Transforming Clinical Trials with Generative AI: Insights from Medable

In this interview, we speak with Tim Smith, co-founder and CTO at Medable, to discuss AI’s role in addressing DCT challenges, enhancing efficiency, and the future of AI-driven solutions. Tim shares insights on study builds, adoption challenges, and collaboration with industry stakeholders.

Moe Alsumidaie: How does generative AI in Medable Studio address pain points for sponsors and CROs in DCTs, and what barriers remain?

Tim: In Medable Studio, generative AI is designed to provide transparency, control, and significantly reduce complexity in clinical trials. By embedding AI capabilities, we manage the configuration complexities inherent in DCTs. Our architecture is based on standardization, with a common study ontology that is easy for language models to consume. This allows AI to effectively process inputs and simplify study builds, reducing timelines from months to days, and with AI, even to minutes. However, transitioning from one solution to another remains a barrier. We address this with our generative eco capabilities, facilitating seamless library migrations, which is crucial as library builds significantly lift any ecosystem. This approach streamlines processes and enhances the overall efficiency and effectiveness of clinical trials.

Moe Alsumidaie: How do you validate metrics like reducing study builds from days to minutes and achieving high eco adherence?

Tim: We continuously benchmark activities within Studio, both internally and among our customers. Medable has an internal delivery team that has been using Studio since its release. This makes it easy to compare timelines for activities like configuring assessment schedules, creating eCOAs, translating, and so on. By comparing these activities with our extensive historical data, we can clearly see the significant step-function change in speed and efficiency. Additionally, we can gather metrics on Studio usage to benchmark common activities across studies. This allows us to track the day-to-day adoption and usage of Studio.

Moe Alsumidaie: How does Medable AI maintain flexibility for diverse therapeutic areas and study designs?

Tim: AI shines in handling diverse therapeutic areas by leveraging comprehensive datasets. Our expert agents are grounded in relevant data, providing real-time guidance tailored to specific therapeutic areas. For example, in oncology versus neurology trials, AI can offer study design recommendations based on historical data and best practices specific to each field. This allows for informed study design recommendations, drawing from validated historical data, and ensures that the unique requirements of each therapeutic area are met. By doing so, we ensure that our AI solutions are adaptable and flexible, capable of meeting the nuanced needs of various therapeutic areas and study designs.

Moe Alsumidaie: What strategies are you using to accelerate AI adoption among smaller CROs and sponsors?

Tim: AI’s ease of adoption is a key advantage. Our roadmap addresses various industry use cases, focusing on streamlining digital data flow. Intelligent agents can normalize data and interoperate with systems in real-time, making the adoption process more efficient and less resource-intensive. For smaller CROs and sponsors, this means they can leverage AI without needing extensive infrastructure changes, as the intelligence is baked into the platform, allowing for quicker and more seamless integration. This approach accelerates adoption and democratizes access to advanced AI capabilities, enabling smaller players to compete effectively in the clinical trial landscape.

Moe Alsumidaie: How do you plan to overcome skepticism from sponsors, CROs, and regulators?

Tim: Despite complexities, there’s a significant eagerness to adopt AI. We apply rigorous validation, traceability, and referenceability to ensure compliance with industry standards. This transparency allows auditors to understand AI-driven processes, fostering trust and confidence in the technology. For instance, by maintaining a clear map of what happened and why, with human oversight, we ensure that AI processes are transparent and accountable, which is crucial for building trust with stakeholders. By addressing these concerns head-on, we aim to build a robust foundation of trust and confidence in our AI solutions, paving the way for broader adoption across the industry.

Moe Alsumidaie: How do you envision collaborating with stakeholders to standardize AI-driven solutions?

Tim: Collaboration is crucial, and we focus on aligning with industry standards like USDM for digital data flow. Our partnerships, such as with Google, enhance our AI product development, ensuring our solutions meet the needs of both our customers and the broader industry. This collaborative approach is integral to driving adoption and innovation. By working with tech providers and regulatory bodies, we aim to standardize AI-driven solutions, making them more accessible and effective across the industry. Through these efforts, we hope to create a unified framework that supports the seamless integration of AI into clinical trials, benefiting all stakeholders involved.

Moe Alsumidaie: Is there anything else you want to add about Medable AI’s future capabilities?

Tim: We’re excited about upcoming Medable AI capabilities in Medable Studio, including automating the entire study build process and enhancing reporting and translations. These innovations promise to accelerate further and streamline clinical trial processes, making them more efficient and effective. For example, our eCOA capabilities allow for real-time library builds, and our work with translation vendors aims to expedite the translation process, which is often a critical path component in study builds. These advancements highlight the transformative potential of AI in clinical trials, and we are eager to see how they will continue to shape the industry’s future.