | Recipients Lists
Research Category: Environmental Statistics , Ecological Indicators/Assessment/Restoration , Health , Ecosystems , Air
This project focuses on statistical investigations into the relationship between air quality and respiratory health in the Chicago area. The specific objectives of this research project are to: (1) identify atmospheric states most likely to be accompanied by elevated ground-level ozone amounts in Illinois; (2) elaborate on and develop statistical models to describe the spatio-temporal structure in ground-level ozone; (3) examine and compare different measures of asthma incidence for use in population-based studies of the effects of air pollution on respiratory health; and (4) develop statistical models to link acute asthma occurrence in Chicago’s Medicaid population to levels of ozone, particulate matter, pollen, and meteorological conditions in both urban-aggregated and spatially resolved ways.
Research results to date are summarized below for each objective. Problems encountered are mostly centered on the spatial, and in some cases the temporal, coverage of specific data sets related to atmospheric conditions, air quality, and health outcomes. The health outcome data are organized spatially at the level of ZIP code. On any given day, ozone is measured hourly at a minimum of 11 sites in Cook County, but this still does not provide sufficiently dense coverage for all of Cook County’s 56 ZIP codes. Pollen is a significant factor in asthma occurrence, but the available data include only one daily value for the entire Chicago area. The data set for large particles, particulate matter (PM10), has sparse spatial coverage and contains information at irregular time intervals, where the period between data points at a specific site can be up to 6 days. Data on PM2.5 are not available for the same time frame for much of our health outcome data.
Therefore, we are faced with significant problems of incompatible space-time scales across the multiple data sources to be used in our research project. In response to the issues posed by such problems of scale, we are investigating a variety of statistical spatio-temporal modeling techniques, the use of independent data sets with higher spatial resolutions, and physical models (MM5, Community Multiscale Air Quality [CMAQ]) for use in generating spatio-temporal structure in localized factors that may influence asthma incidence. The problems identified are significant, and they have served as a part of the wider impetus for developing important new statistical methodology that will be applicable in our study and in other studies with similar study designs. Consistent with the evaluation provided by the external Advisory Board, upcoming work in this research project will make extensive use of new spatio-temporal statistical methods (developed in another Center for Integrating Statistical and Environmental Science [CISES]-supported effort), as well as the capability in physical modeling of meteorology and air quality at high spatial resolutions. In addition, the longer-term objective of developing improved statistical models for linking acute asthma events to air quality indicators is a driving force behind most of our efforts, and work will continue on these models for the duration of the research project. The effort devoted to identifying relationships between meteorological regimes and air quality has been completed (Eshel and Bernstein, 2003), and further work in this area will not be undertaken.
Research Results to Date
Meteorological States That Accompany Elevated Ground-Level Ozone Amounts. A strengthened capability to characterize atmospheric states that promote degraded air quality is important for improving forecasts of conditions that may pose health-related risks. Statistical analyses of ground-level ozone in Illinois, in conjunction with radiosonde data for the decade of the 1990s, defined the regional meteorology most likely to be associated with poor air quality. Results show that elevated ozone amounts tend to occur during periods of strengthened subsidence over the state, leading to a shallow boundary layer. The prevailing downward motions inhibit ventilation of the boundary layer by convection, thereby allowing a buildup of ozone. This work was completed in Year 2 of the project, and a paper describing all aspects of the work currently is in review (Eshel and Bernstein, 2003).
Statistical Models of Spatio-Temporal Structure in Ozone. The asthma outcomes of interest in this work are available daily at the level of individual ZIP codes throughout Chicago, but the air quality data exist on a totally different grid. The objective is to develop and apply spatio-temporal models to generate air quality information with a spatial resolution compatible with the medical information. The quantity predicted to date is the maximum 8-hour average ozone mixing ratio encountered during a 24-hour period at the population centroid of each ZIP code, and also at the ozone monitoring stations. Tests of the models have used ozone measured at a given subset of these stations over a period of time to predict ozone amounts at other stations where physical measurements also exist. In the first analysis, we predicted daily ozone at each ZIP code using traditional spatial kriging applied independently to each day's data. This approach clearly is inadequate because it does not exploit the fact that the data on successive days are highly correlated. Therefore, new models are needed that model the ozone process over both space and time. Such space-time models are challenging to develop and fit. To date, we have fitted preliminary versions of space-time models to ozone data in Chicago over periods of 2 weeks. Tests of these model fits demonstrated that the spatio-temporal models predict the ozone values more accurately than traditional kriging on each day’s data separately. A major challenge is to develop computational tools for space-time estimation over longer periods of time, so that the predicted values can be used in the analyses of human health outcomes.
Validation of Alternate Measures of Asthma Outcomes. This work is still in progress as part of graduate student Xiaoming Bao’s research project. The Medicaid database includes various types of information recorded on a daily basis, including: albuterol prescriptions, and visits to emergency departments (ED) and hospitalizations. The last of these are clear markers of (severe) asthma occurrence because the diagnosis code associated with the billing event is unambiguously asthma. Albuterol prescriptions also may be strongly linked to asthma because this drug is very specific to asthma. In addition, the database for albuterol prescriptions is much larger than that for ED visits or hospitalizations, potentially containing more information about the air quality and respiratory-health relationship. It is, therefore, of value to determine the degree to which these prescriptions may be useful as markers of acute asthma outcomes. The central questions addressed are: (1) Is there a significant relationship between albuterol prescription refills and hospitalizations or ED visits for asthma?; and (2) What is the temporal nature of that relationship? Statistical analyses demonstrate that a positive connection indeed exists and reveal the complexity of the relationship.
We have found, in preliminary results, that albuterol prescriptions rise 1 to 2 days after an ED visit or hospitalization; although a weaker link, ED visits and hospitalizations rise shortly after the filling of prescriptions. A possible explanation for this asymmetry is the fact that ED and hospital discharges almost always are accompanied by the issuance and subsequent filling of additional albuterol prescriptions. Furthermore, stratifying the data on whether or not the subject uses inhaled steroids shows that different patterns of correlation arise for the users as compared to the nonusers. The reason is that the medical condition of asthma appears to be better managed by steroid users, so that albuterol prescription refills are more regular and less strongly linked to acute events. Consistent with this, a stronger link between albuterol prescriptions and ED visits is observed for people who do not use steroids. This study partly validates the use of albuterol prescriptions as an asthma outcome, and it has implications for the temporal resolution that can be obtained in correlations with aspects of air quality. If degraded air quality is a trigger for asthma, a time lag can be expected before the coupling appears in the number of prescriptions filled.
Statistical Models of the Relationship Between Degraded Air Quality and Asthma Incidence. The efforts summarized above support the primary goal of the research project, to develop and apply statistical models to assess the link between the incidence of asthma and measures of air quality in Chicago, IL. The initial study of this link used an aggregate data model based on combining albuterol prescriptions for all ZIP codes in the Chicago metropolitan area. Independent variables were temperature, relative humidity, PM10, pollen, and ground-level ozone amounts. The PM10 was characterized by 24-hour integrals estimated for each day, while the index of ozone was the maximum 8-hour average value, averaged over all stations in the metropolitan area. In addition, the model allowed for unmeasured time-varying confounding variables that depend on only 1 day of the year, with peaks in spring and late summer/early fall. Results from this aggregated model reveal significant links between the number of daily medical claims and temperature (a negative correlation) and pollen levels (a positive correlation). No statistically significant association with air quality as measured by PM10 and ozone appears in the spatially aggregated model. We also have begun efforts on a spatially resolved asthma model. An initial Bayesian model based on zip code level data revealed a statistically significant positive link between ozone and reported asthma outcomes. This result is still preliminary because: (1) work on the model estimation technique is still ongoing; (2) some revisions in the data sources are to be incorporated into a final model; and (3) improvements based on spatio-temporal modeling of ozone and particulate matter are still to be incorporated in the model. These model upgrades currently are under investigation and will be developed during the remainder of Year 2 of the CISES project.
Relevance to the U.S. Environmental Protection Agency's Mission
The components of this project are of direct relevance to important concerns. These involve gaining an improved understanding of relationships between air quality and respiratory health, where this understanding provides a basis for policy formulation and evaluation of past actions. The results obtained thus far, and to be produced in the future, are relevant to: (1) improving the capability to forecast periods of degraded air quality from observations of regional meteorology; (2) identifying new respiratory health outcomes, such as albuterol prescriptions, that are useful in conjunction with air quality information; (3) determining quantitative links between respiratory health and measures of air quality in a major urban area; and (4) developing improved statistical methodologies for relating environmental factors to human health.
We will continue to improve the space-time modeling of air quality data. Major challenges include developing techniques for checking results produced by space-time models, and developing efficient computational tools that avoid construction and inversion of very large matrices so that these models may be estimated for a longer-time series of days. The meteorological and air quality models, MM5 and CMAQ, provide an alternate approach to generating trace gas and particulate matter information at high spatial and temporal resolution. Work in the upcoming period will compare pollutant levels based on MM5 and CMAQ to monitoring data in Chicago for specific subsets of the period 1995-1998. These comparisons will provide the basis for new regression-based predictions of air pollutant levels (ozone, PM10, PM2.5) and meteorological parameters (temperature, humidity). During Year 2 of the project, we will complete the analysis of the 1995-1998 Medicaid database for asthma at the aggregate (city-wide) level, and will complete preliminary modeling at the ZIP code level. Efforts here involve validating the Bayesian Markov Chain Monte Carlo statistical procedure, validating the numerical stability of the nonparametric smoothing procedure, rechecking and cleaning up any errors in the database, and reanalyzing the data. The results of the above work will form the basis of a technical report and a manuscript to be submitted to the epidemiological literature. Finally, upgrades on the ZIP code level model will continue. These include the incorporation of air quality information produced for individual ZIP codes into the model. In addition, efforts will begin on designing a model to relate individual-level respiratory health outcomes to degraded air quality. A test of the individual-level model will be its ability to reproduce results obtained from the completed aggregate model when integrated up to the city-wide level. This work will continue into Year 3, subject to renewal of the project.
Journal Articles on this Report : 1 Displayed | Download in RIS Format
|Other subproject views:||All 21 publications||4 publications in selected types||All 3 journal articles|
|Other center views:||All 115 publications||69 publications in selected types||All 47 journal articles|
||Eshel G, Bernstein JJ. Relationship between large-scale atmospheric states, subsidence, static stability and ground-level ozone in Illinois, USA. Water, Air, & Soil Pollution 2006;171(1-4):111-133.||
Supplemental Keywords:respiratory health, air quality, asthma, ozone, ground-level ozone, particulate matter, PM, pollen, meteorological conditions, atmospheric conditions, atmosphere, Chicago, Illinois, IL, Medicaid., RFA, Scientific Discipline, Health, Economic, Social, & Behavioral Science Research Program, PHYSICAL ASPECTS, Air, Geographic Area, Ecosystem Protection/Environmental Exposure & Risk, HUMAN HEALTH, Applied Math & Statistics, particulate matter, Health Risk Assessment, Ecosystem/Assessment/Indicators, Ecosystem Protection, Ecological Effects - Environmental Exposure & Risk, Monitoring/Modeling, Risk Assessments, Environmental Monitoring, Ecological Effects - Human Health, Health Effects, Physical Processes, Ecological Risk Assessment, Environmental Statistics, Environmental Engineering, Engineering, Chemistry, & Physics, Ecological Indicators, EPA Region, particulates, risk assessment, ecological effects, monitoring, asthma, atmospheric particulate matter, health risk analysis, watersheds, emissions monitoring, ecological health, ozone , particulate, stratospheric ozone, human health effects, ozone, sediment transport, airborne particulate matter, computer models, exposure, air pollution, chemical transport modeling, chemical transport, trend monitoring, environmental health effects, air pollutant-induced pulmonary inflammation, statistical models, human exposure, ecological risk, water, ecosystem health, environmental indicators, PM, ecological models, aersol particles, chemical transport models, Region 5, data models, modeling studies, air quality, statistical methodology, human health risk, stochastic models
Progress and Final Reports:Original Abstract
Main Center Abstract and Reports:R829402 Center for Integrating Statistical and Environmental Science
Subprojects under this Center: (EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
R829402C001 Detection of a Recovery in Stratospheric and Total Ozone
R829402C002 Integrating Numerical Models and Monitoring Data
R829402C003 Air Quality and Reported Asthma Incidence in Illinois
R829402C004 Quasi-Experimental Evidence on How Airborne Particulates Affect Human Health
R829402C005 Model Choice Stochasticity, and Ecological Complexity
R829402C006 Statistical Approaches to Detection and Downscaling of Climate Variability and Change