Background: The Multiple Sclerosis Severity Score (MSSS) is a widely used measure of the disability progression rate. However, the global MSSS may not be the best basis for comparison between all patient groups. Objective: We evaluated sex-specific and onset phenotype-specific MSSS matrices to determine if they were more effective than the global MSSS as a basis for comparison within these subsets. Methods: Using a large international dataset of multiple sclerosis (MS) patient records and the original MSSS algorithm, we constructed global, sex-specific and onset phenotype-specific MSSS matrices. We compared matrices using permutation analysis. Results: Our final dataset included 30,203 MS cases, with 28.9% males and 6.5% progressive-onset cases. Our global MSSS matrix did not differ from previously published data (p > 0.05). The progressive-onset-specific matrix differed significantly from the relapsing-onset-specific matrix (p <0.001), with lower MSSS attributed to cases with the same Expanded Disability Status Score (EDSS) and disease duration. When evaluated with a simulation, using an onset-specific MSSS improved statistical power in mixed cohorts. There were no significant differences by sex. Conclusion: The differences in the disability accrual rate between progressive- and relapsing-onset MS have a significant effect on MSSS. An onset-specific MSSS should be used when comparing the rate of disability progression among progressive-onset cases and for mixed cohorts.