What A Good Risk Assessment Contains
Expert panel reports results of one-year study
Though risk assessments never lack for critics, few question their importance in contributing substantively to decisions regarding social policy. It was against that background in 1988 that an ad hoc study group of professionals in risk analysis and risk management formed to conduct a one-year study aimed at improving risk assessment reports and providing risk managers with background for understanding and using assessment reports.
Cosponsors of the study, entitled, "Presentation of Risk Assessments of Carcinogens," were the American Industrial Health Council, U.S. Environmental Protection Agency, Dept. of Health and Human Services, and Society for Risk Analysis. The report, published in July 1989, is a consensus statement of the study group members, but does not represent official positions of their affiliated organizations
The study's authors said that they focused on assessments of carcinogenic risks, but noted that "many of the attributes and conclusions...appear to apply to other types of health risk assessments..."
The study was organized in accordance with a risk assessment framework developed by the National Academy of Sciences, consisting of hazard identification, dose-response evaluation, exposure assessment, and risk characterization.
The ad hoc committee developed what it termed, "General Desired Attributes of a Risk Assessment Report," as follows:
* The scope and objective of the report are explicitly stated.
* The report's content is laid out impartially, with a balanced treatment of the evidence bearing on the conclusion.
* The risk assessment presentation includes a description of any review process that was employed, acknowledging specific review commentary.
* The key findings of the report are highlighted in a concise executive summary.
* The report explains clearly how and why its findings differ from other risk assessment reports on the same topic.
* The report explicitly and fairly conveys scientific uncertainty, including a discussion of research that might clarify the degree of uncertainty.
The authors of the study focused on hazard identification dealing with qualitative assessment of likely carcinogenic hazards. The ad hoc study group noted that data reviewed might include epidemiologic studies, laboratory animal bioassays, pharmacokinetics and metabolism studies, short-term tests, and analysis of the chemical structure. The study group urged assessors to review data early in the assessment to avoid misclassification.
The designated attributes for hazard identification are: the presentation and review of all relevant information; highlighting of critical aspects of data quality; a weight-of-evidence approach to the likelihood of human carcinogenic hazards; and the identification of research permitting a more confident statement about human hazard.
The study's authors point out that dose-response gets sharper scientific scrutiny and generates more debate than the other components of the risk assessment framework. Therefore, the study group advises, "a complete and accurate presentation of this component of a risk assessment is...critical to the perceived integrity of the overall risk assessment..."
The study group spelled out these specific desired attributes of dose-response evaluation:
* Valid data sets and plausible models for high-to-low dose and interspecies extrapolation are presented in dose-response modeling. …