OBJECTIVES: Payers in Europe and North America have different policies on coverage of multiple sclerosis (MS) disease-modifying agents (DMAs). With the introduction of Medicare Part D and in the presence of substantial variation in analytical methods used to examine cost-effectiveness (CE) of MS DMAs, an assessment of the models' features and parameters is necessary to understand and interpret the CE results for clinical practice and health policy. This study compares the results of CE models evaluating DMAs (interferon beta-1a, interferon beta-1b, and glatiramer acetate) vs. conventional therapy for treatment of MS. METHODS: Search of electronic databases has identified 8 models. We evaluated the following sources of uncertainty: 1) variation in population characteristics (age, gender, country); 2) sources of data on effectiveness, costs, and health preferences; 3) modeling assumptions (choice and duration of treatment, long-term treatment effectiveness, time of treatment initiation and termination); and 4) model structure (number of health states, study horizon, and modeling software). RESULTS: Results for interferon beta-1a varied from cost-saving to $2,558,660 (2005 US$) per quality adjusted of life year (QALY), CE of interferon beta-1b varied from $10,629/QALY to dominated (more costly and less effective), and results for glatiramer acetate varied from $165,201/QALY to dominated. Time horizon and treatment duration varied from 2 years to lifetime. Studies with longer treatment duration reported worse (higher) CE. All studies used country-specific cost data and performed some sensitivity analyses, but only 4 models were evaluated for uncertainty. CONCLUSIONS: Two out of 8 models found interferons cost-effective, while glatiramer acetate was not CE based on societal standards. The differences in models' results were attributed to the lack of evidence on long-term treatment effectiveness and variation in modeling approaches. Use of DMAs could be justified for selected subpopulations, if prices were reduced, or if more information on long-term treatment effect becomes available.