The Environmental Public Health Tracking (EPHT)

Describing Environmental Public Health Data: Implementing a Descriptive Metadata Standard on the Environmental Public Health Tracking Network Jeff Patridge and Gonza Namulanda The Environmental Public Health Tracking (EPHT) Network provides an opportunity to bring together diverse environmental and health effects data by integrating local, state, and national databases of environmental hazards, environmental exposures, and health effects. To help users locate data on the EPHT Network, the network will utilize descriptive metadata that provide critical information as to the purpose, location, content, and source of these data. Since 2003, the Centers for Disease Control and Prevention s EPHT Metadata Subgroup has been working to initiate the creation and use of descriptive metadata. Efforts undertaken by the group include the adoption of a metadata standard, creation of an EPHT-specific metadata profile, development of an open-source metadata creation tool, and promotion of the creation of descriptive metadata by changing the perception of metadata in the public health culture. KEY WORDS: Dublin Core Metadata Set, Environmental Public Health Tracking Network, Federal Geographic Data Committee Content Standard for Digital Geospatial Metadata, International Organization for Standardization 11179, International Organization for Standardization 19115 The prefix meta is Greek and translates as beside or after. Metadata the data used to locate and describe other data are often referred to as data about data. Metadata is typically a structured list of welldefined metadata elements and their meaning. 1,2 The elements can be used to document the content of a dataset and describe and explain the dataset. Metadata can also provide a semantic directory by describing the meaning of dataset elements and their relationships to other data. 2 J Public Health Management Practice, 2008, 14(6), 515 525 Copyright C 2008 Wolters Kluwer Health Lippincott Williams & Wilkins Public health systems collect and generate large volumes of data that are used for several purposes. Locating these data and distinguishing the datasets that are appropriate to a public healthcare practitioner s specific purpose are often challenging. Metadata can be used to facilitate discovery of datasets, identify datasets, assemble similar datasets, distinguish dissimilar datasets, and provide access to datasets. 1 Standardizing the structure of metadata elements enhances the uses of metadata. A metadata standard allows for a systematic way of describing data resources. Therefore, metadata standards will facilitate searching for datasets that have been described using the same standard. Metadata standards also enhance communication among public healthcare practitioners because they promote a common understanding of the meaning and representation of each data element. 3 Common accepted metadata standards include International Organization of Standardization (ISO) 19115 and 11179, Federal Geographic Data Committee (FGDC) Content Standard for Digital Geospatial Metadata, and the Dublin Core Metadata Standard (ISO 15836). Metadata standards have been applied in public health to provide more useful data to answer public Disclaimer: The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention. The authors thank past and present Metadata Subgroup members, without whose efforts the accomplishments mentioned in this article would not have been possible. Corresponding Author: Jeff Patridge, MA, Metadata Subgroup, Missouri Office of Administration, 920 Wildwood, Jefferson City, MO 65109. Jeff Patridge, MA, is a member of Missouri s Environmental Public Health Tracking team, and Cochair, Metadata Subgroup, Missouri Office of Administration, Jefferson City. Gonza Namulanda, MS, is a public health informatics fellow with the National Center for Environmental Health, and Cochair, Metadata Subgroup, Centers for Disease Control and Prevention, Atlanta, Georgia. 515

516 Journal of Public Health Management and Practice health questions. 3 An example is the Common Data Standards metadata descriptors developed by the National Cancer Institute. These metadata descriptors are used to more consistently describe cancer data and other health data to enhance their aggregation and use. 3 The US Health Information Knowledgebase maintains a metadata registry that is based on the ISO 11179 metadata standard. Metadata registry standards support the sharing of data across different systems and the public health organizations that use them. 4 Because no universally accepted definition is available, the term metadata has different meanings to different persons, organizations, and professions. Attempts to refine the metadata definition have produced varied results. Dr William Y. Arms, a recognized expert from Cornell University on digital library development, identifies three types of metadata: descriptive, structural, and administrative. 5 The National Information Standards Organization recognizes these same types of metadata and has added subtypes of administrative metadata known as rights management and preservation metadata. 1 The Getty Institute classifies five types of metadata: administrative, descriptive, preservation, technical, and use. 6 Given all these metadata types and definitions, the fact that confusion surrounds the concept of metadata is not surprising. This is the difficulty the Environmental Public Health Tracking (EPHT) Network Metadata Subgroup faced when it first met in early 2003. The subgroup members who serve on the Centers for Disease Control and Prevention s (CDC) EPHT Standards and Network Development Workgroup had been tasked with developing or adopting a metadata standard for the EPHT Network. Although they did not know it at the time, these geographic information systems (GIS) professionals, epidemiologists, information technology professionals, public health educators, and administrators were beginning a journey that 4 years later would result in the adoption of a metadata standard, the development of a metadata creation tool, and the creation of an EPHT metadata profile. Adopting a Metadata Standard The EPHT Network s primary function is to bring together diverse environmental and health effects data by integrating local, state, and national databases of environmental hazards, environmental exposures, and health effects. 7 Without the ability to describe and locate these data, the functionality of the EPHT Network is hindered. That is why metadata has often been called the EPHT Network s backbone. The first effort undertaken by the subgroup was to define the type of metadata that would be available on the EPHT Network. After significant debate, the subgroup agreed upon the concept of descriptive metadata. They defined descriptive metadata as information that describes the content, quality, and context of a data resource for the purpose of facilitating identification and discovery. It may reference additional information such as quality assurance documents or data dictionaries. Early in the process, the subgroup members decided to focus on adoption of an existing descriptive metadata standard for the EPHT Network. They realized that it would require significant effort and resources to develop, govern, and promote the use of an EPHT Network specific metadata standard. Accordingly, the subgroup members consulted representatives of the CDC s Public Health Information Network and the Environmental Protection Agency s Exchange Network to determine the descriptive metadata standards adopted for use on these networks. From these discussions, three standards were identified for further evaluation: the Dublin Core Metadata Element Set (ISO 15836), ISO 11179, and the FGDC Content Standard for Digital Geospatial Metadata. Initially developed for libraries, the Dublin Core Metadata Set is increasingly being used to describe content on the World Wide Web. It consists of 15 descriptive elements. These elements are outlined in Appendix 1. 8 ISO 11179 was created by the International Organization for Standardization and specifies a basic set of data element characteristics necessary to share data. It places increased emphasis on characteristics such as identifiers, definitions, and classification categories. It also establishes guidelines for the creation and maintenance of a data element registry. ISO 11179 elements are listed in Appendix 2. 9 Although the FGDC Content Standard for Digital Geospatial Metadata was developed to describe GIS data, it also has the potential to describe nongeospatial data. In 1994, Presidential Executive Order 12906 dictated that all federal agencies utilize the FGDC Content Standard for the creation of geospatial metadata. An example of the high-level FGDC elements is shown in Appendix 3. 10,11 As a requirement for receiving federal funds to plan for local network development, the EPHT grantees were required to create an inventory of all health- and environment-related datasets in their states or cities. To evaluate which of the three descriptive metadata standards would best describe EPHT Network data resources, the subgroup members compiled a list of common elements from these data inventories. They reviewed the list to create a minimum information set that would adequately document a dataset, and then mapped the elements to the three standards. Using the results of this mapping effort, the subgroup members selected the FGDC standard, citing the

Describing Environmental Public Health Data 517 following principal reasons: The EPHT Network includes geospatial and nongeospatial data, and the FGDC standard gave the most flexibility for describing both types of data. The FGDC standard had a large number (>195) of total data elements that could be used to describe EPHT Network data. The FGDC standard was well established, having been developed in 1994 and updated in 1998. It is supported by a presidential executive order that requires federal geospatial datasets to be described using this standard. Testing After agreeing on a descriptive metadata standard, the subgroup expanded testing of the standard to the data stewards. They formed a team to oversee the testing and tasked them with (1) testing the FGDC Content Standard against public health datasets, (2) evaluating the usefulness of the US Geological Survey freeware metadata creation tool Tkme, 12 and (3) gathering feedback from the test participants. Seven testers, representing six states and one city, were chosen. The data stewards selected for the test had no previous experience with metadata creation. To complete the testing, each participant was asked to return a copy of the feedback form together with his or her completed metadata records. In their responses, the testers agreed that the FGDC standard adequately described public health data, but they had two recommendations. First, they requested the selection of a core set of FGDC elements that would constitute the bare minimum of data elements required for placing data on the EPHT Network. The testers were concerned that the time necessary to complete the full FGDC template documentation requirements would inhibit the creation of metadata. Second, the testers recommended the development of a tool that would improve metadata creation. The tool had to be user friendly and include built-in constraints (eg, date, state abbreviations). Testers found that Tkme was adequate for testing purposes but difficult for use by those who were not intimately familiar with the FGDC standard. Metadata Profile Development The subgroup members and the data stewards involved in testing reviewed the FGDC standard and developed what is today known as the EPHT Metadata Profile. The current profile contains the minimum FGDC-required elements for compliance and several additional, optional elements the group considered essential for documenting public health data resources. Appendix 4 provides a list of the EPHT Metadata Profile elements and their definition. A total of 52 elements constitute the minimum elements required for placing data on the EPHT Network. The subgroup members recommended that those data stewards who create and manage geospatial data or those who wanted to undertake more in-depth documentation use the full FGDC profile. Additional testing efforts and evaluations by data stewards and stakeholders have further refined the profile. Appendix Figure 1 provides an example of metadata for a fictitious cancer dataset created using the EPHT Metadata Profile. To help guide data stewards on profile completion, the Metadata Content Guidance Document has been created. The document provides a refined, user-friendly definition of each metadata element, recommends ways to complete each element, and contains metadata examples for various types of environmental public health data resources (Figure 1). Metadata Creation Tool Development Before testing, the subgroup researched those tools that were available for editing and creating FGDCcompliant metadata. The subgroup discovered that most tools were either part of proprietary software packages or not user-friendly. As previously mentioned, the subgroup settled on the freeware Tkme. After reviewing the posttesting recommendations, the subgroup reactivated the test team to create a prototype tool to spark further discussion; gather tool requirements; and package these requirements for use by a developer. The team developed a mock-up of a tool and held two requirements-gathering sessions with EPHT grantees and stakeholders. The agreed-upon requirements were documented, and they became the first version of the Metadata Tool Vision. The vision has served as the blueprint for the open-source metadata creation tool currently under development. When the tool is released, data stewards will have the ability to create FGDC-compliant metadata records for their data and be able to export an Extensible Markup Language (XML) record that can be uploaded to the EPHT Network s metadata registry. Metadata Use Cases The subgroup members also developed Metadata Use Cases. The initial five use cases outline creating descriptive metadata, uploading descriptive metadata,

518 Journal of Public Health Management and Practice FIGURE 1 An Example of a Guidance Document Entry updating descriptive metadata, searching descriptive metadata, and managing descriptive metadata. They provide guidance to all grantees and partners implementing a metadata effort. Promoting Metadata During its deliberations, the subgroup realized that a successful descriptive metadata effort would require a change in culture. Initial inquiries revealed that few public health agencies created descriptive metadata for the data they collected. In most of these agencies, GIS professionals were identified as the only group that was creating and maintaining descriptive metadata and that was for geospatial data layers only. To educate data stewards on the benefits of descriptive metadata creation, the subgroup worked closely with the CDC EPHT Program Marketing and Outreach Workgroup, leading to the development of a benefits document describing the value of creating quality metadata and a series of frequently asked questions (FAQs). These documents are part of a larger data steward benefits package that explains the EPHT Network and its functionality. They have also presented their results at several conferences and sponsored a Metadata Brownbag Webinar to further explain the concept of descriptive metadata. In November 2007, the subgroup members conducted a hands-on Web-based training in which more than 80 grantees, data stewards, and EPHT Network partners had an opportunity to create descriptive metadata using the Metadata Creation Tool. Conclusions Although far from complete, the effort to integrate descriptive metadata into the EPHT Network has provided several lessons learned. First, there is a real need for the creation of descriptive metadata for public health datasets. Descriptive metadata can help locate the voluminous disparate and scattered data resources necessary for both surveillance and research. Second, the number of existing metadata standards makes the creation of a homegrown standard not only

Describing Environmental Public Health Data 519 unnecessary but also impractical. Besides the outlay in resources to maintain such a standard, its use also hinders data-sharing efforts. Finally, installing any metadata standard into an organization that has not created metadata or has very little understanding of the concept is a challenge. Remember that work is not complete with the selection of a standard. Those tasked with putting the standard in place must continue to guide the process to the end, or the effort is likely to fail. Continued education, marketing, and technical assistance are requirements for a successful implementation effort. REFERENCES 1. National Information Standards Organization. Understanding Metadata. Bethesda, MD: NISO Press; 2004. http://www. niso.org/standards/resources/understandingmetadata.pdf. Accessed May 14, 2007. 2. O Carroll PW, Yasnoff WA, Ward ME, Ripp LH, Martin EL, eds. Public Health Informatics and Information Systems. New York; Springer Science + Business Media; 2003. 3. National Cancer Institute Center for Bioinformatics. Cancer Data Standards Repository. http://ncicb.nci.nih.gov/ NCICB/infrastructure/cacore overview/cadsr. Accessed November 12, 2007. 4. US Health Information Knowledgebase. http://www. ushik.org/registry/x/index.html. Accessed November 12, 2007. 5. Arms WY. Digital Libraries. Cambridge: Massachusetts Institute of Technology Press; 2000. 6. Baca M, ed. Introduction to Metadata: Pathways to Digital Information. Version 2. Los Angeles, CA: Getty Information Institute; 2000. http://www.getty.edu/research/conducting research/standards/intrometadata/index.html. Accessed May 12, 2007. 7. Centers for Disease Control and Prevention. National Public Health Tracking Program network implementation. CDC-RFA-EH06-601. http://www.cdc.gov/nceh/tracking/ pdfs/rfp06.pdf. Accessed May 20, 2007. 8. Dublin Core Metadata Organization. The Dublin Core Metadata Initiative page. http://dublincore.org. Accessed May 20, 2007. 9. International Organization for Standardization. ISO/IEC 11179 Document page. http://metadata-standards.org/ 11179/. Accessed May 20, 2007. 10. Federal Geographic Data Committee home page. http://www.fgdc.gov/. Accessed May 20, 2007. 11. Federal Geographic Data Committee. Geospatial Metadata Standards page. http://www.fgdc.gov/metadata/ geospatial-metadata-standards. Accessed May 21, 2007. 12. US Geological Survey. Tkme home page. http://geology. usgs.gov/tools/metadata/tools/doc/tkme.html. Accessed May 20, 2007.

520 Journal of Public Health Management and Practice Appendix 1 Dublin Core Metadata Elements Coverage Subject Rights Description Title Date Type Contributor Format Relation Creator Identifier Source Publisher Language Appendix 2 ISO 11179 Metadata Elements Name Permissible values Permissible values Identifier Related data reference Maximum size Version Type of relationship Minimum size Context Data type Keywords Classification scheme Appendix 3 An example of the FGDC Metadata Elements Data resource title Contact information Status Attributes Purpose Citation Spatial domain Distribution Access constraints Time period of data Resource Keywords Process steps Appendix 4 Element EPHT Metadata Profile Elements Definition Identification elements Citation: Information to be used to reference the dataset Originator The name of an organization or an individual who developed the dataset. If the names of editors or compilers are provided, the name must be followed by (ed.) or (comp.), respectively Publication date The date when the dataset is published or otherwise made available for release Title The name by which the dataset is known URL The name of an on-line computer resource that contains the dataset. Entries should follow the uniform resource locator convention of the Internet (complete if applicable) Description: A characterization of the dataset, including its intended use and limitations Abstract A brief narrative summary of the dataset Purpose A summary of the intentions with which the dataset was developed

Describing Environmental Public Health Data 521 Element EPHT Metadata Profile Elements (Continued ) Definition Supplemental info Other descriptive information about the dataset (complete if applicable) Access constraints Restrictions and legal prerequisites for accessing the dataset. These include any access constraints applied to ensure the protection of privacy or intellectual property and any special restrictions or limitations on obtaining the dataset. Use constraints Restrictions and legal prerequisites for using the dataset after access are granted. These include any use constraints applied to ensure the protection of privacy or intellectual property and any special restrictions or limitations on using the dataset. Native dataset environment A description of the dataset, including the name of the software, computer operating system, file name, and dataset size Time period of content: Time period for which the dataset corresponds to the currentness reference Currentness The basis on which the time period of content information is determined Status: The state of and maintenance information for the dataset Progress The state of a dataset Maintenance and update frequency The frequency that changes are made to the dataset after the initial dataset is completed Spatial domain: The geographic area covered by the dataset West bounding coordinate Western-most coordinate of the limit of coverage expressed in longitude East bounding coordinate Eastern-most coordinate of the limit of coverage expressed in longitude North bounding coordinate Northern-most coordinate of the limit of coverage expressed in latitude South bounding coordinate Southern-most coordinate of the limit of coverage expressed in latitude Keywords: Words or phrases summarizing an aspect of the dataset Theme keyword thesaurus Reference to a formally registered thesaurus or a similar authoritative source of theme keywords Theme keyword Common use word or phrase used to describe the subject of the dataset Place keyword thesaurus Reference to a formally registered thesaurus or a similar authoritative source of place keywords Place keyword The geographic name of a location covered by a dataset (includes city, county, state, state acronym, regional descriptions and references) Security information: Handling restrictions imposed on the dataset because of national security, privacy, or other concerns Security classification system Name of the classification system Security classification Name of the handling restrictions on the dataset Security handling description Additional information about the restrictions on handling Data quality information Logical consistency report An explanation of the fidelity of relationships in the dataset and tests used Completeness report Information about omissions, selection criteria, generalization, definitions used, and other rules used to derive the dataset Process step: Information about a single event Process description An explanation of the event and related parameters or tolerances Process date The date when the event was completed Entity and attribute Overview description: Description of the entities, attributes, attribute values, and related characteristics encoded Entity and attribute overview Detailed summary of the information contained in a dataset Entity and attribute detail citation Reference used to the complete description of the entity types, attributes, and attribute values for the dataset Distribution information Distributor: The party from whom the dataset may be obtained Resource description The identifier by which the distributor knows the dataset Distribution liability Statement of the liability assumed by the distributor Custom order process Description of custom distribution services available and the terms and conditions for obtaining these services Metadata reference Metadata date The date that the metadata were created or last updated Metadata standard name The name of the metadata standard used to document the dataset

522 Journal of Public Health Management and Practice Element EPHT Metadata Profile Elements (Continued ) Definition Metadata access constraints Restrictions and legal prerequisites for accessing the metadata. These include any access constraints applied to ensure the protection of privacy or intellectual property and any special restrictions or limitations on obtaining the metadata Metadata use constraints Restrictions and legal prerequisites for using the metadata after access are granted. These include any metadata use constraints applied to ensure the protection of privacy or intellectual property and any special restrictions or limitations on using the metadata Time period information Single date: Means of encoding a single date and time Calendar date The year (optionally month or month and day) Time of day The hour (and optionally minute, or minute and second) of day Range of dates: Means of encoding a range of dates Beginning date The first year (optionally month or month and day) of the event Beginning time The first hour (and optionally minute, or minute and second) of the event Ending date The last year (and optionally month or month and day) for the event Ending time The first hour (and optionally minute, or minute and second) of the event Contact information Contact information: This section provides a means of identifying individuals and organizations and is used by other sections of the metadata standard Contact organization The name of the organization Contact position Title of the individual (complete if applicable) Contact address: The address for the organization Address type Address type Address Contact address for organization City Contact address city State or province Contact address state or province Postal code Contact address, ZIP or postal code Country Contact address country Contact telephone number The telephone number by which individuals can speak to the organization. Contact TDD/TTY telephone The telephone number by which hearing-impaired individuals can contact the organization (complete if applicable) Contact fax number The telephone number of a facsimile machine of the organization (complete if applicable) Contact e-mail address The address of the electronic mailbox of the organization (complete if applicable) Hours of service Time period when individuals can speak to the organization. (complete if applicable) Contact instructions Supplemental instructions on how or when to contact the organization (complete if applicable)

Describing Environmental Public Health Data 523 APPENDIX FIGURE 1 An Example of a Completed EPHT Metadata Record (continues)

524 Journal of Public Health Management and Practice APPENDIX FIGURE 1 An Example of a Completed EPHT Metadata Record (continued)

Describing Environmental Public Health Data 525 APPENDIX FIGURE 1 An Example of a Completed EPHT Metadata Record (continued)