Enhancing Drug Safety with AI and Automation Technologies – with Marie Flanagan of IQVIA

Generative AI has been accelerating transformation across the life sciences, leaving pharmaceutical companies and clinical research organizations facing a dual challenge: modernizing pharmacovigilance while maintaining uncompromising standards for patient safety, data integrity, and global regulatory compliance.

This delicate balance is most critical in adverse event detection and signaling, where a single missed risk can trigger public health crises, a loss of trust in therapies, and regulatory action. OECD data show that avoidable adverse medication events alone cost $54 billion USD annually.

A 2023 McKinsey Global Institute report estimates that generative AI could unlock $60–110 billion in annual value — roughly 2.6–4.5% of revenues — across the pharmaceutical and medical-products industries, driven largely by productivity gains in R&D and knowledge work. The report further notes that generative AI and related technologies have the technical potential to automate activities accounting for 60–70% of employees’ work time across the economy.

For global organizations managing safety data at population scale, the opportunity extends far beyond efficiency. Every automated decision must be traceable and aligned with patient outcomes.

Emerj Editorial Director Matthew DeMello sat down with Marie Flanagan from IQVIA on the ‘AI in Business’ podcast to continue their conversation about enhancing drug safety with AI and automation technologies and how these technologies are redefining pharmacovigilance.

The following article will focus on three key takeaways from the conversation:

• Driving proactive safety with social media data: Leveraging social media data to develop predictive pharmacovigilance models that anticipate adverse events and enable earlier intervention.
• Automating drug safety transformation: Combining optical character recognition (OCR), robotic process automation (RPA), and natural language processing (NLP) to identify adverse events faster and more accurately across vast, unstructured safety data sets.
• Regulatory collaboration in AI-driven pharmacovigilance: Building cooperative oversight models between life sciences operations and regulators to align digital innovation with patient safety.

Listen to the full episode below:

Guest: Marie Flanagan, Director of Product Management in Digital Projects and Solutions, IQVIA

Expertise: Drug Safety, Lifecycle Safety, Regulatory Reporting

Recognition: Marie has held numerous roles at IQVIA, beginning in drug safety in 2007. Over the next two decades, Marie moved through various account manager roles, rising through the ranks to director level, and to her current role. She also holds a Bachelor’s Degree in Microbiology and Immunology from the University of College Cook from 2004.

Focusing on Proactive Safety

When asked about how AI, and specifically NLP, has transformed safety signal monitoring and detection from IQVIA’s vantage point, Flanagan responds. According to Flanagan, IQVIA has been in the AI space for a long time.

Flanagan explains that the product she supports, Vigilance Detect, uses NLP trained on a proprietary bank of safety-trained keywords and patterns for the last 13 to 14 years to find safety events earlier in the data pipeline than what a traditional safety workflow allows.

Flanagan goes on to explain how NLP is used to sift through a vast amount of unstructured data, including very diverse multilingual datasets in an attempt to find adverse events or product complaints. This is a key example of using technology to eliminate very tedious, manual work, according to Flanagan.

She details how IQVIA works very closely with regulators and influential industry players to maintain an open dialogue. Flanagan also adds that the regulatory environment in her industry is changing to be more patient-centric in how it deploys technology and digital channels for patients.

Flanagan tells the Emerj executive podcast audience that IQVIA has achieved two things by opening up patient channels:

• Has enabled more information to come in
• Has given healthcare workers who interact with patients more time in their day to exercise empathy, have conversations, and convey more clinically robust information to regulators

Until very recently, using social media for patient information was scoffed at, according to Flanagan. Also, the burden was too great, and there were no safety signals or events of consequence on social media.

Flanagan notes how that stigma is changing, particularly centered around the ability to harness the information in a different way. She explains how the information can be used for pre-signaling.

Companies like IQVIA can hotspot influenza incidents in specific Manhattan suburbs and correlate that data with later data in the FDA’s Adverse Event Reporting System (FAERS) a couple of months later.

Flanagan insists that correlating data with official resources via FAERS opens the door to using the information by applying analytics to help determine which signaling activities a company could use for its products before they ever become a problem.

Integrating Natural Language Processing and Automation to Transform Drug Safety

When asked about the opportunities and challenges she sees at IQVIA in leveraging automation technologies such as OCR and RPA for adverse event detection, Flanagan offers important insight. She contends that IQVIA would not achieve results nearly as good as they have achieved if they relied solely on AI or NLP in its purest form.​

She then explains that ensuring safety is a highly complex, end-to-end process, and that IQVIA has benefited from combining a mix of automation technologies. Flanagan emphasizes that if the company used NLP as a standalone approach to identify adverse events, it might only result in a 20-30% positive outcome.

“Once we apply OCR, RPA, and various other automation techniques to take that data and move that data around in a way that is usable, we can get up to 70-80% positive outcome in searching social media for adverse events.

So, our best projects with our most favorable outcomes have been because we have mixed RPA, OCR, traditional coding, traditional machine learning, with more advanced AI techniques. That’s when we see the magic happen, and that’s the artistry of this, what we do in safety and particularly in finding safety adverse events in unstructured data.”

– Marie Flanagan, Director of Product Management in Digital Projects and Solutions at IQVIA

Regulatory Collaboration in AI-driven Pharmacovigilance

​Flanagan provides helpful insight when asked about where she sees the dynamic for regulatory bodies like the EMA and FDA emphasizing digitalization in real-world evidence for pharmacovigilance. The conversation highlighted a major transformation in how regulators interact with the life sciences industry.

For decades, regulation has been seen as a necessary safeguard, but it traditionally slows innovation to protect patient safety. However, Flanagan explains that the age-old regulatory dynamic is shifting as regulators are encouraging increasingly digital transformation within safety operations and becoming active participants in the modernization of pharmacovigilance.

She tells Emerj’s podcast audience that, previously, IVQIA had to switch off comments on social media channels because they simply didn’t know how to handle them. More recently, though, they’ve opened up comments on social media to allow patients to directly interact with regulators. As a result, regulators are actively sharing that information with the industry, leading to a much more collaborative relationship.

Flangan, though, is cautious when describing the current dynamic, “That’s not to say it’s become harmonious, in that we do not have harmony of regulations between the major regions,” she tells the Emerj podcast audience. “But we do see them working together, and we do see them setting out a very, very broad set of standards and guidelines for us in our industry in relation to the use of AI.”

Furthermore, she adds that regulators have not given IQVIA or other companies granularity in terms of their expectations. They essentially transfer that responsibility onto the company, according to Flanagan. She explains how regulators are keeping things broad enough not to inhibit her company’s progress in implementing AI.

She offers very insightful concluding advice for life sciences leaders that helps clarify how their organizations can operate effectively in a more collaborative, digitally enabled regulatory environment. She emphasizes the need to assess and optimize workflows before pursuing AI, focusing on high-value, practical digital enhancements, choosing the right use cases, and avoiding over-engineering by starting small.