EMG biofeedback training, relaxation training, and placebo for the relief of chronic back pain.

24 patients with chronic low back pain were randomly assigned to three treatment conditions: EMG biofeedback, relaxation training, and a placebo condition. Patients were seen for eight sessions and were evaluated before Session 1 and after Session 8. Eight analyses of covariance which were adjusted for age and pretest scores were computed on the final scores to find which variables could detect significant difference between treatments. Age was included as a covariate because the differences in age between conditions were significant. Four variables with significant and nearly significant differences were chosen for analysis. The second set of analyses identified the nature of the differences among the three conditions. These included a priori planned comparisons among conditions, and paired t tests. Relaxation-trained subjects were significantly superior to subjects in the placebo condition, in decreasing pain during the function test, increasing relaxation, and decreasing Upper Trapezius EMG. They were superior to EMG Biofeedback training in increasing reported activity. Both Relaxation and EMG trained subjects were able to reduce Upper Trapezius EMG by Session 8. Relaxation-trained subjects showed significant change on eight of the 14 possible comparisons for each treatment condition. EMG biofeedback training showed significant favorable results in only one condition; the placebo condition showed no significant results. Relaxation training gave better results in reducing EMG and pain, and in increasing relaxation and activity than either EMG biofeedback alone or a placebo condition.