
Predicting MS: Can Data Bring Us Closer to Personalised Care?
Can growing amounts of real-world MS data help clinicians predict disease progression and make more personalised treatment decisions? In this episode of the ECTRIMS Podcast, host Brett Drummond speaks with Dr. Will Brown from Cambridge University and Dr. Carmen Tur from CEMCAT about the rapidly evolving field of predictive modelling in multiple sclerosis. They explore why MS is particularly difficult to predict, from its biological and clinical heterogeneity to the lack of sufficiently sensitive, specific and widely available biomarkers. They also discuss how increasingly rich, long-term datasets are allowing researchers to study disease outcomes, treatment journeys and populations traditionally underrepresented in clinical trials. The conversation examines the methodological challenges that come with real-world data, including missing data, non-random treatment allocation, clinical outcome selection and bias, as well as the opportunities and limitations of AI and machine learning. Finally, they ask what must happen before predictive models can enter routine clinical care — including robust external validation, calibration, interpretability, equitable representation and evidence that using these tools actually improves clinical decisions and long-term patient outcomes. --- This podcast episode is supported by an educational grant from Alexion, AstraZeneca Rare Diseases, Bristol Myers Squibb, Novartis, Roche, Sanofi, and UCB. Educational grant providers have no input into the podcast series content.


















