Assessment of Intra- and Inter-observer Measurement Variability in a Radiographic Metacarpophalangeal Joint Osteophytosis Scoring System for the Horse

The study evaluated the intra- and inter-observer measurement variability of an osteophytosis metacarpophalangeal joint scoring system. Ten (n = 10) dorso/palmar, latero/medial, and oblique views of equine metacarpophalangeal joints affected by osteoarthritis were examined. Nine assessment points were graded (scale: 0-3) twice by five veterinary students (inexperienced group, I) and four equine veterinary surgeons (expert group, E). The grades for each of the nine factors were summed to obtain the osteophytosis score. The variability between the two measurements was -2.04 +/- 3.5, 95% CI -3.04 to -1.03 for the I group. For the E group, they were 0 +/- 1.43, 95% CI -0.45 to 0.45. In the evaluation of the same radiographs, the I group had a coefficient of variability (CV) of 37.29%. The correlation was r = 0.90%. The CV between groups was 28.85%. The mean difference between the two observations was -0.03 +/- 0.29 in the E group and 0.22 +/- 0.77 in the I group. The I group showed a greater CV when the score was low (r = -0.78) compared to the E group, where the CV was independent of severity of osteophytosis (r = -0.47). The osteophytosis scoring system is an easily applicable and feasible system to be used by observers with different levels of experience, but inexpert observers may need additional training or may need to be helped by reference images. These data are validated by the low inter- and intra-observer measurement variability results in the E group. Therefore, the scoring system proposed seems to be a repeatable instrument applicable to the radiographic score of the severity of metacarpophalangeal joint osteoarthritis.