The Intelligent Clinical Research Organization: A Conceptual Framework for Artificial Intelligence and Federated Learning in Trial Operations

Abstract

The use of AI (Artificial Intelligence) applications in clinical trials run by CROs (Clinical Research Organizations) has increased significantly, but studies have not yet provided consistent evidence of its effectiveness. This paper discusses AI applications in three key areas: patient selection and screening, genomics turnaround time, and biomarker-based patient stratification. The paper also examines industry analyses of AI implementation in biopharmaceutical development. Some of the findings from the research are: a 34% to 41% decrease in screening time with the help of Natural Language Processing -based eligibility criteria, turnaround times in genomics of several hours because of ultra-rapid sequencing, and AUC results of about 0.65 to 0.809 when predicting the treatment response with the help of AI as opposed to the single biomarker test. Based on these findings, the paper proposes the Intelligent CRO model, which includes NLP-based screening, automated genomic processing, and federated learning.