Spinal range of motion. Accuracy and sources of error with inclinometric measurement.

Abstract

A quantitative construct assessing accuracy and component analysis of sources of error rather than reliability coefficients was tested prospectively in human performance measurements of lumbar spine motion using a cohort of healthy individuals. To evaluate the accuracy of lumbar spine sagittal motion measurements using a computerized inclinometer, which involved progressive analysis of sources of error to identify the most problematic sub-components of the measurement process and device. Many previous studies have described the reliability of inclinometric lumbar motion measurement techniques, but with inconsistent analysis about sources of error to explain identified variability. Similar deficiencies exist in identifying sources of error leading to variability for other human performance measurements (e.g., strength, endurance. lifting capacity, etc.). Yet, range of motion has important clinical applications in monitoring progress and assessing temporary and/or permanent impairment. This makes it especially important for clinicians to be able to recognize and correct factors that limit accurate measurements affecting clinical utility. A computerized inclinometer was used for measuring the sagittal lumbar mobility of 38 healthy individuals after bench testing the device itself for device error. The human performance test conditions were: 1) initial test on study participants by untrained test administrators with no control of human performance or procedural variables, 2) identical tests by procedurally trained test administrators controlling human performance variability by monitoring and controlling total motion, and 3) test by procedurally trained test administrators without controlling for human performance variability. The accuracy of the methodology progressively was degraded by the various sources of error. Device error was negligible relative to error associated with the test process itself. Lack of test administrator training and the magnitude of the measured quantity were the major factors in test degradation. Combined (gross) lumbar flexion was the most accurate measure (worst case > 95% accuracy for overall test conditions), whereas pelvic extension was the least accurate (worst case > 36%). Clinical utility of lumbar spine sagittal motion measurement is highly sensitive to test administrator training to bridge pitfalls to measurement accuracy (bony landmarks, "rocking" of inclinometer on sacrum, etc.). Magnitude of the measurement is another important accuracy factor because absolute error tends to remain relatively constant. Device accuracy is usually an insignificant component of overall test accuracy. Analysis of human performance measurements, such as spinal range of motion, may be facilitated by physics-based assessment of accuracy and procedural error in providing more sophisticated analysis than is customarily accessible through reliability coefficients. Previous studies often failed to recognize correctable procedural errors, rarely addressed them, and almost never quantitated them.