Reliability of video motion-analysis systems to measure amplitude and velocity of shoulder elevation

Dynamic shoulder motion can be captured using video capture systems, but reliability has not yet been established.

To compare the reliability of 2 systems in measuring dynamic shoulder kinematics during forward-elevation movements and to determine differences in these kinematics between healthy and injured subjects.

Reliability and cohort.

Research laboratory.

11 healthy subjects and 10 post-superior labrum anteroposterior lesion patients (SLAP).

Contrasting markers were placed at the hip, elbow, and shoulder to represent shoulder elevation and were videotaped in 2 dimensions. Subjects performed 6 repetitions of active elevation (AE) and active assisted elevation of the shoulder, and 3 trials were analyzed using Datapac (comprehensive system) and Dartfish (basic system).

Amplitudes and velocities of the shoulder angle were calculated. Intraclass correlation coefficient (ICC), standard error of measurement (SEM), and levels of agreement (LOA) were used to determine intersystem and intertrial reliability.

For AE, the amplitude maximum (ICC = .98-.99, SEM = 2-3°, LOA = -9° to 5°) and average velocity (ICC = .94-.97, SEM = 1°/s, LOA = -4° to 1°/s) indicated excellent intersystem reliability between systems. Intratrial reliability for minimum velocity was moderate for Datapac (ICC = .64, SEM = 4°/s, LOA = 7°/s) and poor for Dartfish (ICC = .52, SEM = 20°/s, LOA = 37°/s). Cohort results demonstrated for AE a greater amplitude for healthy v SLAP (139° ± 11° v 113° ± 13°; P = .001) and interaction for an average velocity increase of 2°/s in healthy and decrease of 2°/s in SLAP patients over the 3 trials (P = .02).

Reliability ranges provide the means to assess the clinical meaningfulness of results. The cohort differences are supported when the values exceed the ranges of the SEM; hence the amplitude results are meaningful. For dynamic shoulder elevation measured using video, the assessment of velocity was found to produce moderate to good reliability. The results suggest that with these measures subtle changes in both measures may be possible with further investigations.