Q&A: Assessing the Strides Made So Far in AI’s Clinical Trial Ascent

Mick Ryan, Vice President, IT, ICON plc

Applied Clinical Trials recently spoke with Mick Ryan, vice president of IT at ICON plc, about the emergence of artificial intelligence (AI)-driven approaches in clinical research—from the technology and data analysis component to workflow impact and patient privacy and ethical concerns.

ACT: Where do you think clinical research currently stands with the integration of AI?

Ryan: Collectively, the clinical research industry sees the value of AI and is making strides with AI integration. The industry is keen to adopt tools with promising return on investment and AI’s capacity for handling vast datasets allows researchers to analyze clinical trial data at speeds and depths previously impossible, improving efficiency and accuracy.

Despite progress and growing proof points in support of AI’s advantages, broader adoption still faces barriers, including concerns about data quality, regulatory ambiguity, and interoperability obstacles. Attitudes toward AI adoption remain more cautious than anticipated just a few years ago. The more measured uptake is primarily due to the need for regulatory compliance and reliable data governance. The industry is shifting from experimental use cases and pilot programs to integrating AI-driven tools in routine workflows as a complement to human expertise.

Overall, we see industry interest in AI based in realistic expectations and a growing investment in this space as a potential solution to the industry’s more persistent challenges. As AI matures, its role will likely expand across all phases of clinical development, focused on applications where it will streamline operations and enhance patient-centricity.

ACT: In what areas of clinical research have you seen AI used most effectively?

Ryan: Undoubtedly, AI is most effective in reducing computational burdens on personnel. It can ingest, interrogate, and generate insight from vast datasets much faster than human experts can, freeing up personnel hours and driving efficiency.

Recently, the industry has leveraged AI to accelerate drug discovery with much success. It has also been implemented to reduce risk and tackle challenges at other stages of the development process, for example, during site start-up and feasibility or for post-marketing requirement predictions. Site startup is a crucial stage for trials, and a risky one, offering a prime opportunity for AI to add value. AI uses algorithms and deep datasets to identify eligible patients and the sites most connected to the target patient populations and those most likely to recruit them. By analyzing massive amounts of data to optimize site selection, start-up is accelerated, and trial cost and risk are reduced while diverse trial populations generate more meaningful data for better overall outcomes.

AI also supports real-time data monitoring, improving risk assessment and allowing faster response to adverse events. Predictive analytics, another area of strength, helps forecast trial outcomes and resource needs, enhancing trial design and decision-making.

ACT: What should stakeholders be mindful of when adding AI into their workflows?

Ryan: When implementing AI in clinical research, stakeholders should focus on data quality, ethical considerations, and the ways the organization’s internal infrastructure support the implementation.

High-quality, unbiased data is crucial since AI algorithms rely on accurate inputs to produce reliable insights. Ethical concerns, including patient privacy and avoiding biases in AI models, must be rigorously addressed.

When leveraging a CRO partner’s AI solutions, sponsors should be confident that the CRO’s internal infrastructure is oriented to support scalable, multidisciplinary solutions.

ACT: Clinical research is a highly regulated industry, especially around technology solutions. What challenges does this pose to the successful integration of AI?

Ryan: Strict regulatory standards and the evolving regulations around AI and data integrity and privacy standards do pose challenges for sponsors looking to integrate more AI tools. Compliance with frameworks like GxP, GDPR, and FDA guidance necessitates meticulous oversight of data privacy, quality, and security with AI-driven technologies, which requires strong internal governance and transparency with the interoperable components. For example, regulators are cautious of "black-box" AI models, where decision-making processes are not transparent, which then creates further hurdles for approval. However, the FDA does not yet have finalized regulation on AI, though it has guidance on AI/machine learning as included in the software as a medical device classification without acknowledgement of their use in clinical trials.

ACT: Looking forward five years, where do you hope AI adoption in clinical trials will be?

Ryan: In the next five years, AI has the potential to become more deeply embedded across all stages of clinical development, facilitating data-driven decision-making, enhancing trial efficiency and patient experience while bringing therapies to market faster than ever. A key component of realizing AI’s potential will be its integration within more comprehensive digital ecosystems that enable seamless data interoperability across systems, stakeholders, and solutions throughout development. Large CROs such as ICON have been leading this shift toward a more holistic, integrated approach to implementing innovation, and it will continue to enhance efficiency by facilitating real-time data exchange, improving collaboration and reducing redundant processes.

AI-driven insights will enable precision recruitment, adaptive trial designs, and personalized patient monitoring on a broader scale. More accurate predictive models could support faster, data-driven decision-making, resulting in more effective trials with better risk management. Furthermore, AI's integration into decentralized trials would allow for real-time patient feedback and remote data collection, facilitating more inclusive, diverse trials. As regulatory bodies become more comfortable with AI and sponsors continue to recognize its immediate and long-term value proposition, AI adoption will grow, leading to a more adaptive, patient-centric approach in clinical research.