AI/ML current status of clinical trial implementation

Tufts CSDD conducted a global survey from May to August 2024, assessing AI/ML adoption in clinical trials across 36 activities. The study, which included 302 respondents from 79 companies worldwide, evaluated time reduction in 27 activities and investment in 14 activities. Participants, primarily from clinical operations, development, and data management, represented organizations with varying annual trial volumes, from fewer than 25 to over 100 trials per year.

The January/February Impact Report listed activities under the planning and design, execution and regulatory submission domains for the levels of AI/ML adoption, which included the following highlights:

• Identification of diverse populations and RBM/RBQM activities have been partially or fully implemented among 51.6% and 40% respondents respectively
• Analysis of genetic data and patient narratives, 50% and 47.9% respectively
• Trial Master File filing (38.4%) and Clinical Study Report writing at 31.7%
• Further, the TMF filing showed a $2.2 million of investment, which was only topped by data quality control or data cleaning.

The report includes mean reductions of time saved by AI/ML by activity; drivers of successful implementations; as well as challenges of implementation of AI/ML.