Estimation of normal lumbar flexion with surface inclinometry. A comparison of three methods

The purpose of this study was to determine the magnitude and clinical significance of surface measurement error in the determination of lumbar spinal flexion. Intrarater, inter-rater and intermethod reliability estimates were obtained using single inclinometry, double inclinometry and back range-of-motion inclinometry methods. Eight healthy subjects were examined independently by two experienced observers and three replicates of each measurement were obtained by each observer in a random sequence. In addition, three replicates of lumbar flexion angles were obtained for each subject by a single observer using the B-200. Reliability estimates were determined by intraclass correlation coefficients and were further compared by paired t tests between observation series. The median range of error was 8.5 degrees using the single inclinometer, 10.5 degrees using the double inclinometer and 16 degrees using the back range-of-motion. The intrarater reliability was generally higher than inter-rater reliability and intermethod reliability was low in most cases reflecting the poor cross-validity across inclinometry methods and between each inclinometry method and the B-200. In conclusion, significant measurement error in estimating lumbar flexion by inclinometry may be expected to occur even in a "controlled" setting using experienced observers, standard examination techniques and asymptomatic healthy subjects. These findings appear to undermine the expectation that the clinician can reliably apply surface inclinometry to estimate loss of spinal mobility for purposes of impairment determination.