I agree about not using these measures on their own, but the reliability of subjective outcomes could be improved by, as @Hutan says, following triallists for long enough. Outcomes at 18 to 24 months ought to be much reliable than at two or three; if a treatment has improved symptoms, patients...