#45 – Aaron Mackey, SVP of AI/Data Science at Lokavant & Jonathan Crowther, Head of Predictive Analytics at Pfizer

Introduction

In the latest episode of Bio Breakthroughs, Jared S. Taylor sits down with Aaron Mackey of Lokavant and Jonathan Crowther from Pfizer to delve into the transformative power of causal AI in clinical trials. The conversation explores the potential of causal AI to redefine the pharmaceutical landscape by optimizing clinical trial processes, enhancing decision-making, and minimizing trial delays.

What is Causal AI?

Aaron Mackey begins by clarifying a key term: causal AI, not to be confused with “casual AI.” Causal AI focuses on understanding cause and effect relationships rather than merely generating content or predicting outcomes. Unlike generative AI models like ChatGPT, which analyze vast datasets to predict or create responses, causal AI dives deeper, identifying the root causes and predicting the impacts of specific changes within a system. This depth of understanding is crucial in clinical trials, where outcomes hinge on precise and well-informed adjustments.

Jonathan Crowther elaborates on the practical applications of causal AI in pharmaceutical research and development. He notes how traditional machine learning techniques often result in misleading correlations, which can divert resources and efforts. Causal AI, by contrast, aims to identify confounding factors, ensuring that trial decisions are based on genuine cause-and-effect relationships rather than coincidental data patterns.

Enhancing Clinical Trial Feasibility Assessments

A major theme of the discussion is the impact of causal AI on clinical trial feasibility assessments. Taylor references a report highlighting that 80% of these assessments lack comprehensive data collection mechanisms, leading to inefficient trials and wasted resources. Mackey and Crowther emphasize how causal AI can bridge this gap, providing a more precise and informed approach to feasibility assessments. By leveraging vast datasets and accounting for hidden variables, causal AI allows for a clearer understanding of patient recruitment dynamics, site selection, and protocol development.

Mackey provides a tangible example, discussing how causal AI can differentiate between trial outcomes that result from eligibility criteria and those influenced by external factors like patient registry availability or marketing campaigns. This technology not only predicts outcomes but also offers insights into the specific adjustments needed to achieve desired results, such as speeding up patient recruitment or optimizing trial site selection.

Future of Clinical Trials with Causal AI

Taylor challenges the guests to look ahead and predict how causal AI will reshape clinical trials in the next few years. Crowther envisions a future where clinical trials are increasingly data-driven, with visual tools providing clear and compelling explanations for study strategies. He believes this clarity will enhance collaboration among teams, ensuring that data scientists, strategists, and operational leaders can align their efforts effectively.

Mackey agrees, adding that while causal AI is not new, its modern applications, fueled by advancements in data collection and AI modeling, are expanding its utility. He predicts that within the next two to three years, causal AI will become a mainstream component in clinical operations, helping companies make more informed, efficient, and effective decisions.

Conclusion

As the pharmaceutical industry continues to evolve, the integration of causal AI in clinical trials offers promising potential for streamlining processes, reducing trial times, and enhancing patient outcomes. With experts like Aaron Mackey and Jonathan Crowther at the forefront, causal AI is set to become a vital tool for clinical trial optimization, leading to faster, more effective drug development in the coming years.