Optimizing Patient Recruitment in Multiple Sclerosis Clinical Trials Using Real-World Data and AI

Abstract

With increasingly complex protocols, Multiple Sclerosis (MS) has become one of the most operationally demanding indications in clinical development, frequently suffering from 12-to-18-month recruitment delays and high screen-failure rates. This whitepaper explores how purpose-built artificial intelligence and longitudinal Real-World Data (RWD) are actively solving these bottlenecks. Drawing on a 20,000-patient analysis subset from NDAI's broader database of nearly 50,000 longitudinal MS patients, discover how integrating structured and unstructured data from EHRs and claims to MRI reports and biomarkers enables precise patient identification, radically reducing screen failures and compressing enrollment timelines.

What's Inside

• →

  The Root Causes of MS Screen Failures:

  An analysis of the limitations of traditional, site-dependent recruitment and how fragmented data drives up trial costs.

• →

  The 5-Stage AI Recruitment Pipeline:

  A detailed breakdown of how to aggregate, identify, enrich, and refine patient cohorts using a purpose-built AI architecture.

• →

  Unlocking Unstructured Data:

  How specialized LLMs extract critical eligibility evidence directly from free-text clinical notes, CSF results, and MRI reports.

• →

  Strategic Recommendations:

  Actionable steps for integrating RWD at the protocol concept stage to de-risk trial design and build FDA/EMA-compliant external control arms.

• →

  A 20K-Patient Deep Dive, Backed by a 50K-Patient Platform:

  Insights drawn from a 20,000-patient analysis subset of NDAI's broader ~50,000-patient longitudinal MS database, one of the largest real-world MS datasets available representing real-world disease progression and treatment patterns.