Safety, efficacy, speed and costs must all be prioritized and balanced in the delivery of life-changing therapies to patients. A drug that's quickly and cost-efficiently delivered to market, but isn’t effective and safe is unacceptable. An effective, safe drug that doesn’t get to patients in time to save lives has failed those who needed it most. When it comes to patient health and safety, there can be no compromises.

Fortunately, in a world with abundant data and advanced analytics, we have more tools than ever before to optimize this balance for the betterment of patient safety and outcomes.

In this week’s DIA 2021 Exclusive, SAS Principal Solution Architect Tom Sabo will join a panel of leaders from the pharmaceutical industry and the FDA to discuss "Artificial intelligence: Real applications and regulatory perspectives."

Deep learning and machine learning in action

Sabo's expertise is in using deep learning and machine learning to improve efficiency in identifying drug safety signals in adverse event (AE) narratives. Patient narratives reported in clinical study reports provide clinical evidence of AEs that occurred to a patient and help scientific safety reviewers. The manual review of these narratives is a daunting task because it is time consuming and resource intensive.

Sabo is exploring the use of deep learning to scale back on manual reviews of patient narratives. In a recent study, he implemented deep learning technology on freeform AE narrative data with the goal of accurately categorizing one AE: Serotonin Syndrome. The results of the study are exciting because they indicate that deep learning and machine learning can improve the speed, accuracy and interpretability of medical coding for adverse events.

Ethical AI

I recently had the opportunity to chat with Sabo about AI and the future of drug safety. “AI gives us the opportunity to get closer to the optimal balance between efficacy, which includes safety, speed and costs in drug development,” he said.

We also talked about the importance of ethics in AI and the need for human experts to be able to understand the results. “It’s critical for non-technical experts to be able to review the results of AI models and ensure that they make sense. This is one of the key strengths of SAS® Viya®. It makes the AI modeling process accessible and understandable for all stakeholders, which empowers ethical AI,” he said.

Sabo’s recent work is just one example of an area of exploding growth for the application of AI in drug safety for the pharmaceutical industry and regulatory bodies alike.

Expanding partnership with the FDA

SAS recently announced a major expansion in our 40-year partnership with the US Food and Drug Administration (FDA) that will give the agency access to new capabilities in natural language processing, AI and machine learning through SAS Viya. The FDA’s Center for Drug Evaluation and Research (CDER) will use SAS to advance digital transformation efforts such as modernizing regulatory programs and driving analytical manufacturing facility surveillance.

Next steps


About Author

Cameron McLauchlin

Senior Product Marketing Manager, Global Health and Life Sciences Practice

Cameron McLauchlin is the global lead for industry product marketing for SAS’ business within the life sciences industry. In this role, she manages global marketing and sales enablement initiatives, with an emphasis on developing consistent and effective messaging to promote the benefits of SAS solutions to customers. Prior to joining SAS, Cameron worked in ad agency account management for 13 years where she led marketing, branding and business development initiatives for a variety of clients in health care and pharma. Her specialties include strategic planning and persona-driven message development. She is also a seasoned marketing project manager with experience across a wide range of tactics including traditional and digital advertising, sales collateral, email marketing, website development and tradeshow activation. Cameron has a BA in Journalism and Mass Communication from the University of North Carolina at Chapel Hill.

Comments are closed.

Back to Top