Paradigm Health recently presented at the inaugural SCOPE X conference, sharing real-world lessons from deploying AI-enabled technologies across oncology clinical trial operations at scale.

In the talk, “From Startup to Scale: Lessons Learned from Widespread Deployment of AI-Driven Technology for Cancer Clinical Trial Operations,” Ivy Altomare, MD, Vice President of Clinical Research at Paradigm Health, discussed how large language models (LLMs), workflow integration, and operational infrastructure can improve study planning, patient identification and recruitment, and data capture.

Drawing from deployments across a nationwide community oncology research network, key areas of focus included:

• AI-enabled patient identification and trial matching
• Real-time screening and longitudinal patient tracking
• Harmonization of structured and unstructured clinical data from disparate sources
• Customizable integration across sites and research teams
• Operational lessons from deploying AI technologies at scale

The discussion highlighted how AI can reduce manual screening burden, improve screening precision, and accelerate enrollment timelines when paired with strong clinical workflows, operational integration, and continuous refinement.

Real-World Lessons from Deploying AI in Clinical Trial Recruitment

A central takeaway was that AI performance depends on far more than model selection or prompt quality alone. Paradigm Health shared product development and deployment learnings from implementing LLM-enabled matching and screening technologies across real-world oncology workflows:

• AI can drive meaningful efficiencies across recruitment workflows, but performance depends on continuous refinement
• Prompting for unstructured clinical data and complex inclusion/exclusion criteria requires ongoing iteration and empirical testing
• Performance improves through continuous feedback loops with sites and real-world usage
• Match precision can vary significantly across trials depending on eligibility complexity and data quality

Accurate matching alone, however, does not ensure operational impact. To drive enrollment outcomes, matches must be surfaced at the right moment in the patient journey, integrated directly into clinical and research workflows, and aligned with how sites actually operate. Key considerations included:

• Surfacing matches at key care decision points and ahead of scheduled visits
• Embedding notifications within existing clinical and research workflows
• Supporting deployments with site-specific customization, operational partnership, and continuous refinement

Ultimately, successful AI deployment in clinical research requires not only strong technology, but also deep integration with clinical operations and frontline research workflows.

Considerations for Scaling AI-Enabled Recruitment

The presentation concluded by emphasizing that scalable AI-enabled recruitment requires pairing high-performing technology with deep operational partnership, continuous refinement, and end-to-end infrastructure spanning study planning, patient identification, recruitment, and evidence generation.

Several points resonated strongly with attendees, including:

• High-performing products are only the starting point
• Transparency around product limitations builds trust and accelerates improvement
• Successful deployment requires alignment across both site workflows and sponsor objectives
• Workflow integration and operational support are critical for sustained adoption
• End-to-end solutions drive deeper engagement, stronger partnerships, and more scalable trial operations

Paradigm Health’s participation in the inaugural SCOPE X conference reflects an ongoing commitment to building pragmatic, scalable AI-enabled infrastructure that improves trial access and operational performance across oncology research.