Abstract
Background: A wide range of outcome measures is available for the assessment of individuals with shoulder disorders. Since outcome measures are used to orientate treatment decision-making, clinicians must make sure that the ones they intend to use present adequate properties for their clinical purpose.
Objectives: To review the psychometric evidence relating to five categories of shoulder outcome measures: range of motion, scapular position and motion, strength, functional performance, and self-report questionnaires. Validity, reliability, and responsiveness were reviewed for each category.
Major findings: For most of the shoulder outcomes reviewed, evidence relating to relative reliability (intraclass correlation coefficients) and construct validity was favorable. Globally, intrarater and test–retest reliability indices were excellent, while interrater indices were good to excellent. Favorable evidence on responsiveness (effect size, standardized response mean, clinically important difference) and absolute reliability (minimal detectable change) was also found for self-report questionnaires. However, for range of motion, scapular position and motion, strength and functional performance tests, evidence on responsiveness and absolute reliability was limited or non-existent.
Conclusions: Self-reported scales, such as the DASH, ASES, WORC, or WOSI, should be first line tests as they are highly responsive. Furthermore, they are the only shoulder outcomes for which most useful clinical benchmarks are established, such as minimal detectable change or clinically important difference. As for more conventional measures such as range of motion and strength, while still clinically relevant, they require further investigations to define their responsiveness, minimal detectable change, or clinically important difference.