Industrial application of a credibility framework: two case studies

Abstract The usefulness in industrial applications of a credibility framework for establishing confidence in model outputs in the absence of real-world data has been demonstrated using two case studies from regulated industries. One based on a placement in the team of an engineering company responsible for the safety case for graphite bricks in a nuclear reactor and the second based on interviews with individuals involved in a safety assessment for propylparaben as an ingredient in cosmetics. The previously published framework involves evaluating the possession by the model and its outputs of seven factors related to assumption confirmation, qualitative concordance, quantitative concordance, explanatory power, internal consistency, external consistency and simplicity. It was concluded that possession of all seven factors is unlikely in the absence of data from the real-world and might be impossible in some scenarios. However, it was demonstrated that the process of evaluating whether the factors are present, facilitated by the framework, drives a systematic approach to collation of the available evidence and identification of potential gaps in the evidence, leading to a more informed and structured discussion of safety assessments.