A new strategy for ranking chemical hazards. Framework and application

This paper outlines a strategy, based on statistical design, for assessing and ranking the hazards of chemicals for which little or no useful data on toxic effects are available. Missing data of series of chemicals are predicted on the basis of test data generated for a minimum number of specific compounds that are adequate representatives of the relevant series: such critical compounds are identified by means of a factorial design. Accordingly, the aquatic toxicities of 100 monosubstituted benzenes were predicted from the experimental values of 8 selected representatives. The reliability of the predictions was checked by comparing the predicted and experimental toxicities of six additional compounds in the series.