The objective of the work reported in this paper was to critically assess how sequential disease-modifying anti-rheumatic drugs (DMARDs) have been modelled in the context of economic evaluation of the use of DMARDs for treatment of rheumatoid arthritis (RA). A secondary purpose was to identify the methodological challenges of modelling sequential therapies. Systematic searches of 10 databases were undertaken in February 2013. Studies were included if they were in the English language and a full comparative economic evaluation was reported. They were appraised by use of the Drummond checklist (Appendix to this paper). Data extracted included economic evaluation data, data relating to sequential treatment, and data on the modelling methods used. Fifty-seven studies were identified, with 25 (44 %) modelling a sequence of treatments. Forty-three (75 %) were cost-utility analyses. Eleven (19 %) were UK studies and 11 (19 %) were US. The remainder were mainly European (26 (46 %)). A distinction was made between studies of recent-onset RA (14 (25 %)) and those of established RA (42 (74 %)). One study (1 %) was unclear. Individual-level models were more likely to meet the Drummond criteria and evaluate sequences. No study identified an optimum sequence of multiple treatments given a set of treatment options. The level of reporting of the methods and evidence used to assess the effect of downstream treatments in the sequence was generally poor. When lifelong models and downstream treatment sequences were considered, evidence gaps were identified. The review discovered that methods have not been consistently applied, leading to varied estimates of cost-effectiveness. Treatment sequences have not been fully considered and modelled, potentially resulting in inaccurate estimates of cost-effectiveness.