Absolute, not relative, changes are important when interpreting trial data: Reply