From data to decision : A case study of controversies in cancer risk assessment

Abstract: Risk assessments serve as the foundation of policy decisions on whether to take measures to reduce a risk or not. However, different risk assessors frequently come to divergent estimates of the magnitude and even the nature of risks. Few attempts has been made in the past to describe and understand the reasons for these differences. This thesis reports the results from a detailed comparison of 30 different cancer risk assessments made of one and the same chemical substance, namely the chlorinated solvent trichloroethylene (CAS no. 79-01-6). The purpose of the present study is to discuss (1) why risk assessors come to different conclusions, (2) how scientific data are used in risk assessment, and (3) how scientific uncertainty is handled in the process. The overall objective is to contribute to increase the transparency and reliability of risk assessments so that they better serve the needs of risk managers and the public. In the first part of this study the different conclusions drawn in these risk assessment documents are identified and described. This is made within the framework of a proposed cancer risk assessment index (CRAI). The CRAI categorization shows that these risk assessors come to divergent conclusions about the trichloroethylene potential to cause cancer. To enable an analysis of the reasons for these differences, detailed information from the trichloroethylene risk assessment documents was stored in a database. This information made it possible to compare the risk assessment documents in terms of data availability (a time dependent factor), data selection, data interpretation, data quality evaluation, and (animal to human) extrapolation of data, and to analyse how these parameters influenced the overall conclusions. The analysis of these data indicates that the differences in conclusions cannot exclusively be explained by an evolving database (data availability). The data sets utilized by the trichloroethylene risk assessors are surprisingly diverse and incomplete, and biased data selection may have influenced some of the risk assessors conclusions. The TCE risk assessors often interpret and evaluate scientific data in different ways. These differences are considered to be within the scope of the scientifically acceptable. In the second part of this case study the European Union regulatory process for classification and labeling served as study object of the risk assessment process in a setting where risk assessors from different affiliations evaluated exactly the same data. This part of the study indicates that there is a scope of possible interpretations of the primary data in relation to the classification criteria and thus that there may be more than one possible alternative for classification of individual substances. The main controversies in this process are also identifed and they are found to concern issues that include policy considerations and thus are not readily resolved by ftirther research. It is concluded that the uncertainty inherent in scientific data opens up a scope of possible interpretations and conclusions and that differences in the assessment and handling of this scientific uncertainty have the potential to influence the overall assessment of risk. It is furthermore concluded that even if an enormous amount of resources were spent on testing and assessment of individual substances (orders of magnitude more than what is required according to existing and proposed regulations), significant uncertainty about their potential to cause harm may still remain.