Risk estimation in rheumatoid arthritis: from bench to bedside

Abstract

The prognosis for patients with rheumatoid arthritis (RA) who were diagnosed in the years since 2010 is much better than for individuals who were diagnosed with the disease 20 years ago. This improvement in the long-term outcome of disease is the result of earlier initiation of therapy, disease-activity-guided modification of treatment and the availability of new, and effective, drugs. Nonetheless, current treatment strategies remain population-based, rather than individualized. Decision-making processes relevant to the provision of individualized treatment require appropriate prognostication with regard to a number of variables. Here, the methods available to evaluate the performance of predictive models are discussed. In addition, I highlight the advances in risk estimation that have been made concerning three treatment decisions relevant to the management of RA that are made daily in the clinic: when to initiate treatment with DMARDs in patients in the early stages of arthritis; the ideal intensity of initial treatment; and the likely responsiveness of the patient to a particular therapy. Apart from a model predicting the development of RA, the majority of prognostic tools derived in arthritis and RA are not accurate or not validated. Hence, personalized treatment decisions in arthritis and RA are still far from bedside.