The added value of predictive biomarkers in treat-to-target strategies for rheumatoid arthritis patients: a conceptual modelling study

Abstract

Abstract Objectives To quantify the additional value of a hypothetical biomarker predicting response to treatment for RA regarding efficacy and costs by using a modelling design. Methods A Markov model was built comparing a usual care T2T strategy with a biomarker-steered strategy for RA patients starting biologic therapy. Outcome measures include time spent in remission or low disease activity (LDA) and costs. Four additional scenario analyses were performed by varying biomarker or clinical care characteristics: (i) costs of the biomarker; (ii) sensitivity and specificity of the biomarker; (iii) proportion of eligible patients tapering; and (iv) medication costs. Results In the base model, patients spent 2.9 months extra in LDA or remission in the biomarker strategy compared with usual care T2T over 48 months. Total costs were €43 301 and €42 568 for, respectively, the usual care and biomarker strategy, and treatment costs accounted for 91% of total costs in both scenarios. Cost savings were driven due to patients in the biomarker strategy experiencing remission or LDA earlier, and starting tapering sooner. Cost-effectiveness was not so much driven by costs or test characteristics of the biomarker (scenario 1/2), but rather by the level of early and proactive tapering and drug costs (scenarios 3/4). Conclusions The use of a biomarker for prediction of response to b/tsDMARD treatment in RA can be of added value to current treat-to-target clinical care. However, gains in efficacy are modest and cost gains are depending on a combination of early proactive tapering and high medication costs.