Characterization of Data Uncertainty and Variability in IRIS Assessments Pre-Pilot VS Pilot/Post-Pilot
This report is a scientific review conducted in support of an EPA evaluation of the characterization of data variability and uncertainty in its Integrated Risk Information System (IRIS) health effects information database. See the full EPA report for complete details about this evaluation: Characterization of Data Variability and Uncertainty: Health Effects Assessments in the Integrated Risk Information System (IRIS) EPA/635/R-00/005F, September 2000 (PDF File 117K).
This report presents the results of a scientific review of uncertainty and variability analysis and documentation in health assessments contained in the Integrated Risk Information System (IRIS). The review was conducted by six individuals with expertise in various areas of toxicology, human health risk assessment, and related fields. The objective was to perform an in-depth study of uncertainty and variability documentation in IRIS assessments, based on sixteen chemicals randomly selected from the 537 assessments currently in the IRIS database. Specifically, the review addressed uncertainty documentation in sixteen assessments that spanning the years 1988 to 2000. The first group of eight were selected from the assessments generated by EPA during the time period 1987 through 1994, also referred to as "pre-Pilot"assessments. The second group of eight was selected from assessments released from 1995 to present, also referred to as "Pilot/post-Pilot assessments." Each IRIS assessment (the summary sheets and the Toxicological Review) was reviewed to determine how uncertainties in the data bases, modeling extrapolation, and other factors were considered in the development of toxicity values. Comparisons between the pre-Pilot and Pilot/post-Pilot groups were made as part of this review. Such a comparison facilitates a better understanding of the evolution in the procedures used to develop IRIS assessments and the extent to which uncertainty and variability have been incorporated into IRIS assessments.