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

Orthopaedic trials frequently rely on patient-reported outcomes (PROs) to measure primary end points. Thresholds for clinically meaningful score differences are then used to interpret PRO scores and support result interpretation. At the patient level, thresholds are used to determine if an individual patient has experienced a clinically meaningful improvement or deterioration, which evaluates whether or not they are a treatment responder. At the group level, thresholds are applied to interpret mean score differences between groups (e.g., trial arms) or between time points and determine if a treatment effect is meaningful. While patient-level thresholds are frequently available for commonly used PROs, interpretation of between-group-level differences is far less established. In the absence of well-defined group-level difference thresholds for PRO scores, patient-level thresholds are frequently used to interpret the difference between groups, such as trial arms. However, meaningful difference thresholds at the patient level are typically larger than relevant differences at the group level. As such, this leads to an underestimation of treatment effects reported in orthopaedic trials.