Maxime Verhoeven

228 Chapter 13 clinical classification (i.e., active or inactive disease based on (adjustments of) DMARD therapy). In general, the performance of DAS-OST was good compared to DAS28 and the rheumatologist’s clinical interpretation, chapter 10 and 11 . In addition, the gender effect, which is seen in OST-scores, plays no role when using DAS-OST in which gender is taken into account. However, the clinical applicability of DAS-OST may be limited as more patients with active disease according to the rheumatologist were incorrectly classified as having LDA, compared to when using DAS28. This indicates that we should optimize the cut-offs of DAS-OST in future research, as we used the established DAS28 cut-offs, since DAS-OST was modelled to reflect DAS28 as reference standard. Perhaps such a new cut-off could be validated in a currently ongoing trial at our department. Before results of the current thesis were present, we started a clinical non-inferiority trial in 2017, to investigate HandScan-guided treatment aimed at ‘HandScan remission’ versus clinically guided treatment aimed at ACR/EULAR 2011 Boolean remission. 28 HandScan remission in this trial is based on a single OST-score, but in case of a continued discrepancy (at three consecutive time points) between HandScan-guided decisions and decisions that would have been taken based on the Boolean criteria, as safety measure in this trial, the Boolean-guided decision will be followed. It would be interesting to perform post-hoc analyses in this trial by calculating DAS-OST and simulate treatment decisions. In this analysis, we can investigate whether other/more appropriate decisions would have been made by applying the established as well as optimised cut-offs for DAS-OST. Moreover, it would be interesting to replace the joint count assessment in more recently developed disease activity indices, e.g., CDAI, by the OST-score. Especially CDAI is of interests as in this measurement an APR is not included, and since CDAI (as well as simplified disease activity index(SDAI)) uses equal weighting for all components. 29 Thus, if a suitable cut-off for DAS-OST can be defined in future research, i.e., if DAS-OST can distinguish between patients having LDA or not, DAS-OST could be implemented in daily practice. Then it may be a faster and more objective alternative measurement compared to the current indices, as a DAS-OST measurement can be performed by a non-healthcare professional. This allows the rheumatologists to focus on patients with active disease (treat-to-target). We conclude that a more promising approach regarding the HandScan than using only the resulting OST-scores seems to be replacement of (time consuming) joint counts by the OST-score in the composite DAS28 measurement, generating a new measurement instrument; DAS-OST. Although this measurement seems to perform reasonably, several challenges have to be overcome before implementation in clinical practice. Until this time, we recommend to (at least in parallel) use the current composite measurements to monitor disease activity.