Comment from Parexel International

AnonymousSupportBusiness
Summary: Parexel International, a global clinical development partner, supports the proposed AI-Enabled Optimization of Early-Phase Clinical Trials Pilot Program. They advocate for including Contract Research Organizations (CROs) as direct participants, prioritizing specific use cases like safety signal detection and dose-escalation, and establishing rigorous pre-specified metrics for evaluating AI performance and decision quality.
Parexel International appreciates the opportunity to respond to the Food and Drug Administration's Request for Information on the proposed AI-Enabled Optimization of Early-Phase Clinical Trials Pilot Program (FDA-2026-N-4390). Parexel operates early-phase trials at scale, currently managing clinical trials globally with 22,000+ experts across regulatory science, biostatistics, clinical operations, and data science. Critically, we serve as the operational partner for many small and mid-sized biotechnology sponsors conducting early-phase IND studies; the very population whose participation is essential to the pilot's success and whose findings would have the greatest practical impact. Our perspective is distinct from that of a drug sponsor or a technology vendor. We do not develop drugs whose programs might benefit from accelerated decisions, and we do not sell AI models. Instead, we integrate AI tools into operational workflows daily and see, on a recurring basis, the gap between an AI capability as it is described in a vendor demonstration and the same capability as it performs against real protocol amendments, real site behaviors, real electronic data capture heterogeneity, and real adverse-event narratives written under time pressure. Because we operate across dozens of sponsors and hundreds of sites, we possess an operational evidence base on AI deployment that no single sponsor can match. The pilot's success will depend on whether its design accounts for that operational reality and explicitly includes CROs as direct participants rather than invisible subcontractors. Our comments are organized to follow the RFI's question numbering in Sections A and B, with particular depth on pilot design, participant selection, AI performance, trustworthiness, and decision quality; the areas where the CRO perspective is most differentiated. We have also included observations on the Agency's Real-Time Clinical Trials (RTCT) initiative, which we understand to be a primary use case of interest within this RFI. We thank FDA for its continued leadership on this important initiative and welcome the opportunity to engage further as the pilot is finalized.

View on Regulations.gov