VARIABILITY OF SCOLIOSIS MEASUREMENT FROM SPINAL ROENTGENOGRAMS

The authors conclude spinal roentgenogram measurement does indeed vary between testers - not only in degree of lateral curvature, but also in identification of curve location. In this study, measurement of 10 roentgenograms by two physicians skilled in taking such measurements revealed a mean absolute difference of 4.2 degrees between the readings of the two testers. Because of these findings, 4 degrees was chosen as the cirterion for assessing whether lateral spinal curvature had changed in our study on the effects of an exercise program on the minimal idiopathic scoliosis curve. Changes of less than 4 degrees were considered to have resulted from measurement variability, while those of 4 degrees or greater were regarded as representing actual change. This criterion was established by a simple procedure that required no additional equipment or facilities and less than 10 man-hours for preparation, testing, and data analysis. The results have also made us more keenly aware of possible variability in measurements for all radiographs and for other types of patient-evaluation tools.