Collaboration - one control arm, 2+ active drugs from multiple companies (why not 6+?). Fewer patients exposed to placebo, less paperwork, same statistical power, more so if you randomise 2:1:1:1:1:1 rather than a standard trial with 1:1 or 2:1