Critical disparities in the diagnostic timeline of MS quantified

In a large retrospective real-world analysis, process mining and deep learning techniques were applied to characterise patient trajectories associated with delayed MS diagnosis. The findings identified substantial disparities in time to diagnosis, driven by gaps in early recognition, insurance-related barriers, and variability in referral and treatment pathways.

Early intervention in MS is essential, as prompt initiation of disease-modifying therapy (DMT) reduces relapse frequency, delays disability progression, and improves long-term neurological outcomes. Dr Arjun Gurjar (University of California, CA, USA) and colleagues therefore sought to quantify critical disparities in the MS diagnostic timeline and determine whether insurance claims could be used to model diagnostic trajectories and identify drivers of delay [1].

The investigators retrospectively analysed a cohort of 447,222 MS patients, using the Komodo Health longitudinal claims dataset, the largest cohort to date analysed for modelling MS healthcare trajectories. An open-source electronic health records (EHR) foundation model was used to generate patient-level embeddings that capture longitudinal healthcare interactions. These embeddings were subsequently clustered to identify subgroups characterised by diagnostic delay, referral sequences, and comorbidity-related confounding.

The analysis demonstrated marked disparities in diagnostic timelines across healthcare entry points. Patients entering the healthcare system through emergency medicine experienced the longest time to MS diagnosis, followed by those referred by general practitioners. In contrast, direct entry via neurology rendered the shortest diagnostic interval.

The authors hypothesised that several mechanisms may contribute to these delays. Misdiagnosis appears to be common, with early MS symptoms frequently attributed to psychiatric, orthopaedic, or rheumatologic conditions. Geographic and structural barriers, described as a “desert phenomenon”, reflecting limited access to specialist care, may further prolong time to diagnosis. Emergency department data suggested additional contributors: patients presenting with demyelination symptoms often did not undergo appropriate MS diagnostic workups, with CT performed instead of MRI scans. Additionally, individuals with prodromal cognitive symptoms were frequently managed within psychiatric channels prior to neurological evaluation.

Together, these findings highlight the need for improved early recognition strategies, streamlined referral pathways, and equitable access to specialist care to reduce diagnostic delays and optimise outcomes in MS.

Gurjar A, et al. A claims-based analysis of 447,222 MS patients: Revealing pre-diagnostic trajectories with process mining and representation learning. P403, ACTRIMS Congress 2026, 4–7 February, San Diego, California, USA.