Key Variables Ascertainment and Validation in RW Setting

Ascertainment of key variables is a major component in the FDA Real-World Evidence (RWE) framework (FDA. Framework for FDA’s real-world evidence program. https://www.fda.gov/media/120060/download , 2018). Key variables are those that define a key patient characteristic necessary to identify the study population of interest, the treatment/exposure, the outcome, or a key confounder whose effect must be adjusted for. With Real-World Data (RWD), often (1) exposure, outcomes, and key confounders cannot be measured directly, requiring the use of proxies and linkage to external data sources to determine these; (2) no single variable can measure a desired characteristic, requiring characteristics be ascertained through computations or combinations of many data fields using a manually pre-defined rule or computer-generated algorithms; (3) information on the characteristic of interest is contained within another data element such as text field, requiring natural language processing techniques such as information extraction be applied; (4) misclassification is a concern, requiring validation studies. This chapter will cover the above topics and walk through an example study for which the ascertainment of key variables was found to be acceptable from a regulatory standpoint.