A model for human and animal data integration: Weight of evidence strategy.

BACKGROUND Integration of animal and human data to assess potential risks of the use of medications in pregnancy is important. A qualitative weight of evidence process enables all available evidence to be considered in a consistent, systematic manner. METHODS We aim to describe the weight of evidence methodology utilized by the authors, a summary of which was presented at the 59th Annual Meeting of the Teratology Society entitled "Integration of Human and Animal Data to Inform Medication Use in Pregnant Women." The qualitative weight of evidence process evaluates data that inform on a potential relationship between an adverse pregnancy outcome and a medication exposure. An interdisciplinary panel evaluates all available human and animal data related to the question of interest. Study quality assessments of both human and animal data are incorporated. The evaluation assesses gaps in the data from the following areas: (a) strength, (b) specificity, (c) consistency, (d) dose response relationship, (e) methodological considerations, and (f) biological plausibility for the potential association of interest. RESULTS The panel integrates all the information to arrive at an assessment of the evidence and provides recommendations, which may include obtaining more specific information. We provide examples of how the authors apply this process at a pharmaceutical company for evaluation of potential postmarketing safety issues regarding medications and pregnancy outcomes. CONCLUSIONS This weight of evidence method improves the ability to integrate published literature and other data sources to assess the potential risks of medication use in pregnant women and inform future drug safety studies.