Laying hens often suffer from keel bone damage (KBD) that includes pathologies with different etiologies, like diverse forms of fractures and deviations. Since KBD is a problem in all countries and housing systems, methods for the assessment of deviations are urgently needed. Comparisons between genetic lines and between studies are important to detect underlying mechanisms. Field researchers often use palpation as a low-cost and feasible technique for the assessment of KBD. In contrast to palpation, radiography is effective and highly precise at least in detecting keel bone fractures. The aim of this study was to: i) develop a scoring system to assess keel bone deviations from radiographs, ii) to assess inter- and intra-observer reliability of this scoring system, and iii) to investigate whether fractures and deviations of the keel are correlated. In total, 192 hens were used for the investigation. Digital radiographs were taken and evaluated for all hens after slaughter. We developed a tagged visual analog scale with two extreme images as anchors and four intermediate tags, resulting in six images representing the range from “no deviation” to “highly deviated” on a 10 cm line. Eleven participants scored 50 radiographs of keels with varying degree of severity, whereas five images were scored twice to assess intra-observer reliability. Intraclass correlation coefficient for inter-observer reliability was 0.979 with a confidence interval of 0.968 < ICC < 0.987 (F49,268 = 54.2, p < 0.0001). Intraclass correlation coefficient for intra-observer reliability was 0.831 with a confidence interval of 0.727 < ICC < 0.898 (F54,55 = 10.8, p < 0.0001). Individual intra-observer reliability ranged from 0.6 to 0.949. The Spearman correlation showed a strong positive correlation of fractures and deviations (sroh= 0.803, p < 0.001). The tagged visual analog scale could be a reliable instrument for the scoring of keel bone deviations. Our results support the assumption that the majority of highly deviated keels suffer from fractures as well. Further research is needed to investigate the correlation of palpation scores with the evaluation on radiographs.
