A systematic approach for evaluating and scoring human data

Abstract

An approach is described for how the quality of human data can be systematically assessed and categorised. The approach mirrors the animal data quality considerations set out by Klimisch et al., in order that human data quality can be addressed in a complementary manner and to help facilitate transparent (and repeatable) weight of evidence comparisons. Definitions are proposed for the quality and adequacy of data. Quality is differentiated into four categories. A description of how the scheme can be used for evaluating data reliability, especially for use when contributing entries to the IUCLID database, is shown. A discussion of how the criteria might also be used when determining overall data relevance is included. The approach is intended to help harmonise human data evaluation processes worldwide.