Plain language summary:
Much work has discussed how studies that partner with multiple separate databases can address things that create bias for estimates from specific databases. Little attention has been paid, however, to how these studies consider differences in characteristics between the databases and specific groups of people treatment effects may be relevant to (sometimes referred to as “external validity”). This work describes various concepts related to external validity and how they relate to these multi-database studies, discusses how Sentinel, CNODES, and PCORnet (three such database partnerships) address these concepts, and considers how each of their approaches reflects the structure and purpose of that specific partnership.