Science Advisory Board Memorandum on the Integrated Risk Information System and EPA Response
Honorable William K. Kelly
U.S. Environmental Protection Agency
401 M Street, S.W.
Washington, D.C. 20460
Subject: Science Advisory Board's review of the Integrated Risk Information Systems
Dear Mr. Reilly:
The Environmental Health Committee of the Science Advisory Board (SAB) was given a presentation by EPA staff on the Integrated Risk Information System (IRIS) at its meeting on October 26, 1989. The presentation also included discussion of the activities of the Carcinogen Risk Assessment Verification Endeavor (CRAVE) and the RfD (Reference Dose) Review Group.
While it is our understanding that the IRIS was developed primarily for use within EPA, the Committee believes that the IRIS would be of great utility both within EPA and other organizations concerned with the potential health impacts of toxic chemicals in the environment. IRIS has the potential to provide a summary of toxicological data for a large number of chemicals in readily accessible form, either from an EPA on-line computer data bank, from access through existing routes such as the National Library of Medicine's TOXNET, or from
regularly updated computer diskettes distributed to IRIS users. Many state and local regulatory agencies, as well as scientists working in the field of regulatory toxicology, would find IRIS to be a valuable reference source.
The IRIS files contain not only the toxicological data, but also EPA's summary of these data, which may be in the form of the weight-of-evidence characterization for carcinogenicity, unit risk numbers for substances judged to have sufficient evidence for carcinogenicity in animals or humans, and reference dose numbers. This type of information may be widely used both within EPA and by other environmental regulatory agencies as the basis for regulatory decisions. It is therefore very important that the information in IRIS be carefully reviewed for its accuracy, timeliness, and completeness, and that appropriate caveats regarding the data and EPA's evaluation of the data be included in the IRIS files.
We recommend that SAB reviews of Agency documents on specific substances be referenced in the IRIS files for these substances. A short summary of the SAB evaluation of EPA conclusions, especially as to the weight-of-evidence characterization, unit risk, or reference dose, should also be included in the IRIS file, and a short summary of any subsequent communication from the Administrator back to SAB in response to its evaluation.
We understand that Federal Register notices of proposed regulatory actions and final regulatory actions for chemicals in IRIS are now included in the regulatory summaries of IRIS files for those chemicals, a step forward which we commend. In the same vein, major EPA scientific reports such as health advisories, health assessment documents, criteria documents, and Risk Assessment Forum reports should also be cited in IRIS files, and we understand that this will occur in the future. Checks of the files for individual chemicals indicated that IRIS currently lacks citations to some key EPA reports on specific chemicals.
The current computer implementation of IRIS is somewhat cumbersome. For example, capabilities such as returning to earlier text in files or doing searches for specific words or phrases are not available in the current implementation. We understand that the computer implementation of IRIS will be upgraded, and we urge EPA to develop an implementation that is flexible, and ''user friendly" for the spectrum of anticipated users both inside and outside of EPA. EPA should also consider the need for, and potential benefits from, developing more training materials and on-line help capabilities to assist users unfamiliar with IRIS to learn how to use the system. In any such efforts, EPA should remain cognizant that an increase in users should be expected, and the system designed accordingly.
The Agency needs an overall strategy on computerized lists of chemicals,
one which takes into account the differing needs of various segments of the user community. While IRIS may be vary helpful for those wishing to know about the toxicological data, other users may simply wish to know what regulatory actions EPA has taken on a specific chemical, or how to deal with an emergency response in the event of chemical spills. EPA either has or is developing other computerized lists of chemicals, but the planning and coordination among these efforts could be improved. EPA should consider what computerized chemical lists are needed, and, more broadly, how modern computer and telecommunications technology can assist in the processes of risk assessment and risk management for the thousands of chemicals that are of interest to EPA. The Agency should then take steps to assure coordination, cross referencing, and standardization in access procedures for the various computerized lists of chemicals it is, and will be, developing.
The Environmental Health Committee is pleased to have had the opportunity to review IRIS and to offer its advice. We would appreciate your response to the major points we have raised:
We will be pleased to assist the Agency further as it proceeds with the development of IRIS and other computerized chemical lists.
Dr. Raymond Loehr, Chairman
Science Advisory Board
Dr. Arthur Upton, Chairman
Environmental Health Committee
Dr. Raymond Loehr
Science Advisory Board
U.S. Environmental Protection Agency
401 M Street, S.W.
Washington, D.C. 20460
Thank you very much for your letter of March 14, 1990, and your comments on the Environmental Protection Agency's (EPA) Integrated Risk Information System (IRIS). I greatly appreciate, and share, the Science Advisory Board's interest in IRIS and its future.
As you correctly state, IRIS is an important risk information resource both for the Agency, and for other organizations concerned with the potential health impacts of toxic chemicals in the environment. Because the summary risk information on, to date, 397 chemicals represents authoritative EPA consensus positions on the adverse health effects of these chemicals, the Agency is aware of its obligation to the user community to provide system oversight and quality assurance. I share your concerns that the IRIS risk information be as accurate, timely, and complete as possible, and that appropriate discussion and/or caveats be included in the IRIS files.
In your letter you raise several interesting points which I have asked the Office of Health and Environmental Assessment in the Office of Research and Development, responsible for IRIS development and management, to reply. Please see the enclosure for a detailed response.
Again, thank you very much for your letter and your interest in the Agency's IRIS data base. We welcome your comments and appreciate your offer to work with the Agency as it proceeds with future development of IRIS.
William K. Reilly
The Integrated Risk Information System (IRIS) is one of the Agency's major risk information resource tools containing summaries of health risks and EPA regulatory information on, to date, 397 chemicals. Updated monthly, it is used by EPA to provide high quality, timely scientific and technical information to Agency scientists, and promote Agency-wide coordination and consistency of risk assessments. Because IRIS, containing authoritative consensus EPA positions on chemical-specific potential adverse health effects, is used extensively both inside and outside the Agency, we recognize the need to maintain, and improve, the quality of the system, its access and delivery systems, and sufficient oversight. We welcome the Science Advisory Board's (SAB) interest and comments on IRIS and take this opportunity to respond to the major points raised in your March 14 letter.
1. Data Review - As you know, two Agency work groups develop the risk information summaries that appear in IRIS. Each work group is comprised of approximately 20 senior Agency scientists and statisticans from risk assessing program offices, laboratory facilities, and regional offices. During, and subsequent to, the work group deliberations, there are several levels of quality control and internal review built into the IRIS information development process. First, particularly in the case of the Carcinogenic Risk Assessment Verification Endeavor (CRAVE), an emphasis is placed on the use of external/ and or SAB peer-reviewed documents (e.g., Health Assessment Documents, Drinking Water Criteria Documents) to support these summaries and the quantitative risk values they contain. While the Reference Dose (RfD) Work Group process is different, actually developing the oral RfDs for each chemical, they use the same consensus procedures as the CRAVE Work Group. Also, the Oral RfD Work Group methodology has been peer-reviewed and receives SAB oversight.
Second, an extensive technical quality control process is part of each work group's operating procedures. Technical quality control includes internal work group draft Summary review, final Summary review, final check prior to IRIS loading, and a further check after the summary is online. This final consensus summary sheet development is the primary goal of the work groups and reflects the diligence and hard work of the group Chairs and members.
Third, an editorial quality check is conducted prior to loading on the System. This check, performed by a contractor, is being done on all chemical files currently on IRIS and on new files before they go online. It includes an edit for clarity, style, continuity, and typographical errors.
Finally, since 1986 when IRIS was made available to the Agency, and 1988 when IRIS was made available to the public, its use has grown far beyond earlier expectations. We acknowledge that additional oversight of the system is war-
ranted. To that end, EPA's Risk Assessment Council, which is chaired by Deputy Administrator F. Henry Habicht II, has established a subcommittee for IRIS. This subcommittee, chaired by Dr. William H. Farland, Director, Office of Health and Environmental Assessment, Office of Research and Development, will address both generic and chemical-specific issues concerning IRIS and its associated work groups. Also, IRIS status will be an agenda item at each Council meeting.
These various levels of review and oversight help to assure that IRIS remains an important resource toll and that the quality and validity of the information continues to improve.
2. SAB Evaluations - Preliminary discussions with SAB Director Dr. Donald Barns regarding addition to IRIS of short summaries that could include SAB evaluation of and comments on the principal EPA documents that support the CRAVE and RfD findings, have taken place. The inclusion of the SAB information would underscore the argument that while individual IRIS summaries are not peer-reviewed, the reports and documents on which the summaries are based have received external review. The process and management details on how to accomplish this task will be worked out in conference with Dr. Barnes, Dr. Farland, and the IRIS staff.
3. EPA Reports - Only citations for EPA scientific reports and other references used in developing the RfD and/or CRAVE summaries are included in IRIS. Full bibliographies listing those references are currently being prepared and loaded on the System. Thus far, bibliographies for 251 chemicals are online, with 146 to go. Once the addition of all bibliographies are complete, a user will have citations for all reports, studies, and documents used by the two work groups. Also, summaries of Drinking Water Health Advisories are included on IRIS in Section III: HEALTH HAZARD ASSESSMENT FOR VARIED EXPOSURE DURATIONS. A backlog of the Drinking Water summaries currently exists. The IRIS staff is in the process of putting them on the system.
When IRIS was initially developed in 1986, EPA Regulatory Actions (Section IV) were part of the system. These regulatory action sections provide information, including applicable Federal Register citations, for the Clean Air Act, Safe Drinking Water Act, Clean Water Act, Federal Insecticide, Rodenticide and Fungicide Act, Toxic Substances Control Act, Resource Conservation and Recovery Act and the Superfund Reauthorization Act. Because this regulatory information is subject to change, we are aware that this section needs to be carefully reexamined to insure that it is up-to-date and complete. Working with the Risk Assessment Council's IRIS Subcommittee, the IRIS staff is in the pro-
cess of developing a proposal to review and update the present regulatory action section. This work should commence in the near future.
4. Delivery Systems - Currently, Agency scientists and SAB members access IRIS using the EPA Electronic Mail system (EMAIL). IRIS on EMAIL is slow, cumbersome, and offers little or not reporting capabilities. In 1986, IRIS was a new Agency resource tool containing both qualitative and quantitative risk assessment information. The EMAIL delivery system, by design, obliged users to look at the whole chemical file, not just selected small sections, thus providing a wider chemical profile. At that time, there was concern that only the quantitative risk values would be accessed and not the qualitative discussion of the underlying studies, reports, assumptions, and limitations which is critical in evaluating and understanding the derivation of the risk values. As risk assessment methodologies have become more sophisticated, so too have IRIS users become more experienced and sophisticated in interpreting, evaluating, and using the IRIS risk information. Therefore, the time is right to provide them with a greatly enhanced delivery system that is fast, flexible, interactive, and user friendly.
On March 5, 1990, IRIS became available on the National Library of Medicine's (NLM) Toxicology Network (TOXNET). TOXNET is an online system that is highly regarded and easily accessed. IRIS on TOXNET provides many of the sophisticated functions requested by users. For more information on TOXNET, please refer to the enclosed NLM IRIS Fact Sheet.
Also, a Personal Computer (PC) based version of IRIS is being developed. The PC delivery system will provide the user with sophisticated user capabilities including easy movement within files, reliable keyword and string searches, reporting options, and a fast, accurate, and easily accessible system. We anticipate its availability in early 1991.
5. Training Materials - Your comments on the need for more and better developed IRIS training materials and online help are correct. The current user guide was inadequate for the users needs and did not provide clear, concise, and complete instructions. A revised user guide has been completed and the final version will be available both online and in paper copy by the end of May 1990. Also, development of new online help and other training materials are under consideration, including a revised case study, fact sheets, and interactive demonstration diskettes.
Training has been an important part of IRIS from its inception. A large training program both at Headquarters and in the Regions accompanied IRIS's availability in 1986. Presently, each Region has its own IRIS coordinator who
conducts training as needed, and the IRIS staff conducts workshops and seminars both inside and outside the Agency on a regular basis. A joint symposium, sponsored by EPA and the Chemical Manufacturers Association, on IRIS and some of its underlying risk assessment methodologies is being considered as another opportunity to stress appropriate use of the system. Also, finalization and distribution of the PC version of IRIS will result in another round of intensive Agency-wide user training.
6. List Coordination - The Agency recently took a major step forward in coordinating and cross-referencing regulatory and regulation-like lists, by approving development of a pointer system that will contain references to all chemicals and other pollutants regulated by EPA and to all the lists on which each chemical or pollutant occurs. This system, tentatively called the Registry of Lists, is currently under development; a prototype should be built during this calendar year, and the system should be generally available in one to two years. It will be designed as a pointer system, telling users where other information is available, because each individual list has been compiled for different programmatic reasons, and there is generally not a uniform set of data elements across the lists. IRIS chemicals will be referred to explicitly in the Registry of Lists, and IRIS and the Registry will be compatible to ensure that IRIS users can get complete cross-reference information.
If you have further questions or comments regarding any of the responses included above, please contact Linda Tuxen, EPA IRIS Coordinator, at 202-382-5949 (FTS 382-5949).