Maxime Verhoeven

198 Chapter 10 SUPPLEMENTARY FILE Supplementary Data S1: Optimal cut-off estimation of CRP ≤10mg/L based on ESR Since no data about CRP was available in the dataset of the MMC cohort, it was not possible to calculate directly remission using the Boolean criteria. Therefore, we used ESR to estimate CRP ≤10mg/L, and thereafter this estimate to calculate remission, according to the Boolean criteria. We obtained data including ESR and CRP (collected at the same moment) of 1646 (590 male; 1056 female) RA patients visiting the out-patient- clinic in the Máxima MC from April 2017 to including September 2019, but who were not part of our MMC cohort. We used a linear regression model with a natural log (LN) transformed CRP as outcome and a LN transformed ESR and gender as independent variables. The acute phase criterion of the Boolean remission criteria was defined as LN transformed, estimated CRP ≤10mg/L (i.e., ≤2.302585093). LN (estimated CRP)= -1.08297 + 0.92 * LN(BSE) + 0.55 * male Supplementary Table S1 Established OST-score cut-off per joint using Youden’s index for swelling per joint with diagnostic values in the internal validation cohort. Joint OST-score cut-off* AU-ROC Sensitivity Specificity PPV NPV wrist 1.0 0.55 0.37 0.72 0.13 0.92 MCP1 1.4 0.54 0.20 0.89 0.07 0.96 MCP2 1.0 0.65 0.42 0.87 0.27 0.93 MCP3 0.7 0.66 0.50 0.82 0.15 0.96 MPC4 1.3 0.58 0.23 0.96 0.07 0.99 MCP5 1.5 0.63 0.35 0.91 0.08 0.98 IP1 1.7 0.51 0.00 0.98 0.00 0.99 PIP2 0.9 0.65 0.35 0.90 0.19 0.96 PIP3 0.5 0.60 0.40 0.81 0.15 0.94 PIP4 1.0 0.63 0.39 0.88 0.09 0.98 PIP5 1.2 0.54 0.15 0.93 0.07 0.97 Internal validation cohort consisted of 1120 observations fromMMC cohort, not used in development cohort. *if OST-score for specific joint ≥ cut-off, then that specific joint is estimated as inflamed. AU-ROC= area under the receiver operating characteristic curve; PPV= positive predicted value; NPV= negative predicted value.