Objective: To investigate the reliability of a radiographic measurement procedure that uses a computer and sonic digitizer to determine projected spinal displacements from an ideal, normal position.
Design: A blind, repeated-measure design was used. Anteroposterior cervicothoracic spine radiographs were presented in random order to each of 3 examiners. Each film was digitized, and the films were randomized for a second examination.
Setting: Private, primary care chiropractic clinic.
Main outcome measures: Intraclass correlation coefficients for intraexaminer and interexaminer reliability for measures on radiographs comparing the perpendicular distance (T(x)) from a vertical axis line drawn through the center of T4 and the center of C2, the linear distance (vertebra(apex)) from the center of the vertebra most displaced from a line connecting the centers of C2 and T4, the angle (Rz) formed by the intersection of the vertical axis line and the upper thoracic line, and the angle of intersection (CDA) between the upper thoracic line and the cervical line.
Results: Intraexaminer reliability for T(x) distance was 0.99 to 1.00, with confidence intervals from 0.98-1.00; for vertebra(apex) was 0.96 to 0.97, with confidence intervals from 0.92-0.98; for Rz was 0.94 to 0.98, with confidence intervals from 0. 89-0.99; and for CDA was 0.92 to 0.95, with confidence intervals from 0.84-0.97. Interexaminer reliabilities for the 3 examiners ranged from 0.97 to 0.99.
Conclusions: Measures similar to those described in this study are commonly used to quantify and categorize spinal displacements from true vertical alignment (i.e., scoliosis measurements). Intraclass correlation coefficient values >0.70 are considered accurate enough for use in clinical and research applications. The measures tested here would fit within these guidelines of reliability. Establishing reliability is an important first step in evaluating these measures so that future studies of validity may be undertaken.