FDA AI Office Faces Test Turning Policy Into Practical Rules
FDA’s decision to establish its new Office of Innovation and Clinical Trial Modernization (OICTM) and form dedicated divisions for artificial intelligence (AI) and real-world evidence (RWE) analytics (see earlier story) could provide drug developers with a clearer regulatory pathway for emerging technologies, but the agency now faces the harder task of translating that structure into practical guidance, according to an analysis by Clinical Trial Vanguard.
Clinical Trial Vanguard chief editor Moe Alsumidaie says the office and division structure suggests FDA increasingly views AI and RWE as interconnected regulatory issues rather than separate disciplines. For example, an AI system that selects clinical trial participants using electronic health records raises questions both about the validity of the algorithm and whether the underlying real-world data are fit for their intended purpose. Housing AI and RWE policy within one office could allow FDA to develop those regulatory frameworks together, he contends.
The challenge, Alsumidaie writes, will be keeping regulatory policy abreast of rapidly changing clinical trial technology. Sponsors are already using AI-assisted monitoring, electronic clinical outcome assessments, wearable devices and other technologies for which regulatory expectations remain incomplete or inconsistent. Among the unresolved questions are how sponsors should establish that an AI model is credible for a particular use, address changes in underlying data over time, maintain human accountability for AI-influenced decisions and document those decisions so FDA reviewers can reconstruct and evaluate them.
A centralized office could also reduce inconsistencies among FDA review divisions by giving sponsors a clearer point of engagement on AI-related questions, Alsumidaie notes.
Whether OICTM ultimately accelerates innovation or creates another regulatory layer will depend heavily on the policies it produces, Alsumidaie concludes. The key test will be whether FDA can establish safeguards for AI use while developing rules quickly enough to keep pace with the technology.