Using Generative AI to Streamline and Automate Clinical Protocol Development

Morteza Maleki, MS previously at the College of Computing Georgia Institute of Technology and now global consultancy PwC investigates whether clinical trial protocol development be revolutionized via the integration of advanced AI technologies such as large language model-based generative AI? The leveraging of generative AI, specifically GPT-4, a recent study involved an attempt to streamline and enhance the efficiency and accuracy of clinical trial protocols. The methodology encompassed a detailed analysis and preparation of comprehensive drug and study level metadata, followed by the deployment of GPT-4 for automated protocol section generation. What happened? The results are compelling, not only demonstrating pragmatic application and benefits of generative AI in clinical trial design, but as well establishing the platform for incremental value-based innovation in this critical field.
Maleki raises some impressive points that caught our attention at TrialSite. In this paper, the author explains the various elements that need to be present to execute on the task for protocol generation, from study and drug metadata collection, clinical trials protocol introduction and study design sections extraction, data preprocessing, prompt engineering and evaluation of the final results.