Use of kinematic algorithms to distinguish people with chronic non-specific low back pain from asymptomatic subjects: A validation study

Abstract

Objective: To determine whether kinematic algorithms can distinguish subjects with chronic non-specific low back pain from asymptomatic subjects and subjects simulating low back pain, during trunk motion tasks.Design: Comparative cohort study.Subjects: A total of 90 subjects composed 3 groups; 45 chronic non-specific low back pain patients in the CLBP group; 45 asymptomatic controls people in the asymptomatic controls group. 20/45 subjects from the asymptomatic controls group composed the CLBP simulators group as well. Method: During performance of 7 standardized trunk motion tasks 6 spinal segments from the kinematic spine model were recorded by 8 infrared cameras. Two logit scores, for range of motion and speed, were used to investigate differences between the groups. Group allocation based on logit scores was also calculated, allowing the assessment of sensitivity and specificity of the algorithms. Results: For the 90 subjects (pooled data), the logit scores for range of motion and speed demonstrated highly significant differences between groups (p<0.001). The logit score means and standard deviation (SD) values in the asymptomatic group (n = 45) and chronic non-specific low back pain group (n = 45), respectively, were -1.6 (SD 2.6) and 2.8 (SD 2.8) for range of motion and -2.6 (SD 2.5) and 1.2 (SD 1.9) for speed. The sensitivity and specificity (n = 90) for logit score for range of motion were 0.80/0.82 and for logit score for speed were 0.80/0.87, respectively. Conclusion: These results support the validity of using 2 movement algorithms, range of motion and speed, to discriminate asymptomatic subjects from those with low back pain. However, people simulating low back pain cannot be distinguished from those with real low back pain using this method.