Screening-Level Assessment of the Currentness of U.S. EPA’s IRIS Database. K. C. Osborn, J. B. Welham, B. P. Anderson, J. Conrad; ICF Consulting, 9300 Lee Hwy., Fairfax, VA 22301; and American Chemistry Council, 1300 Wilson Blvd., Arlington, VA 22209
U.S. EPA’s Integrated Risk Information System (IRIS) is a database of the Agency’s consensus positions on potential human health effects from chronic exposure to environmental contaminants. IRIS values have become increasingly central to the practice of risk assessment in the United States. In 1999, the Senate Appropriations Committee raised questions about the extent to which IRIS values reflect "the most current data." A screening-level review of IRIS was performed to estimate the proportion of chemicals in IRIS for which the data cited there do not appear to reflect all current toxicity studies available in the published literature. The approach involved reviewing title, key word, and abstract information to make a preliminary professional judgment as to whether a study could potentially be relevant for updating a chemical’s IRIS assessment. The actual quality and significance of the more current studies were not assessed. A representative sample of 35 chemicals was taken from a subset of the population of 536 IRIS chemicals. (Chemicals updated in 1998 and 1999 (12 chemicals) and chemicals without CAS numbers were excluded from the population.) A literature search for toxicity studies in several publicly available electronic databases was performed in December 1999, and the citations were initially categorized as "appears useful," "possibly useful," and "not useful" for updating the IRIS cancer or non-cancer assessment based on key word criteria. Then, the initial results were reviewed by a senior scientist, who downgraded studies that were unlikely to be useful although they met key word criteria for "appears useful" (e.g., studies of mixtures). The proportion of chemicals in the IRIS subset that had one or more recent citations (i.e., studies published since the most recent IRIS update for that chemical) that "appeared useful" for updating IRIS was estimated to be 91% (95% C.I.: 83% to 97%). The proportion of chemicals in the IRIS subset that had at least five recent citations that "appeared useful" was estimated to be 69% (95% C.I.: 54% to 84%). The mean and median number of total citations (i.e., both cancer- and/or non-cancer-related studies) per chemical that "appeared useful" were 12 and 9, respectively. The median year of last revision for the sample population was 1990.
Go to . . .
2000 SRA Annual Meeting Table of Contents
2000 SRA Annual Meeting Author Index
Main Abstracts Menu Page
RiskWorld Home Page