Here we outline the manner in which we defined the scope and content of the paper, determined the taxonomy, and collected information. Because this was somewhat of a simultaneous needs assessment and research project, we continually refined each of these areas as we obtained additional information, so that the document would be as useful as possible. Thus, we describe our methodology chronologically.
With the mission of the Center for Healthy Environments and Communities in mind, we began with the idea of a report describing the state of regional environmental health, based upon an earlier draft.[7] Initial exploration revealed that there already existed local reports or projects that did the following:
· Touched upon certain aspects of environmental factors or health in reasonable depth,[8] without necessarily linking the two fields.
· Gathered expert input to define indicators based upon a portion of health and environmental datasets.[9]
· Gathered input and reported on summary indicators for several broad areas, including environmental health.[10]
· Listed large numbers of online links and resources for health or environmental issues--or both--on a larger geographic scale, often nationwide.
There was not yet available, however, a descriptive inventory of the types and quality of existing environmental health data for the region. Such information, as described in the introduction, forms the foundation for many other endeavors.
We established a custom database for storage of information on various types of environmental health data sources, organizations, and contacts. This was partially modeled upon the past highly-detailed “metadata database” of the Baltimore City Data Collaborative, once managed by one of the authors (DW). This database not only helped us to organize our information, but forced us to determine a taxonomy for it. We populated the database with already-collected documents and links, and began to add additional information through the following:
· Browsing websites and already-gathered information sources
· Attending meetings or conferences (including online conferences)
· Contacting experts in the local, county and state-level government; non-profit organizations; private organizations; and academic institutions. This included individuals listed on websites or other reference materials, individuals with whom we had already had prior contact, and individuals referred to us through prior conversations
The types of questions we asked of experts, and which we had in mind when collecting information from other sources, are as follows:
· After hearing a description of our project, does it sound like we’re duplicating the efforts of anything that’s currently being done or that has already been done? If so, whom do we contact?
· What data have you or your organization collected that are related to environmental health? Where can these data be obtained, and who can obtain them?
· What are some major strengths and limitations of the data with which you’ve worked?
· Is there a more recent version of the work you did previously? Do you plan to do future versions, and if so, how often?
· What types of data related to health and the environment would you or your organization like to obtain, that are currently difficult to obtain, don’t exist, or have severe limitations?
Appendix H: Comprehensive Question List uses one health outcome (cancer) to illustrate what might data might be collected in a more comprehensive environmental health data inventory and needs assessment. We had hoped to use that longer list, but realized that time and staff resources were too limited, and rarely did one person have the answers for more than a few of the questions. While we considered developing a mass-distributed online survey to save time and gather a larger volume of input, we decided that the relatively closed format could limit the types of information collected, as we might be excluding important questions.[11]
In gathering this information, we initially focused on three broad areas that seemed to encompass most environmental pollutants posing potential health risks: land, water and air. As we continued to gather information, however, we realized the following:
· There were far too many sources to be able to provide information for these broad areas in much depth. For example, many national and state-level sources could also be applied to the Pittsburgh region; and within each of the three areas were numerous modes of gathering data (e.g., surveys, mechanical monitoring of the environment, reporting of hospital diagnoses).
· Areas such as the human-built environment didn’t seem to fit the “air, water, land” taxonomy.
· We realized that the information would probably change more rapidly than we could update it, and that projects focused upon creating online data collection, linkage and reporting instruments might make more sense as an updateable holding point for more detailed information in the longer term.
· Due to the breadth of the universe of environmental health data, even readers who are seasoned experts in one area might know very little about most of the other areas.
· Several very recent or still-in-planning endeavors already dealt with portions of what we originally set out to do. These projects, described elsewhere in this report, included a new statewide plan to inventory environmental health data and two recent or not-yet-published reports on regional water quality and related data sources.
We therefore made several changes in our approach. These included the following:
· We halted expansion of the metadata database in Microsoft Access, transferring much of the information into the relatively popular reference information package EndNote, and focusing energies upon producing a written document. This wouldn’t allow for as broad a range of information to be tracked on each source, but it would allow a large range of data sources to be entered more quickly—and perhaps eventually be transferred, with much of this report, to a “live” source where access by a large number of experts allows for maintenance of more in-depth information.
· We changed the taxonomy of data types to its current form, organized by points at which data on pollutants and outcomes can be collected, with air, water and land being merely subcategories below the environmental monitoring section. This seemed to cover the range of environmental health data more comprehensively and also seemed to make more sense from an environmental health data collection perspective—the previous taxonomy had seemed to fit primarily a non-health environmental perspective.
· We decided to look at a few additional topics:
o The human-built environment, because it didn’t fit the same model as environmental pollutants and related outcomes, as described elsewhere in this report.
o Endeavors to build data-related tools that could be useful from an environmental health perspective. Along with discussing data, we realized that to promote collaboration, we needed to help generate awareness of these projects.
o Other considerations that spanned beyond the data themselves, but that impact data quality, such as political and environmental justice considerations.
· We decided to report less in-depth information and more general and explanatory information for each of the sections then we had originally planned, so we could cover a broader range and help readers to see how it all “ties together.” We began to treat our project more as a “starting point” that gives readers some initial direction and understanding, and refers them to other resources for more detail.
· In line with the previous decision, we decided to arbitrarily limit the total number of experts with whom we spoke due to time and resource limitations, even if it meant covering some topics in less depth. We decided to use distribution of this first document as a means of encouraging input from additional experts and organizations, regarding both content and possible future direction and format of the project.