Original paper :AATEX 5(1-2):59-73,1998
In the inter-laboratory validation study organized by JSAAE to evaluate the practicability of five cytotoxicity assays as feasible alternatives to the Draize eye irritation test, we have carried out thorough data cleaning before data analysis in order to be assured of the reliability and correctness of the results of the study, since so many errors on protocol understanding and data recording were found which were assumed to have primarily originated from the participating laboratories. We call these errors "human errors." In this article, various human errors were categorized and analyzed for their possible causes such as lack of description of serum that might have led to the use of mismatched serum types. Based on our experience on data cleaning, we also described proposals with key points that should be taken into consideration at the planning stage of further validation studies to assure the quality of data and to attain reliable results.
Key words: alternatives, cytotoxicity assay, colony formation assay, crystal-violet staining assay, Draize eye irritation test, inter-laboratory validation, LDH release assay, MTT assay, neutral red uptake assay, quality of data.