Reflections on the Red/Mis-Read Book, 20 Years After

Abstract

The National Research Council's (NRC's) report Risk Assessment in the Federal Government: Managing the Process, known as the “Red Book,” had a large impact on the U.S. Environmental Protection Agency (EPA) and other regulatory agencies. In part this impact resulted from EPA Administrator William Ruckelshaus, whose speeches implied that risk assessment following the NRC's paradigm could facilitate better use of science in regulatory decision-making. Unfortunately, Administrator Ruckelshaus and many others who followed him did not read the Red Book as carefully as they might have. The four-element risk assessment process is the most frequently cited aspect of the Red Book. The ten recommendations in the report have received less attention. The “conceptual separation” of risk assessment and risk management called for in Recommendation 1 has been interpreted as organizational separation, with a detailed risk assessment done by risk assessors preceding risk management, rather than an iterative process between risk managers and risk assessors to support timely decision-making in the face of existing uncertainties. The importance, the value judgment aspects, and the evolving nature of “inference guidelines” (now usually called “default assumptions” or simply, “defaults”) imply that risk assessment has an important subjective or policy component. These assumptions are used to bridge uncertainties in the science available to support regulatory decision-making. The Red Book recognized that risk assessment could not be based entirely on objective science, given the pervasive uncertainties. There was a need to make policy judgments and to make these judgments consistently. The report therefore recommended “detailed, comprehensive, and flexible” guidelines. While Federal and EPA guidelines were adopted following the report, there has been only modest progress in refining default assumptions and in taking advantage of flexibility to use improving knowledge of biological mechanisms as the basis for departures from default assumptions. Despite encouragement from subsequent National Research Council reports and from the Presidential/Congressional Commission on Risk Assessment and Risk Management, EPA and other regulatory agencies still use the “Red Book paradigm” in a way that is slow, rigid, and overly fixated on numerical “bright lines.” More progress is needed in following the original recommendations of the Red Book and those of the subsequent reports to create a more flexible and timely risk assessment process to inform regulatory decision-making.