OBJECTIVES: Expanded Access (EA) programs are primarily intended to help patients with unmet medical need access investigational medicine. Secondarily, data can be collected as there is a moral obligation to track outcomes of patients treated with medicine in development. Hence, real-world data collection in such programs is an area of arising interest. This study assessed the role of real-world evidence derived from data collected in EA programs in regulatory approvals from EMA and FDA. METHODS: We downloaded all application documentation (label, summary and medical review) from the FDA database Drugs@FDA. For the EMA, we extracted all public assessment reports (EPAR), labels and/or scientific discussions. To perform text extraction from scanned documents, we used optical character recognition (OCR). Subsequently, we searched for terms related to EA (‘compassionate use, named-patient, pre-approval access, early access, expanded access’). For all documents where at least one term appeared, we assessed whether data from the associated EA program was used to support clinical efficacy. RESULTS: The study included 22,467 documents, related to 1024 drugs (EMA) and 3031 New Drug Applications (NDAs), (FDA). We found that EMA used efficacy data from EA programs in 29 products and the FDA in 21, a minority of the total number of products with a related EA-term (138). The first EA data were used in 1996, however, the majority stems from the last decade. For 12 products, data were used both by both agencies. In three cases, the EMA and FDA drew different conclusions on the same EA data. For 5 products, evidence from the EA programs was described as ‘pivotal’. CONCLUSIONS: This study provides evidence on the role of data from expanded access programs in regulatory approvals. Although the EA data are relatively seldom used in EMA and FDA decisions, these data play an important role in a select subset of regulatory decisions.