A method to compare prospective and historical cohorts to evaluate drug effects. Application to the analysis of early treatment effectiveness of intramuscular interferon-β1a in multiple sclerosis patients

Background: Disease modifying therapy have changed the natural evolution of multiple sclerosis (MS), with efficacy demonstrated in randomized clinical trials. Standard-of-care effectiveness is needed to complement clinical trial data and highlight outcomes in real-world practice, but comparing prospective patients with historical cohorts likely introduces biases. To address these potential biases, assigning a patient with a score that expresses his/her disease prognosis before starting a therapy may make it possible to evaluate the unbiased ability of the therapy to modify disease natural history. Thus, we aimed at analyzing the effectiveness of intramuscular interferon-β1a (im IFN-β1a) matching by BREMSO score (Bayesian Risk Estimate for Multiple Sclerosis at Onset) a prospective real-world cohort of treated patients with a historical cohort of untreated patients. Material and methods: We observed 108 newly diagnosed, treatment naïve MS patients over 12 months of treatment with im IFN-β1a. BREMSO score was used to assign a value to each patient, giving the real-world treated patients comparable with the Historical untreated patients, on the basis of the same risk to have unfavorable evolution. Results: A significantly higher percentage of relapse-free patients is observed in IFN-β1a treated cohort vs. Historical untreated cohort (79.6% vs. 59.3%, p < 0.01). Clinical relapses risk is reduced by 2.2 times in treated patients (p = 0.01). Conclusions: We propose a promising method to manage observational data in a relatively unbiased way, in order to analyze real-world treatment effectiveness.