Interrater Reliability of Lumbar Accessory Motion Mobility Testing

BACKGROUND AND PURPOSE: This study examined the interrater agreement, or reliability, of accessory motion mobility testing of the lumbar spine in patients with low back pain. SUBJECTS: Subjects were 18 patients with low back pain referred to the physical therapy outpatient department of a university teaching hospital. METHODS: Six orthopedic physical therapists evaluated the posterior-anterior (P-A) accessory motion mobility at each of six levels, L-1 to the sacral base, on each subject. The mobility was recorded on a nine-point scale, and reproduction of pain was noted. The physical therapists noted any level at which mobility or pain findings were of significance to treat. To evaluate agreement on the identification of spinal levels, therapists were asked to identify one spinous process, which was arbitrarily marked on each subject. Kappa analyses and intraclass correlation coefficients (ICCs) were calculated to evaluate agreement on the level of the marked segment and the mobility at that level, respectively. RESULTS: The ICC for determination of the marked level was R(2,1) = .69 (95% confidence interval = .53-.82). The ICC for mobility findings at the marked level was R(2,1) = .25 (95% confidence interval = 0-.44). A secondary Kappa analysis to determine agreement on treatment decision making demonstrated similarly low levels of agreement. CONCLUSION AND DISCUSSION: There is poor interrater agreement on determination of the segmental level of a marked spinous process. There is poor interrater reliability of P-A accessory mobility testing in the absence of corroborating clinical data. Caution should be exercised when physical therapists make clinical decisions related to the evaluation of motion at a specific spinal level using P-A accessory motion testing.