Mylène Jansen

KIDA performance in severe OA 227 11 While the minimum JSW SDD is comparable between mild and severe OA, the CV shows a large difference, because the severe OA patients show a smaller mean value for minimum JSW. Surprisingly, for minimum JSW, a higher absolute difference between measurements was significantly associated with a higher mean values (and thus less severe OA), although this comparison could have been complicated by the extremely small values, as a result of a truncation effect (1-sided limitation at 0) and limitations with respect to pixel size. Nevertheless, also in these cases, performing the analyses in a short time frame greatly decreases this variability. The SDDs calculated in this research indicate the smallest change that can be interpreted as a real change, as opposed to a measurement error, with p< 0.05. It is important to note that the SDDs described in this research are relevant on an individual level. When using these SDDs on a group level, for example when evaluating groups of patients before and after treatment using KIDA analyses, the group SDD should be calculated by dividing the SDDs calculated here by the square root of the number of observations in the group. 20,21 Apart from intra-observer differences between measurements, it was shown that also in more severe OA, osteophytes and JSW parameters were significantly associated with KL grade. As such, as for mild OA, also for severe OA KIDA is a valid method to evaluate radiographic characteristics of OA. In conclusion, while the variability of some parameters may depend on severity, and without precautions bias may develop, KIDA has been shown to be a useful tool also in patients with severe OA. Its use, like most image analyses techniques, needs to be performed with caution. In order to decrease variability and be able to detect smaller differences, images should be analyzed in a limited time frame and randomized order.