With countries across the globe preparing for a potential second (or third) wave of COVID-19 cases, drug developers are working around the clock to uncover a vaccine. Many are collaborating in new and different ways to quickly bring a safe and effective vaccine to market. Biopharmaceutical companies and health authorities worldwide are purposefully moving faster and more flexibly than they ever have before.
It traditionally takes 10-15 years to bring new vaccines to market. However, with COVID-19, regulators have been working with the pharmaceutical industry to try to drastically reduce timelines. Given the circumstances, regulatory bodies like the Food and Drug Administration (FDA) have flexed regulatory requirements with the hope of more quickly uncovering effective treatments.
When one or more vaccines are ultimately approved for distribution, they will need to be delivered to most of the planet’s seven billion inhabitants. While there is tremendous urgency to create and scale these treatments quickly, this cannot come at the detriment of safety and efficacy, lest it contribute to a much larger problem than we currently face. Safety and pharmacovigilance teams must be more equipped than ever to monitor and report potential adverse events in real time from clinical development to the post-approval phases – as the number of vaccine recipients rockets from tens of thousands to millions and billions, involving every kind of population cohort imaginable. These kinds of big data and advanced analytics scenarios on such crash timing cry out for a major assist from artificial intelligence (AI) and other emergent technologies and approaches.
Prior to the advent of COVID-19, the pharmaceutical industry had been slow to embrace advanced technologies including AI in safety and regulatory functions. The pandemic is shining a light on how these technologies can help companies swiftly and accurately navigate complex administrative requirements especially in the face of a massive influx of new data. Here are a few ways adding AI and other novel approaches to pharmacovigilance practices can aid in delivery of a safe COVID-19 vaccine to the world.
Reducing complexity to improve data management
Health data has grown by 878 percent in the past four years. Terabytes of additional data points are compounding by the day. Simply put, this means that humans are now physically incapable of manually processing all necessary data points.
AI technology addresses this challenge by helping pharmaceutical companies process massive amounts of data from structured as well as unstructured sources. It also helps pharmacovigilance professionals decipher which signal detection data points are indicative of potential adverse events and report to regulatory bodies. This means that PV and case processing can be performed in much shorter amounts of times and with a much higher degree of accuracy.
In turn, these technologies provide insights that regulators, industry bodies, and internal experts need to make evidence-based decisions about the safety of developmental and marketed products and future opportunities for commercialised treatments. The agility that AI provides is vital in the case of COVID-19, as it helps the pharmaceutical industry evaluate risks and predict problems before they become crises. AI will be a core tool as the pandemic situation continues to evolve at a rapid pace.
Finding adverse event data in unstructured sources
While data from clinical settings, including clinical trials, are key for bringing new treatments to market, unstructured informational sources contain a wealth of information as well. With the advent of patient reported outcomes (PROs), a multitude of new channels have emerged, including chat groups, registries and social media. These unstructured sources had previously been untapped by the industry – as they require information to be culled from narratives rather than directly lifted from forms and reports.
Today, natural language processing (NLP), has enabled the pharmaceutical industry to convert this information into structured formats. This information contains a variety of insights about patient experiences with commercialised treatments which are necessary for adverse event reporting.
With this capability in place, AI can monitor multiple platforms at once in any language, in real-time. By refining the algorithm over time, these technologies can work together not only to simply pull from keywords but recognise further context that we are able to discern as humans in order to deepen our sense of whether a reported outcome is in fact an adverse event. What’s more, AI pulls relevant context from de-identified data, meaning that while organisations utilise insights from broader data sets, it does so after stripping out personally identifiable information in-line with global data privacy regulations.
AI’s role in maintaining compliance for commercialised treatments
Pharmaceutical companies know all too well that getting a drug to market is only half the battle: keeping it there is an ongoing challenge. Even before COVID-19, pharmacovigilance cases were steadily rising by 20 percent annually. In tandem, regulatory requirements for safety reporting have become more expansive as well as more demanding.
As a result, there is a growing need to scale operations and maintain costs. Compliance has long been viewed as a cost centre for the pharmaceutical industry. Yet when pharmaceutical companies start to see compliance as value-add as opposed to a necessary evil, they will reap the benefits of avoiding delays or denials of product approvals, license withdrawals, financial penalties and, most importantly, protect patients.
AI-powered automation in adverse event monitoring and case processing will be critical in supporting the continued delivery of safe and effective treatments – reducing the administrative burden as well as the potential for human error. This will be critical for COVID-19 treatments to market and as we move into this new normal.
Conclusion
In this data-driven world, embracing technological innovations will be not only beneficial but necessary to deliver the promises of patient safety and treatment efficacy – especially in the case of COVID-19. Equipping the industry with real-time pharmacovigilance information increases the ability to predict and respond to safety concerns as they happen, not only heightening regulatory compliance but also improving patient safety in both the short- and longer-term.
In the near term, these technological advances deliver significant time and cost savings, reduced risks, and lower burnout for pharmacovigilance professionals currently spending most of their time on administrative tasks. In the longer term, aiding in the establishment of proactive safety practices will give pharmaceutical companies the competitive edge they need to see gaps and opportunities in the life sciences landscape and propel the innovation of medicine for the world at large.
Joe Rymsza is vice president of Global Pharmacovigilance and Regulatory Technology Solutions at IQVIA
