A FEDERATED ANALYSIS OF CLINICAL REAL-WORLD DATA LINKING TREATMENT PERSISTENCE AND SOCIO-ECONOMIC OUTCOMES IN MULTIPLE SCLEROSIS ACROSS DIFFERENT HEALTHCARE SYSTEMS
Author(s)
Vittorio P. Illiano, PhD1, Markus C. Elze, PhD1, Gabriele Zilorri, PhD1, Annalisa Morgan, MD2, Tina Görbing, Dr3, Noga Orr, Dr4, Erin Longbrake, MD2, Tjalf Ziemssen, MD3, Lawrence Steinman, MD4, Björn Tackenberg, MD1, Matthias Antonin, MSc1;
1F. Hoffmann-La Roche AG, Basel, Switzerland, 2Yale University, Department of Neurology, New Haven, CT, USA, 3Dresden University of Technology, Department of Neurology, Dresden, Germany, 4Stanford University, Department of Neurology and Neurological Sciences, Stanford, CT, USA
1F. Hoffmann-La Roche AG, Basel, Switzerland, 2Yale University, Department of Neurology, New Haven, CT, USA, 3Dresden University of Technology, Department of Neurology, Dresden, Germany, 4Stanford University, Department of Neurology and Neurological Sciences, Stanford, CT, USA
OBJECTIVES: Multiple Sclerosis (MS) is a highly heterogeneous disease of the central nervous system. The therapeutic landscape of MS has evolved significantly in recent years. Treatment persistence is critical for effectiveness and is associated with socio-economic benefits. However, more real-world evidence is needed to understand socio-economic outcomes among people with MS (pwMS), especially when looking at differences between healthcare systems. In this study we applied a Federated Analysis (FA) approach across three sites in two countries. To evaluate the relationship between Healthcare Resource Utilization (HCRU) outcomes and treatment persistence in MS across different healthcare systems by leveraging a federated architecture for real world data analysis.
METHODS: We included pwMS initiating a Disease Modifying Treatment (DMT) between 2018 and 2021 with at least 2 years of follow-up data. Persisters were defined as pwMS having continuous records of a certain class of DMT within the 2-year follow-up period. Clinical and socio-economic outcomes were collected retrospectively, harmonized, and made available for FA. FA keeps individual-level data localized at the institutions and transmits only aggregated data. By decentralizing analytics, FA protects sensitive personal data and has the potential to accelerate data collaborations in the healthcare sector.
RESULTS: We analyzed data of 2528 pwMS across three clinical sites (Dresden University of Technology, Stanford University and Yale University). 76.5% of pwMS were persisters. The likelihood of persistence was positively associated with higher treatment efficacy and conversely, with a lower risk of hospitalization or need for walking aids/wheelchairs, which are significant drivers of HCRU. FA was demonstrated as a reliable approach to simplify multi-site comparisons in the real-world.
CONCLUSIONS: We explored the association of HCRU outcomes with persistence in different healthcare systems through FA. The results will foster the understanding of the influence of treatment approaches on socio-economic outcomes and support the design of future studies on long-term effects on HCRU.
METHODS: We included pwMS initiating a Disease Modifying Treatment (DMT) between 2018 and 2021 with at least 2 years of follow-up data. Persisters were defined as pwMS having continuous records of a certain class of DMT within the 2-year follow-up period. Clinical and socio-economic outcomes were collected retrospectively, harmonized, and made available for FA. FA keeps individual-level data localized at the institutions and transmits only aggregated data. By decentralizing analytics, FA protects sensitive personal data and has the potential to accelerate data collaborations in the healthcare sector.
RESULTS: We analyzed data of 2528 pwMS across three clinical sites (Dresden University of Technology, Stanford University and Yale University). 76.5% of pwMS were persisters. The likelihood of persistence was positively associated with higher treatment efficacy and conversely, with a lower risk of hospitalization or need for walking aids/wheelchairs, which are significant drivers of HCRU. FA was demonstrated as a reliable approach to simplify multi-site comparisons in the real-world.
CONCLUSIONS: We explored the association of HCRU outcomes with persistence in different healthcare systems through FA. The results will foster the understanding of the influence of treatment approaches on socio-economic outcomes and support the design of future studies on long-term effects on HCRU.
Conference/Value in Health Info
2026-05, ISPOR 2026, Philadelphia, PA, USA
Value in Health, Volume 29, Issue S6
Code
PCR132
Topic
Patient-Centered Research
Topic Subcategory
Adherence, Persistence, & Compliance
Disease
SDC: Neurological Disorders