Drug Safety: A Holistic View
While individual case reports will always have an important role to play in ensuring drug safety, pharmacovigilance professionals are moving towards an aggregate analysis of real-world data in their effort to understand how drugs and vaccines affect patients.
This “is a huge shift for pharmacovigilance,” asserts Stewart Geary, MD, Global Head of Safety at Eisai Co., Ltd. Dr. Geary joined the Safety Signals Podcast to discuss how scientists and regulators are gaining a more holistic view of drug safety in both the clinical development and post marketing environments.
Listen to the whole conversation.
“During a clinical trial we receive individual adverse event reports from the sites, both non serious and serious reports,” he said. “But in addition to that there’s a great deal of clinical data being collected, which…is available in databases for searching and review for potential safety issues.”
According to Dr. Geary, regulatory authorities are setting expectations for sponsors to look at the entire picture of the data at their disposal. He noted specific requests for an independent body–such as a data safety monitoring committee or board–to review aggregate incidences of events to investigate issues of concern across different study arms.
In the post marketing environment, Dr. Geary said that he’s seen an increased interest in post marketing observational studies and the use of real-world data to inform safety signal detection and evaluation.
Real-World Data: The Next Big Thing
While Dr. Geary believes that “real-world data and the attention to real-world evidence is…the next big topic for making organized assessments of safety and efficacy questions in the post marketing environment,” he cautions that data must be of sufficient quality in order to avoid bias.
Social media can be a way to discover if people are talking about safety concerns, for example, but it is often colored by news reports and other cultural phenomena. “If I’m posting on social media I might be posting about a rumor I heard from my mom who talked to her friends at the Social Club about some very scary event, which may or may not be true,” he said.
According to Dr. Geary, health authorities are “exhibiting an interest in higher quality sources of information and higher quality evaluations of data before they’re willing to consider that real-world evidence for regulatory decision.”
Claims databases and other forms of electronic health records are potentially rich sources of real-world data, but the data needs to be very carefully handled to avoid confounding factors that could fool scientists one way or another about a safety or efficacy issue. Readily available databases can create problems in terms of which patients are being followed consistently over time, which patients are dropping out, and whether data is being counted twice.
“The right way to control for variables like that is to understand the questions you’re asking of the data,” Dr. Geary said. “I think it’s very difficult to do a very vigorous study resulting in real-world evidence if you don’t have the question that you’re asking in advance.”
Dr. Geary said that safety scientists could screen real-world data to see what diagnoses are being made for patients using a particular drug. If that process resulted in any potential safety issues, a more deliberate investigation of the data could look for answers to the questions raised. “That then involves a more pharmacoepidemiologic approach of considering the database, what confounders you might have that are associating your event with that potential safety issue, how you’re going to control for that in the database, perhaps why you might look at not just that database, but a separate database to confirm what you’re looking at,” he said.
AI Tools Can Help
Acknowledging that sponsors “have always had some issues with the data that we have available and how we look at it,” Dr. Geary is hopeful that new artificial intelligence (AI) tools can help. This is especially necessary given the volume and complexity of data that can be generated when new treatments hit the marketplace. Dr. Geary noted that the CDC’s v-safe tool alone generated 1.6 million reports in a single month after people started receiving COVID-19 vaccines.
“It really points to the need to use tools or big data to evaluate that in some way besides using just human eyeballs to look at a line listing of reports,” he said. “When you’ve got a huge amount of data and you’re relying on some interface to tell you what’s in that data, if that interface is wrong you can be badly misled…So I think having tools that we know how to use, and we know how to critically evaluate as we use them, is the thing I feel we’re struggling with right now.”
Dr. Geary is eager to work with technologists to find tools that can look at aggregate data and identify patterns in more sophisticated ways. “I see more technology in our future,” he said. “And hopefully technology which is going to be easier for us to use after we go through this digestion period.
Listen to the whole conversation.
This post is based on an episode of the Safety Signals podcast. To hear more, check us out on Apple Podcasts, Spotify, or on our website.
The views of the hosts and guests featured on Safety Signals are their own and do not necessarily reflect the views of Saama or the individual companies for which the guests may work.