Estimating three-dimensional spinal repositioning error: the impact of range, posture, and number of trials

Spinal repositioning sense was tested in normal subjects using a balanced within-subject study design.

The study had three objectives: first, to document the number of trials required to derive a representative value of accuracy and precision in spinal repositioning; second, to document the effects of range on spinal repositioning sense; and finally, to document the effect of different lower limb postures on the repositioning performance.

Joint position sense and kinesthesia play an important role in the control of normal movement of the spine. This has important implications for the diagnosis and assessments of specific movement disorders in individuals with spinal pain syndromes. The literature is varied in methods and results in assessing spinal repositioning sense. For some studies, the inability to determine effects for range or differences between patients with low back pain and normal control subjects may be related to the fact that too few trials were performed to detect a statistical difference.

Twenty-three subjects were tested in standing on a repositioning task for spinal position sense. After a familiarization period, each subject performed 10 matching trials in three ranges (20%, 50%, and 80% of available range) during spinal flexion. The flexion task was performed with knees fully extended, with knees partly flexed, and with the pelvis rotated at 45 degrees to incorporate an asymmetric flexion rotation movement pattern. The three-dimensional coordinates of the repositioning tasks were used to determine accuracy (mean and median of trials) and precision (variable error-standard deviation of trials). The coefficient of variation and statistical power analysis using variables derived from progressively larger numbers of trials were examined. Analysis of variance was used to detect differences for the three ranges and three postures.

After the familiarization period, no learning effect was demonstrated across trials. The coefficient of variation and statistical power for the accuracy and precision tended to stabilize after six trials. Using derived variables from six trials, there was a statistically significant range effect. Accuracy in the inner range was worse than that in the outer range (P < 0.05). There was little evidence of a range effect for precision. Posture had little overall impact. Trunk flexion with the knees flexed improved three-dimensional accuracy in the middle range compared with accuracy with the knees extended and during the flexion rotation task.

It was concluded that increasing the number of trials increases the statistical power and stability of the derived variables. In normal subjects, the accuracy of trunk flexion repositioning improves as one moves further into range.