Too Much of the "Red Book" is Still (!) Ahead of its Time
The committee that prepared the National Research Council's (NRC's) report Risk Assessment in the Federal Government: Managing the Process, called the "Red Book," envisioned that ad hoc risk assessments would be supplanted by assessments that relied on "uniform inference guidelines;" that is, on a comprehensive "road map" through the four-step process that explained why particular assumptions or models would be used at each inference point in the absence of data to the contrary. Twenty years later, no Federal agency has yet developed such guidelines. The EPA has come closest, but its much-debated Guidelines for Carcinogen Risk Assessment fall far short of revealing what it regards as the "default" assumption at each inference point (and why!), which alternatives are potentially credible, and what type and quality of information will be necessary to discard the default in favor of a more sophisticated or case-specific alternative. Nor do the EPA guidelines shed any light on the central science-policy question hinted at in the Red Book and discussed in detail in the subsequent 1994 NRC "Blue Book" (NRC 1994)--should the choice among assumptions be dictated solely by opinions about scientific plausibility, or should assessment strive also to avoid errors wherein risks are greatly underestimated? Various excuses have been offered up to defend risk assessments produced without guidelines-- assessments that as a result are not uniform, transparent, cognizant of uncertainty and inter-individual variability, and encouraging to those considering research to develop and validate new approaches--but none of these excuses holds up to more than casual scrutiny.
The other major set of Red Book recommendations clearly warned against institutional separation of risk assessment from risk management. Fear of assessment "polluting" the decision process has led instead to increased isolation of both functions. If anything, the Red Book Committee did not go far enough in warning against the most damaging aspect of organizational separation-the proliferation of analyses that preclude decision-driven risk assessment. To fulfill its potential, risk assessment must reveal the potential costs and benefits of different feasible decision options, in order to discriminate among them; this requires not only full dialogue, but the willingness even to invite assessors in to help scope out the options they will then analyze.inference guidelinesdepartures from defaultsRed Bookmodel uncertaintyoptions analysesassessment/management separation.
The other major set of Red Book recommendations clearly warned against institutional separation of risk assessment from risk management. Fear of assessment "polluting" the decision process has led instead to increased isolation of both functions. If anything, the Red Book Committee did not go far enough in warning against the most damaging aspect of organizational separation-the proliferation of analyses that preclude decision-driven risk assessment. To fulfill its potential, risk assessment must reveal the potential costs and benefits of different feasible decision options, in order to discriminate among them; this requires not only full dialogue, but the willingness even to invite assessors in to help scope out the options they will then analyze.
Keywords: Red Book; assessment/management separation; departures from defaults; inference guidelines; model uncertainty; options analyses
Document Type: Research Article
Affiliations: U.S. Occupational Safety and Health Administration, Denver, Colorado, USA
Publication date: 01 September 2003
- Access Key
- Free content
- Partial Free content
- New content
- Open access content
- Partial Open access content
- Subscribed content
- Partial Subscribed content
- Free trial content