Minimal clinically important difference, low disease activity state, and patient acceptable symptom state: methodological issues.