Season 1 Episode 4: Building a Connected Safety Platform for the Future

This podcast episode, featuring Veeva's David Kološić and Marius Mortensen, principal architect, explores the strategic evolution and future vision of Veeva Safety, focusing on the advantages of building a modern pharmacovigilance (PV) platform on a validated cloud foundation. The discussion traces the origins of Veeva Safety (starting development around 2018) and contrasts the traditional, siloed PV ecosystem—characterized by five-year upgrade cycles, numerous point solutions, and complex integrations—with Veeva’s platform approach. The core argument is that leveraging the established Veeva Vault platform allows developers to bypass foundational concerns (like workflows and basic configuration) and focus immediately on complex safety specifics, such as end-to-end case processing, configurable gateways, and regulatory submissions (E2BR3, VAERS).

A major theme is the paradigm shift from infrequent, massive upgrade projects to continuous improvement via three validated releases per year. This continuous delivery model is essential for keeping pace with rapidly changing global regulations (like the transition to E2BR3 standards) and evolving technology. The speakers emphasize that the platform strategy aims to eliminate the "behemoth of applications" stitched together by IT, which often leads to data inconsistencies and validation headaches. Instead, the unified data model allows for a holistic, worldwide system where regional requirements (like those for Japan) are managed within a single instance, rather than requiring separate, potentially out-of-sync databases.

The conversation heavily details the power of "Veeva Connections," which enable seamless, process-driven integration across different domains (Safety, RIM, Clinical, Medical Affairs). Unlike traditional, "lossy" third-party integrations (which often require manual data reconciliation after an E2B exchange), Veeva Connections focus on specific business use cases—such as literature case intake or updating product labels based on signal detection—by leveraging an "apples to apples" data mapping across the shared platform. This connectivity facilitates cross-domain workflows, allowing PV teams to shift their focus from administrative minutiae to the science of safety, ultimately moving toward predictive analysis of benefit-risk profiles.

Finally, the experts address the role of automation and AI in handling the increasing volume and complexity of adverse event reporting. They advocate for a layered approach: first, ensuring a trusted, high-quality data foundation (the platform); second, automating structured processes (like data exchange and rules execution); and third, applying AI to specific use cases. Mortensen notes that early AI attempts often just shifted the burden to QC by providing non-contextualized data entry. The current focus is on building trust through targeted applications like MedDRA auto-coding and ensuring that AI knows when to "raise its hand" and ask for human intervention, preventing the system from generating untrustworthy or inconsistent outputs. The ultimate vision for 2030 is a "safety golden age" where fragmentation is gone, and PV teams focus on predictive science.

Key Takeaways: • Platform-First PV Strategy: Building PV systems on a robust, validated cloud platform (like Veeva Vault) accelerates development by solving foundational IT challenges (workflows, security, validation) and allowing teams to focus immediately on complex safety-specific requirements. • Continuous Regulatory Compliance: The shift to frequent, validated releases (e.g., three times per year) is crucial for maintaining compliance with rapidly evolving global standards (ICH, E2BR3, FDA) without requiring disruptive, multi-year upgrade projects typical of legacy systems. • Eliminating Data Silos: Traditional PV ecosystems rely on multiple siloed point solutions, leading to complex integrations, data inconsistency, and high IT maintenance stress; a unified platform eliminates these handoffs and provides end-to-end process management. • Veeva Connections for Process Efficiency: Integrations within the Veeva ecosystem are not just technical data exchanges; they are designed around specific business processes (e.g., intake from Medical Affairs, label updates via Regulatory) to enable cross-domain workflows and operational efficiency. • Global System, Single Instance: A modern PV platform must support worldwide operations and regional requirements (like those for Japan) within a single system, preventing data synchronization issues that arise when running separate databases for different geographies. • AI Adoption Requires Trust and Context: Early AI applications often failed by merely shifting the burden of quality control (QC) to human reviewers; successful AI implementation in PV requires contextualized data processing, targeted use cases (like MedDRA auto-coding), and a framework that ensures consistency and reliability. • Automation Precedes AI: The strategy for efficiency involves first automating structured data exchange and rules execution (e.g., auto-filling, sending questionnaires) before layering on advanced AI solutions for complex tasks like narrative generation or complex case intake. • Future Focus on Predictive Safety: By 2030, the goal is for PV teams to move away from the "minutiae of efficiency" and administrative tasks, leveraging automation and high-quality data to focus on the science of safety, risk-benefit analysis, and predictive pharmacovigilance.

Tools/Resources Mentioned:

• Veeva Vault Platform (Foundation for Safety, RIM, Clinical, Quality)
• Veeva Safety
• Veeva SafetyDocs
• Veeva SiteVault
• MedDRA (Medical Dictionary for Regulatory Activities)
• E2BR3/R2 Standards (ICH standards for data exchange)
• FDA VAERS (Vaccine Adverse Event Reporting System)

Key Concepts:

• Veeva Connections: Process-focused integrations between different Veeva applications (e.g., Safety, RIM, Medical) that leverage a unified data model to streamline cross-domain workflows and eliminate manual handoffs.
• Validated Cloud Platform: A system that provides continuous validation and compliance (e.g., GxP, 21 CFR Part 11) through regular, controlled releases, contrasting with the high-risk, infrequent upgrades of on-premise legacy systems.
• Digital Safety Golden Age: The future state (projected by 2030) where PV systems have achieved high-quality, consistent data and seamless connectivity, allowing safety teams to prioritize scientific analysis and predictive risk management.
• Universal Receiver (O Positive): An analogy used to describe the ideal safety system's ability to seamlessly integrate and receive data from diverse internal and external sources (CROs, affiliates, other systems).