OBJECTIVES: A cohort Markov model based on disability scores was originally constructed. Due to constraints of a Markov structure and to test structural uncertainty, a discrete event simulation (DES), based on time to event, was subsequently developed. METHODS: The same inputs were used in both models, except that in the DES a cohort of individual patients that reflected the patients from the main fingolimod trials was used and risks of some events were linked to baseline characteristics. For both models, published post hoc clinical data in the HA RRMS subgroup were taken from the pivotal trials for fingolimod and DMF vs placebo. Utility data for each health state and for relapses were used in line with previous similar models. Published costs were inflated to NHS cost year 2013–14 and UK list prices used for both drugs. Possible Patient Access Scheme (PAS) discount scenarios were investigated. RESULTS: In the base case, using list prices, the average probabilistic incremental cost-effectiveness ratio (ICER) for fingolimod vs DMF was found to be £14,076 per QALY using the Markov model (incremental cost: £10,358, QALYs: 0.74) and £11,449 per QALY using the DES (incremental cost: £8,694, QALYs: 0.76), with a 73% and 72% chance of fingolimod being cost-effective at a willingness-to-pay threshold of £30,000/QALY, respectively. Both models were most sensitive to treatment effect on disability progression, but robust to other changes including incorporation of a wide range of PAS estimates. CONCLUSIONS: The Markov and DES models produced similar results, both concluding that fingolimod remains cost-effective in HA RRMS following the introduction of DMF to the UK market. This validates the use of a DES in this situation. DES has greater potential than the Markov model to be easily adapted in the future to deal with changing assumptions on long-term efficacy, treatment sequences and chronic adverse events.