Excellent work by @benmeg et al. Hope for phase 2. And yes, totally agree: Data heterogeneity issues when "sources allowing free-text input" twitter.com/opentrials/sta…