J K Jones1, A Krow, S Hariharan, L Weekes. 1. Caribbean Centre for Orthopaedic Surgery, #3 Sixth Avenue, Belleville, St Michael, Barbados. jkjcmj@hotmail.com
Abstract
OBJECTIVE: To introduce the "virtual goniometer", a method of measuring angles on digital images using Microsoft PowerPoint, a readily available and inexpensive software programme. METHODS: Twenty-six X-rays of scoliosis curves were photographed with a digital camera. Six examiners measured the angles of curvature on their computers using the goniometer (Set 1). Under a blinded protocol, repeated measurements on these digitalized X-rays were done three weeks later (Set 2). Intra-observer differences were analyzed. To assess validity, four examiners also measured the angles using the Cobb method. Measurements achieved by both methods were analyzed by the paired samples t-test. To assess inter-observer differences, the Pearson correlation coefficient was calculated. RESULTS: Pearson correlation coefficients were significant, r (24) > or = 0.975, p < 0.001. For intraobserver variability, the average 95% CI range was 2.23 degrees between Set 1 and Set 2. The average 95% CI range was 2.38 degrees for the difference between the digital and Cobb methods. CONCLUSIONS: Clinicians using this technique can reliably assume that repeated measurements of scoliosis curvatures will vary in the range of less than 3 degrees. The 95% CI range for intra-observer variability, an index of the technique's repeatability, was > or = 2.4 degrees. A high correlation of measurements can also be expected between different observers with the goniometer. This new technique allows practitioners to utilize an easily accessible computer programme to evaluate angular deformities on digitalized radiographic images accurately and hence reliably make clinical decisions based on these measurements.
OBJECTIVE: To introduce the "virtual goniometer", a method of measuring angles on digital images using Microsoft PowerPoint, a readily available and inexpensive software programme. METHODS: Twenty-six X-rays of scoliosis curves were photographed with a digital camera. Six examiners measured the angles of curvature on their computers using the goniometer (Set 1). Under a blinded protocol, repeated measurements on these digitalized X-rays were done three weeks later (Set 2). Intra-observer differences were analyzed. To assess validity, four examiners also measured the angles using the Cobb method. Measurements achieved by both methods were analyzed by the paired samples t-test. To assess inter-observer differences, the Pearson correlation coefficient was calculated. RESULTS: Pearson correlation coefficients were significant, r (24) > or = 0.975, p < 0.001. For intraobserver variability, the average 95% CI range was 2.23 degrees between Set 1 and Set 2. The average 95% CI range was 2.38 degrees for the difference between the digital and Cobb methods. CONCLUSIONS: Clinicians using this technique can reliably assume that repeated measurements of scoliosis curvatures will vary in the range of less than 3 degrees. The 95% CI range for intra-observer variability, an index of the technique's repeatability, was > or = 2.4 degrees. A high correlation of measurements can also be expected between different observers with the goniometer. This new technique allows practitioners to utilize an easily accessible computer programme to evaluate angular deformities on digitalized radiographic images accurately and hence reliably make clinical decisions based on these measurements.