OBJECTIVES: The advent of new, costly disease-modifying-therapies (DMT) for multiple sclerosis (MS) requires consideration of all evidence on comparative effectiveness with existing therapies to inform both health policy and clinical care.  This study compares the relative effectiveness of DMTs within a mixed treatment comparison (MTC) framework, and explores the contribution of real-world evidence from observational studies to the evidence-base. METHODS: Sixteen randomised controlled trials (RCTs) and four observational studies incorporating nine DMTs were identified according to specified inclusion criteria following a systematic review of the literature.  A Bayesian MTC model was fitted in WinBUGS, for the outcome Annualised Relapse Rate (ARR).  Alternative methods of combining data from different trial designs were used. The model was extended to a meta-regression to include baseline covariates. RESULTS: Natalizumab and Fingolimod, two recently approved DMTs, were significantly more effective than other DMTs for the ARR outcome versus placebo, while the least effective option was Interferon beta-1a 6MIU.  Minimal differences were observed among the other DMTs.  Baseline covariates had no significant impact on the results.  Observational data was available only for the older DMTs.   An MTC of these trials supported the ranking in effectiveness obtained from the RCT MTC, although individual estimates of effectiveness were different to RCT estimates and uncertainty was substantial.  As an alternative to naïve pooling of both RCT and observational data, utilising the observational data as prior information allowed for adjustment of bias due to trial design.  CONCLUSIONS: Relative clinical effectiveness of DMTs is an important component in the assessment of cost-effectiveness of these agents.  RCTs provide the foundation for evaluating comparative effectiveness of DMTs.  However, observational studies can contribute complementary evidence on important comparative effectiveness questions.