Text this: Treatment effects in orthopaedic trials are underestimated by applying patient-level PRO thresholds for meaningful differences at the group level.