While many Pfizer colleagues worked to, develop and manufacture a vaccine for COVID-19, one Pfizer team was hard at work monitoring the safety of the COVID-19 vaccine trial. And they did so while also maintaining safety monitoring operations for more than 1,000 other products in the Pfizer portfolio.

Pharmacovigilance is the name for the area of science which scientists and medical professionals monitor medicine and vaccine-related safety issues. “Medicine and vaccine safety is one of those topics where if it’s done right, it’s not necessarily at the forefront of everyday people’s minds,” says Jamie Wilkins, Head of the Risk Management Center of Excellence in Worldwide Safety at Pfizer.

As the new vaccine was authorized and made available to people around the world, real-time monitoring would be critical, says Wilkins. And technology—specifically, artificial intelligence—would be a crucial tool in processing large amounts of information 24 hours a day for analysis by Pfizer scientists.

Fortunately, the foundation for that technology had been laid years prior, enabling fast and responsive action.

“We were able to evolve the technology we already had in order to meet the demands of the pandemic. The groundwork was really critical, because it allowed us to handle a large additional case volume,” says Wilkins, “and simultaneously helped us to remain compliant with regulatory standards and to continue monitoring the safety profiles of our other vaccines and medicines.”

The science of medicine and vaccine safety

For Pfizer, medicine and vaccine safety is always a top concern. And since 2014, the company has been pioneering the use of artificial intelligence in pharmacovigilance and safety monitoring to help sort through and categorize the "case reports" that people file when they experience something called an “adverse event,” which could be an unexpected side effect in conjunction with a Pfizer product. Because the technology automates repetitive tasks, it can collect and analyze unlimited amounts of data, presenting it to trained experts to view and act upon. Leveraging technology to automate that kind of repetitive work frees up scientists and other team members so that they can focus their attention on more sophisticated tasks centered around patients.

This was imperative when the pandemic hit, says Bhavin Patel, who is Head of Case Processing Automation and Intelligence with Pfizer Worldwide Safety. Statistically speaking, Patel says, because of the unprecedented volume of COVID-19 vaccinations, amounting to billions of people suddenly sharing information about their individual experiences, an increase in his team’s workload was to be expected.

Following emergency use authorizations and approvals, the vaccine went on to become available to billions of people. Processing information received about any adverse events for a vaccine or medicine —such as pain at an injection site, a headache following getting a vaccination or taking a medicine, or any number of other experiences— is an important part of the pharmacovigilance process and a regulatory requirement.

In fact, Patel and his team encourage people to report their experiences, because that helps them better understand potential side effects that people might experience. “We want people to report events to us,” says Patel. “This helps us work with the scientists, physicians and other professionals within Pfizer to ensure all prescribers, physicians, pharmacists,and patients have accurate information about the benefits and risks of our medicines and vaccines. That enables them to make an informed decision about the best use of our medicine for them.”

Collecting information at scale

As the vaccine rolled out, Patel’s team needed to devise a way to best apply artificial intelligence so that they could collect and analyze experiences with the vaccine, while simultaneously monitoring the rest of the global portfolio—numbering over 1,000 medicines and vaccines, with more in development.

To prepare, Patel says his team looked to the past. “We tried to use the closest historical experience that we had, which was the H1N1 pandemic. We looked at that, and the publications from that time and anticipated what the increase in reporting would be,” he says. It quickly became clear, based on that modeling—as well as real-time reporting—that hiring more employees to do the work wasn’t feasible. “We couldn’t just keep adding people to the problem,” he says.

“We hired additional trained pharmacovigilance professionals. However, it quickly became clear that our experience during this pandemic was very different than the prior H1N1 flu pandemic experience, Patel says. “We couldn’t just keep adding more people to manage the increasing workload.”

So, employees within Pfizer’s Safety organization came up with creative solutions. “We started thinking outside the box, and asking 'How can we streamline processes to be more efficient?' 'How can we introduce new technology?' 'How do we train the people we have to handle a broader spectrum of responsibilities within case processing?'” says Patel. “The fact that we already had some of the tools in place gave us a great foundation.”

The team came up with close to 120 ideas, and about half of those were green-lighted for implementation. “We formed nimble and empowered teams that included subject matter experts from the safety team and also people from the digital and information management organization to deliver technology and process efficiencies going forward,” says Patel.

Bringing new medicine and vaccine safety solutions to life

Within about six months, many of those ideas were put to work, and they have been instrumental to medicine and vaccine safety monitoring, says Patel.

He shares examples such as an online adverse event reporting portal called COVAES, that was available in multiple languages, 24 hours a day 7 days a week for patients to directly report their experiences with the vaccine. COVAES is accessible from www.pfizersafetyreporting.com, and it asks questions tailored to the individual about their experience associated with vaccination.

“It’s targeted to collect complete and relevant information from the persons who receive the vaccine” says Patel. Those who don’t want to use the portal can chat with Agent LEO, an AI solution that collects information from automated voice conversations. In addition, the team designed a bot to fully automate the downloading, triage and data entry of adverse event reports filed with the European Medicines Agency. Another technology solution provided significant efficiencies during review of medical journals for potential reports of adverse events related to Pfizer products.

By automating these tasks, this technology has saved countless hours, says Patel. “These are all things that need to be done, but they’re repetitive,” he says. “With the help of AI, we're able to reduce the case processing cycle time and manage a much larger volume of reports. Efficient management of the large volume of reports enables the scientists and professionals within Safety to have timely access to this data so they can communicate any important messages to the patients, physicians and regulators.”

Building for the future

Artificial intelligence is becoming more integrated into everyday peoples’ lives and also the way businesses operate in every area—including healthcare and, in particular, medicine and vaccine safety operations. As one example, Pfizer has been working with a data and analytics company that collects anonymized, de-identified, and secure health data from electronic medical records of around 80 million people in near real time.

Heather Rubino, who is Head of Safety Surveillance Research at Pfizer, says that by accessing that volume of information, the company can derive insights and monitor patient safety en masse. “Almost one-third of the United States population is covered by these systems, so we get a really nice geographic representation of people that are taking our products, and that helps us measure the safety of our products across the entire United States, in near-real time,” says Rubino. In the past, she adds, analysis at that scale previously may have taken years. Now, she says, her team is able to glean and act on insights in months, or even weeks.

Looking ahead, Rubino says that Pfizer’s technology will continue to evolve, in preparation for the next challenge in healthcare. Because, she says, there will always be a “next.”

“It’s critical to build infrastructure and lay groundwork in times of calm, so it’s ready for use in a crisis,” she says. “Before you can get an ambulance to a hospital, you must build roads."