Intra- and inter-observer reliability of Cobb angle measurements in patients with adolescent idiopathic scoliosis.

Abstract

The Cobb angle method is used to determine the severity of scoliosis. Therapeutic decisions for adolescent idiopathic scoliosis (AIS) are guided by the Cobb angle. Therefore, high reliability is crucial. The objective of this study was to determine the intra- and inter-observer reliability of the digital Cobb angle measurements and the definition of end vertebrae in patients with AIS. Moreover, the influence of the observer's medical specialty and experience on Cobb angle measurement was evaluated. Intra- and inter-observer reliability of the digital Cobb angle and end vertebrae is assessed in postero-anterior radiographs of 39 patients with AIS. Measurements were performed blinded and twice by six observers, with a two to 3 week interval. Intra- and inter-observer reliability was analysed by means of intraclass correlation coefficients (ICC). Both intra- and inter-observer reliability analyses resulted in ICC's higher than 0.864 for the Cobb angle and definition of end vertebrae. In addition, for the observer's experience and medical specialty group the inter-observer ICC's were higher than 0.984. The average inter-observer variability for the Cobb angle were 3°, and 1.1-1.6 levels for the cranial and caudal end vertebrae selection. The variability in measured Cobb angle was 1° for the experience group and 2° for the medical specialty group. Cronbach's alpha varied from 0.990 to 0.996. Bland-Altman plots showed moderate variation with a few outliers. The digital Cobb angle measurement as well as the definition of end vertebrae show excellent reliability. According to our results, medical specialty and experience do not affect Cobb angle measurements and definition of end vertebrae.