Observer agreement study of cervical-vertebral ratios in horses.
Academic Article
Overview
Research
Identity
Additional Document Info
Other
View All
Overview
abstract
REASON FOR PERFORMING STUDY: Different examiners or the same examiner were observed at different times producing slightly different results when obtaining cervical-vertebral ratios. If the difference is substantial, then measurement variability would have an important impact of radiographic interpretation. OBJECTIVE: To quantify agreement and repeatability of these measurements. METHODS: An observer agreement study was performed using 75 horses. Measurements were made at C3-4 and C6-7 by a board-certified radiologist and an imaging resident. Intra- and interobserver agreement was quantified using Bland-Altman plots. Repeatability was assessed as the percentage of differences between duplicate measurements by the radiologist that were within 2 s.d. of the differences. RESULTS: At C3-4, the limits of agreement for the intra-vertebral ratio were between -5 and 4% for the intra- and -5 and 6% for interobserver comparison. For the intervertebral ratio, they were between -9 and 8% for the intra- and -10 and 10% for interobserver comparison. At C6-7, the limits of agreement for the intra-vertebral ratio were between -6 and 5% for the intra- and -6 and 8% for interobserver comparison. For the intervertebral ratio, they were between -7 and 7% for the intra- and -6 and 13% for interobserver comparison. At C3-4, all measurements were 95% repeatable (differences typically 4% and always 8%) for the intra-vertebral ratio and 96% repeatable (differences typically 8% and always 11%) for the intervertebral ratio. At C6-7, all measurements were 98% repeatable (differences typically 6% and always 7%) for the intravertebral ratio and 92% repeatable (differences typically 6% and always 10%) for the intervertebral ratio. CONCLUSIONS: Cervical-vertebral ratios typically varied by 5-10% within and between examiners. POTENTIAL RELEVANCE: When using cervical-vertebral ratios in practice, the impact of measurement agreement should be factored into the interpretation of the test result because measurement variability may lead to misdiagnosis and limit the clinical usefulness of these tests.