A survey of discriminant methods used in toxicological histopathology

A cluster survey of United Kingdom (UK)-based toxicological pathologists has been made of the methods used to discriminate treatment-related from spontaneous variation in histopathological material from toxicity studies. The response rate was 78% of 102 polled. Respondents overwhelmingly preferred methods based on grading or scoring lesions and comparatively avoided methods based on ranking, measuring, counting, pairwise comparisons, or overviewing. This preference was in spite of regarding parametric and ranking methods as both more powerful and specific than some commonly used scoring methods. Respondents were evenly divided on the use of blind examination and tended to avoid using historic material to form an opinion.