OBJECTIVES: Adherence to cost-effectiveness analytic guidelines requires careful assessment of clinical trial results to ensure the most appropriate cohort data selection.  This study examines the impact of cohort selection on the results of the Goldberg, et.al. 2009 cost-effectiveness study of disease-modifying therapies (DMTs) in multiple sclerosis (MS).   METHODS: Using intent–to-treat (ITT) two-year data from pivotal trials, Goldberg compared cost per relapse avoided of first-line DMTs.  However, there are important differences in the ITT cohorts among the trials.  Due to lower than expected subject drop-out rates, the interferon beta-1a IM (INF-b-1a-IM) trial was able to meet its primary endpoint with a reduced sample size and a decision was made to terminate the trial early.  This resulted in almost half of the 301 ITT (or “all-patient”) cohort being on study drug for less than two years.  A second, “two-year” 172 patient cohort consisted of patients who had completed two years on drug therapy.  Goldberg used the INF-b-1a-IM “all-patient” cohort in his cost study.  To test the impact of this cohort selection, we recreated Goldberg’s model using the “all-patient” relapse rate (0.67 for INF-b-1a-IM vs. 0.82 for placebo) and then substituted the “two-year” cohort data (relapse rate of 0.61 for INF-b-1a-IM vs. 0.90 for placebo) and compared results. RESULTS: This study’s cost per relapse avoided was 45% lower for INF-b-1a-IM while the results for the other DMTs were comparable to those reported by Goldberg.  Ranked from most to least cost-effective, the model results were $77,980 for INF-b-1a-IM, $80,121 for interferon beta-1a SC, $86,572 for interferon beta-1b, and $87,767 for glatiramer. CONCLUSIONS: Our analysis demonstrates the importance of cohort selection in cost-effectiveness analyses in MS.  By selecting comparable cohort data from patients who had completed at least two years on study drug, we found that INF-b-1a-IM had the lowest cost per relapse avoided of all first-line DMTs.