Taking Action on AI in Clinical Trials
Contributing Expert: Mike Stocks, Chief Technology Officer at Medrio
Artificial intelligence (AI) is steadily advancing toward widespread adoption across many industries, including clinical trials. Its potential to revolutionize processes, from protocol design to data analysis, has sparked significant interest.
Yet, in the cautious culture of clinical research, there is a lot more talk than action.
This measured pace is understandable since the stakes are high. But how do we shift from theoretical conversation to practical application?
Learn from Other Regulated Industries
For decades, financial technology (fintech) has used AI to enhance credit assessments with alternative data and prevent fraud by monitoring transaction patterns in real time. AI also powers algorithmic trading, automates investment strategies, and streamlines compliance with complex regulations.
Meanwhile, in insurance, AI customizes underwriting by analyzing risk profiles and improves customer retention through data-driven insights.
In national defense, AI is widely used to enhance anomaly detection, enrich decision-making, strengthen cybersecurity, and improve autonomous systems.
The clinical trial industry should draw inspiration from how other highly regulated industries have adopted AI. Just as these industries found meaningful use cases for AI, we need to take advantage of realistic opportunities to leverage AI in clinical research.
Embrace a Human-in-the-Loop Approach
Humans generally trust their own judgment more than machines. That is why it's critical to adopt a human-in-the-loop (HITL) approach to AI in clinical research.
The HITL approach blends AI with human expertise. A responsible person or team oversees AI's outputs, interprets insights, and makes necessary adjustments. By using AI to identify anomalies quickly, researchers remain empowered to make the final decision and determine the appropriate action.
Define clear checkpoints
Implementing a HITL approach in clinical research involves defining clear checkpoints where human review is essential, such as during protocol design or risk assessment. Providing the right tools and support is imperative, as it helps users quickly adjust results and make decisions. The system then uses that feedback to learn and improve.
Recommended by LinkedIn
For example, clinicians or data managers can periodically assess AI-generated insights, validate data outputs, and fine-tune algorithms as trial data evolves. Then, AI can use that input to improve and adjust future results.
The benefits of HITL
HITL ensures that AI-driven decisions or recommendations in clinical trials are accurate, relevant, and ethically sound. Combining AI's analytical power with human judgment maximizes both efficiency and safety. It keeps the human element front and center where it belongs.
The HITL approach also:
Using a HITL approach for AI in clinical trials makes building stronger relationships between clinicians and technology easier. This approach also creates solid foundational patterns of learning and growing. Supervised learning in the form of HITL is a better approach for clinical trials due to the risks associated with a bad decision.
Lean on Experts to Find Balance
Navigating the infancy of AI in clinical research requires a responsible and measured approach. As an industry, our focus should be on proven outcomes, regulatory compliance, and ethical responsibility.
We also need to find the right pace. Move too quickly, and you could disrupt a solid trial with unexpected challenges. Move too slowly, and you risk falling behind on innovation and efficiency.
It’s never been more important to lean on the right experts. Medrio’s AI Incubator Lab is tackling today’s challenges head-on. We’re inviting sponsors, CROs, and industry experts to shape how AI can streamline data management for more efficient trials.
Connect with Medrio’s product leadership team, share insights to shape our AI product strategy, and test new solutions to meet real-world needs.