Visual estimation of shoulder posture: accuracy and reliability across five planes of motion

Background: Visual estimation is still widely used by physiotherapists in clinical practice. The accuracy and reliability of visual estimation are under question and vary with the joint being evaluated. The utility of using visual estimation to evaluate shoulder angular posture is inconclusive and has yet to be evaluated across a comprehensive range of planes and postures.

Objective: The purpose of this observational, cross-sectional study was to determine whether visual estimation is an accurate and reliable procedure for determining shoulder angular postures in a range of shoulder motions.

Methods: Sixty-three physical therapy students viewed digital photographs of an asymptomatic volunteer to visually estimate degrees of arm abduction, flexion, extension, internal rotation, and external rotation. Six photographs from each plane of movement were shown, at positions spanning the available range of motion, and participants were asked to estimate the angular posture. Estimates were averaged, and errors from computer measurements were calculated. Differences between estimates and measurements were determined with t-tests.

Results: The overall accuracy of the visual estimates is considered poor to fair. Root mean square errors were reasonable, from 5° to 17°, but the 95% spread of the data reached a maximum of 53°. Estimates were most accurate at the abduction and flexion postures closest to 90°. Most estimates for extension were underestimated, while more internal rotation estimates were overestimated. Reliability ranged from poor to fair (0.12–0.63).

Conclusions: Using visual estimation to evaluate shoulder planar postures results in a wide range of error. Physiotherapy students should be encouraged to use more objective measurement methods.