Failure is endemic to clinical trials, with roughly 9 in 10 drug candidates not reaching market. But could virtual patients — vis-à-vis, synthetic data — make for more successful research questions?