Objective. To survey rheumatologists' preferences for the choice of a second-line disease-modifying antirheumatic drug (DMARD) after inadequate response with methotrexate (MTX) therapy in rheumatoid arthritis (RA).Methods. Thirty-six rheumatologists stated their preferences for RA treatment after inadequate response with MTX therapy (optimal dose at least 6 months). From the initial scenario, we derived 54 vignettes varying by rheumatoid factor or anti-cyclic citrullinated peptide antibody presence, swollen joint count, Disease Activity Score in 28 joints, and structural damage. Respondents stated their preference among 5 therapeutic options: MTX continuation, switch to another conventional DMARD, addition of another conventional DMARD, addition of anakinra, or addition of a tumor necrosis factor (TNF) blocker. Presentation by pairs yielded 10 combinations of strategies for each variant, totaling 540 vignettes; participants evaluated a random sample of 180 vignettes. Determinants of each top-ranked option were analyzed by logistic regression. The compilation of these data served to define a therapeutic algorithm. Results. The responses of 33 rheumatologists were analyzable. Therapeutic preferences corresponded to the top-ranked options. For patients with mild or moderately active RA, either a switch or step-up strategy to another conventional DMARD was top ranked. TNF blockers were preferred for RA patients with high disease activity or progressive structural damage. On the basis of these preferences, we developed a simple decision tree for use in daily clinical practice. Conclusion. Our simple, easy-to-use decision tree developed from rheumatologists' preferences for therapy after failure of MTX therapy in RA treatment may guide rheumatologists in daily practice to choose a second-line DMARD.