The Constant-Murley score (CMS), one of the most popular clinician-rated measurement tool for shoulder patients, has 4 components: 2 subjective (pain and activity) and 2 observational (mobility/strength). Some studies suggest that patient-related psychological factors may also influence the scoring. The purpose of this study was to measure which components may be related with psychological factors.

Inclusion criteria were patients with chronic shoulder pain (> 3 months) treated in rehabilitation from 01.05.2012 to 30.08.2017. Exclusion criteria were other upper limb injuries, upper limb neuropathies, and inability to complete the questionnaires. At entry, the CMS was performed by trained physiotherapists, and the following questionnaires were completed by patients: Brief Pain Inventory (BPI), Hospital Anxiety and Depression (HADs), Pain Catastrophizing Scale (PCS), Tampa Scale of Kinesiophobia (TSK), and Disability of Arm-Shoulder-Hand (DASH). Correlations between the 4 components of CMS and the questionnaires were measured with the Pearson coefficient (weak correlation: 0.20–0.40, moderate: 0.41–0.60, strong: ≥ 0.61).

In total, 735 patients were included [mean age (sd): 47 (11) years; men: 85%; rotator cuff lesions: 72%]. Median duration of symptoms was 14 months (IQR 9–22). The pain component of the CMS was correlated with disability (DASH: −0.43) and pain (BPI: −0.56), but also with psychological factors: anxiety, depression (HADs: −0.28 and −0.32, respectively), catastrophizing (PCS: −0.45), and kinesiohobia (TSK: −0.25). The activity component was correlated with disability (−0.42) and pain (−0.29). Mobility and strength were only correlated with DASH (−0.40 and −0.33, respectively).

A clinician-rated measurement tool should be independent of patient-related psychological factors, which may limit its validity. This study suggests that the pain component of the CMS should be measured separately to the others in order to reduce the risk of measurement bias.