Menu Back to Data-in-Clinical-Development

Want to learn more?

If you or your organization want to learn more about DIA’s Research projects or Think Tanks please contact

Contact Us

Data in Clinical Development

Advancing the uses of AI in biopharmaceutical development

To further understand how Artificial Intelligence impacts therapeutic development, DIA is conducting research to demonstrate its potential.

Phase 1. 2017-2018. PharmaTech

In 2018 DIA embarked on Phase 1 of this project with 8 pharmaceutical partners and Tufts University with a goal to further understanding of current and future uses of AI in biopharmaceutical development. The study demonstrated that AI was being utilized in every major function across healthcare, with the highest use occurring in clinical operations functions, followed by pharmacovigilance, safety, and risk management functions.1 (Figure 4) The study also identified that one major hurdle to AI utilization is the lack of validation, which has led to skepticism of AI and hindered widespread adoption.

Figure 4. Phase 1 highlights: Most common uses of AI across Drug Development functions

Figure 4. Phase 1 highlights: Most common uses of AI across Drug Development functions

Phase 2. Currently Launching!

Artificial Intelligence for Adverse Events Prediction: DIA is currently seeking funding partners for this study.

Drug-associated adverse events add hundreds of billions to healthcare expenses every year. The FDA, EMA, and other regulatory agencies are requiring biopharmaceutical companies to implement signal detection systems that identify and analyze adverse events to identify potentially causal relationships between drugs and events. This requires companies to use a monitoring system that is comprehensive and systematic.

In this upcoming study, DIA aims to help further the adoption and implementation of AI in adverse event identification and signal detection in a safety organization. The proposed project aims to:

  • Develop a predictive model for Immune-related adverse events (irAEs) using electronic health records (EHRs) and machine learning.
  • Identify and predict Adverse Drug Reactions (AEs which are causally related to the administration of a drug) using AI/ML methods.
  • Test the viability of AI/ML solutions in the application of signal detection and drug safety monitoring by testing AI/ML algorithm(s) in the ability to extract critical information from source data.

The effort will leverage unique EHR data from MedStar Health and Hackensack Meridian Health as the data source for the study. The Georgetown-Lombardi Comprehensive Cancer Center will be participating as the Academic and Clinical partners in the effort.