Multiple sclerosis (MS) is the most common cause of neurological disability in young adults in the developed world. Given the clinical heterogeneity in MS, it is difficult to predict the rate at which an individual will accumulate disability. Natural history studies in MS have provided insight but many questions remain unanswered. MS registries fulfil an important role in providing longitudinal data from multi-centric MS cohorts. The analyses described in this thesis were performed using an international MS database, the MSBase Registry, which currently contains records of over 21,000 people with MS from clinics in 64 countries. Expanded Disability Status Scale (EDSS) scores recorded in the MSBase Registry were used to examine disability progression, by ranking scores at specific disease durations, as used to devise the MS Severity Score (MSSS). I demonstrated strong correlation in EDSS rank over five-year periods, from four years after MS onset, suggesting that the concept of EDSS ranking could be useful in predicting later disease severity. I then sought to determine which clinical features influenced EDSS rank change, observing that factors such as gender, relapse rate, treatment exposure and brain imaging findings predicted this outcome, using quantile regression models. These results may prove useful for clinicians in the design of treatment algorithms and clinical trial endpoints in MS. Using the MSBase Registry, the relationship between pregnancy and MS was examined. I confirmed that pregnancy has a favourable effect on relapses but that relapse rate increases in the. postpartum period, as shown in previous studies. A novel association between pre-conception treatment exposure and postpartum relapses was observed. This could allow neurologists to employ a strategy to minimise risk of postpartum relapses in women with MS who are planning conception. Identifying those people with MS who are at high risk of developing disability is becoming increasingly important given the advent of new MS therapies. It is also vital to improve our understanding of the effect of pregnancy on MS, given that onset is common during childbearing years. With these issues in mind, I performed analyses of longitudinal data from the MSBase Registry, producing findings of clinical interest to neurologists worldwide.