The EMA identified ways to address the drawbacks of using electronic healthcare records and real-world evidence to support drug development and regulatory decision-making.

Cloud-based systems and machine-learning algorithms, for example, could give drugmakers and the EMA the capacity to store, access and analyze the massive datasets from electronic healthcare records or other real-world sources.

Missing elements and other noise in the data can be overcome through analysis, corroboration, randomized trials and selectivity, according to an EMA report, which follows a big data workshop held late last year.