Season 3 Episode 4: AI and Clinical Transformation: High-Value or Hype?

This video provides an in-depth exploration of the role of AI in clinical transformation, featuring Ibrahim Kamstrup-Akkaoui, Vice President for Clinical Data Operations and Insights at Novo Nordisk. The discussion, hosted by Veeva Systems, highlights the accelerating pace of innovation in clinical development and the challenges posed by the exponential growth of data. Kamstrup-Akkaoui shares Novo Nordisk's journey from a historically conservative industry to one embracing digital transformation, emphasizing the critical need for sustainable scaling through technology, particularly AI.

The conversation delves into the evolution of technology adoption in pharmaceuticals, contrasting the slow uptake of early systems like EDC with the current imperative for advanced solutions. A central theme is the shift towards a user-centric approach, acknowledging the burden placed on clinical sites and patients by disparate systems. Kamstrup-Akkaoui advocates for a platform strategy to streamline operations and improve data quality, moving away from the current model where sites may interact with dozens of different systems across multiple sponsors. This foundational change is seen as essential for managing the sheer volume of data, which has grown from millions to billions of data points annually.

Kamstrup-Akkaoui distinguishes between automation and AI, asserting that while automation has its place, AI is indispensable for breaking the traditional proportionality between data volume and the human resources required to manage it. He champions a "think big, start small" philosophy for AI implementation, encouraging experimentation and creative application of technology to solve significant industry challenges. Concrete examples from Novo Nordisk illustrate the practical value of AI, including the generation of meaningful test data for system validation and the development of a "Study Builder" application that digitizes protocol development, standardizes processes, and automates system setup, ultimately aiming for submission-ready data sets even before a study begins. The discussion concludes with a forward-looking vision for a fully automated clinical trial setup, driven by AI and a collaborative industry approach.

Key Takeaways:

• Pharma's Digital Catch-Up: The pharmaceutical industry, historically conservative, is now undergoing a rapid digital transformation to catch up with other sectors. Initial technology adoption, like EDC, was slow due to lack of knowledge and a conservative mindset, but current insights into processes and technology applications are game-changers.
• Exponential Data Growth: Clinical development faces an overwhelming increase in data, with Novo Nordisk experiencing a jump from 5-7 million data points annually to approximately 2 billion. This exponential growth necessitates a fundamental shift in how data is managed and leveraged.
• AI for Sustainable Scaling: To manage this data explosion sustainably, AI is crucial for breaking the traditional proportionality between the volume of data and the number of people required to process it. Automation alone is insufficient for future scaling needs.
• User-Centric Platform Approach: Clinical sites and patients are burdened by interacting with numerous disparate systems (up to 20 per sponsor, potentially 30-40 across multiple sponsors). A platform approach that consolidates systems and prioritizes user experience is vital for improving data quality and efficiency.
• AI is Not Hype: Concrete examples demonstrate that AI is already delivering tangible value in clinical operations, moving beyond mere hype. Organizations should embrace AI as a practical tool for innovation.
• "Think Big, Start Small" with AI: The recommended strategy for AI adoption involves starting with small, creative experiments to understand the technology's potential, then scaling successful initiatives to address larger challenges.
• AI for Test Data Generation: Novo Nordisk successfully implemented an AI algorithm that learns from past studies to generate meaningful test data for system setup, testing, and validation, significantly saving time and improving quality.
• Automated Study Build with Metadata: The "Study Builder" application digitizes protocol development, allowing stakeholders to collaborate in a standardized environment. This process generates rich metadata, enabling the automation of clinical system setup and the creation of submission-ready data sets (including SDTM annotations) early in the trial lifecycle.
• Transforming UAT: Automating the setup and testing processes, particularly through AI-generated test data, allows UAT (User Acceptance Testing) to focus on complex study configurations rather than tedious, repetitive tasks, thereby enhancing overall quality.
• Elevated Role of Data Management: Data management has evolved from a supporting discipline to a key strategic player, gaining recognition for its technical understanding and critical role in digitizing and digitalizing clinical operations.
• Vision for Full Automation: The ultimate goal is a "magic button" that automates the entire setup of clinical trials, allowing teams to focus solely on operating and improving the core functionality.
• Industry Collaboration: Sharing historical data and collaborating across sponsors, organizations, and authorities is essential for gaining deeper insights and accelerating the development of new treatments.
• Prioritizing Patient and Site Perspective: Technology solutions must be built with the patient and site perspective at the forefront, ensuring they genuinely support daily life and processes rather than adding complexity.

Key Concepts:

• EDC (Electronic Data Capture): Early technology for digital data collection in clinical trials.
• UAT (User Acceptance Testing): The process of verifying that a system meets user requirements, often a time-consuming manual task.
• MDR (Metadata Repository): A centralized system for storing and managing metadata, crucial for standardizing and automating system setup.
• SDTM (Study Data Tabulation Model): A standard for organizing and formatting clinical trial data for submission to regulatory authorities like the FDA.
• Clinical Transformation: The comprehensive overhaul of processes, technologies, and organizational structures within clinical development.
• Sustainable Growth: Scaling operations in a way that is efficient, cost-effective, and environmentally/socially responsible, particularly in the face of exponential data growth.

Examples/Case Studies:

• Novo Nordisk's AI for Test Data Generation: A small team developed an algorithm that learns from past study data to generate realistic and meaningful test data for setting up and validating new clinical systems, significantly reducing manual effort and improving quality.
• Novo Nordisk's Study Builder Application: This application facilitates digital protocol development by standardizing language, leveraging catalogs, and enabling collaboration among stakeholders. It generates digital specifications and metadata that can then be used to automate the setup of clinical systems and prepare submission-ready data sets.