Air Quality Data Workshop

spacer

The Health Effects Institute

HEI/EPA Workshop on Air Quality Data in Health Effects Research

Newton Marriott Hotel, Newton, MA
November 30-December 1, 2006

A number of epidemiology studies are getting underway to evaluate the role of PM components and sources in increasing disease risk in the US. These studies will be using air quality datasets such as the urban STN and rural IMPROVE monitoring networks to develop exposure assessments. Early efforts have revealed several challenges to achieving uniform methods for exposure assessment. These challenges raise concern for the interpretation and comparison of risk estimates across the studies. Users of the air quality data are finding they need more information to understand the basis for the final measurements reported in air quality datasets. The purpose of this meeting is to bring together epidemiologists, exposure assessors, atmospheric scientists and other relevant researchers to consider how to best use the accumulating air quality data on PM components to address the important hypotheses of interest for health outcomes.

Presentations available
Please click on the links in the program below to download presentations (6 slides per page for easy printing). All files are secure PDF and are less than 2 MB unless indicated otherwise.

Thursday, November 30

Introduction: Goals of the meeting

Jon Samet (Johns Hopkins)

a) To bring together epidemiologists, exposure assessors, atmospheric scientists and other relevant researchers to discuss potential problems in using the available PM components data in epidemiologic research.
b) To understand what PM_2.5 components – and other PM characteristics – data are currently available, the limitations of the existing data, and what pollutant data will be available in the future.
c) To understand how existing air pollutant data and other information are being used to develop classifications of exposure to PM characteristics.
d) To determine how air pollutant data and other information are being linked to health outcomes in different types of health outcome studies.
e) To recommend common data handling approaches that would be helpful to improve comparability of exposure classification in epidemiology studies, and to establish an approach for cataloguing problems that arise in use of the data.

Current and future measurements of air pollutants

Moderator: Ken Demerjian (SUNY Albany)

Joann Rice and Neil Frank (EPA, OAPQS)
a) What PM_2.5 mass, speciated PM_2.5, and gaseous pollutant data are currently being collected and made publicly available?
b) What biases and/or limitations exist in the reported measurements of individual air pollutants? Presentation will discuss in detail key issues of importance to epidemiology studies including;
    i) information concerning the use of blanks data to correct
        concentration measurements for carbon
    ii) measurement of organics by different networks (e.g., IMPROVE,
        STN) planned changes by EPA: implications for continuity in
        longitudinal analyses
    iii) information concerning detection limits and measurement of
        transition metals.
c) Future of the monitoring networks: What changes in monitor locations and methods for PM mass, components, and gases are expected in the EPA/state air quality networks for the next 5 – 10 years? Will there be sufficient stability/consistency for epidemiologic research?

Betty Pun (Atmospheric and Environmental Research Inc) - The HEI Air Quality Database

Progress in using PM components data in epidemiology studies.

Moderator: Dan Greenbaum (HEI)
Discussion of the scientific questions and hypotheses to be tested at the ability to address them using available air quality data. Issues will include: (1) How are raw pollutant data preprocessed into an analytical dataset useful for health outcome studies? (2) What processing decisions are most influential in determining the final exposure classifications? (3) How have spatial and temporal variation been estimated for PM mass and/or species?

Michelle Bell (Yale) – Using speciated PM_2.5 data in mortality studies

Kaz Ito (New York University) – Using speciated PM components data in morbidity and mortality studies (1)

Sverre Vedal (University of Washington) – Using speciated PM components data in morbidity and mortality studies (2)

Friday, December 1 Back to top

Using air pollutant and other data to address important hypotheses about the role of PM components in determining health outcomes.

Moderators: Mark Utell (U Rochester) and Phil Hopke (Clarkson University)

Mark Utell – Introduction: What are the key health effects questions?

Paige Tolbert (Emory) – Lessons from ARIES

Lucas Neas (EPA) – Challenges to testing hypotheses in PM components studies

Panel discussion: Key issues in data collection and databases, statistical modeling, epidemiologic analyses - Synthesizing, prioritizing and proposing solutions. Panelists: Lucas Neas, Paige Tolbert, Joel Schwartz (Harvard University) and Bryan Hubbell (replaced by Beth Hassett-Sipple, both EPA, OAQPS).
    a) What are the challenges to using existing data in different types of
         epidemiologic studies (time series, case-crossover, longitudinal
         cohort). Detailed discussion of some key limitations of existing
         datasets to characterize spatial and temporal variation: imputation
         of missing data over time and space; correlation among components,
         handling of data below detection limits, other sources of measurement
         error.
    b) What current issues are most significant in affecting comparability
          of results?
    c) Other challenges; for example, ultrafine and coarse particles.
    d) What are the issues for which we can agree on a solution?

Phil Hopke (2MB)– Approaches to using PM components data to identify sources

Continued discussion of how to arrive at solutions
    e) For issues that require further information/discussion, what are the
         next steps needed to arrive at a solution?
    f) What further pollutant information or other data that are lacking
         in the present datasets will be useful for future health studies?

Concluding comments: How successful was the meeting in achieving its goals?

Dan Greenbaum, HEI and Dan Costa (EPA, ORD)

Possible recommendations for action items:
a. A written document for investigators using air quality data to delineate key strengths and limitations of the data and future recommendations?
b. What are the best ways to continue dialogue between air quality monitors and health study researchers?
c. Develop an online resource/bulletin board for air quality information users?

Organizing Committee:
Barbara Glenn, Stacey Katz, Gail Robarge - EPA;
Geoffrey Sunshine, Amara Ezeamama – HEI;
Ken Demerjian, SUNY Albany;
Scott Zeger, Johns Hopkins;
Michelle Bell, Yale;
Betty Pun, Atmospheric and Environmental Research, Inc.

Last updated December 18, 2006