FDA plans master protocols to monitor COVID vaccine safety, efficacy
After a vaccine against COVID-19 is authorized or approved, how will its safety and efficacy be assessed? In addition to ongoing postmarket surveillance requirements, the US Food and Drug Administration (FDA) and the Centers for Disease Control and Prevention (CDC) will strive for near real-time safety and efficacy surveillance of COVID-19 vaccines.FDA is currently developing master protocols to guide its safety and effectiveness oversight, said an FDA official during the FDA’s first COVID-19 vaccine-focused advisory committee hearing. “We’ll be posting the protocols for public comment,” said Stephen Anderson, PhD, director of the Office of Biostatistics and Epidemiology at FDA’s Center for Biologics Evaluation and Research (CBER).
Anderson, speaking to the Vaccines and Related Biologics Products Advisory Committee (VRBPAC), reviewed the agency’s resources and plans to track how vaccinated individuals fare, emphasizing the big data approach and resources FDA will draw on as it collaborates with sponsors and CDC to follow the COVID-19 vaccines’ track records. The committee met on 22 October to consider general approaches in authorizing or approving vaccines against COVID-19, without discussing any particular vaccine candidate.
By way of background, Anderson noted that in addition to ongoing passive surveillance through the Vaccine Adverse Event Reporting System (VAERS), the FDA Amendments Act of 2007 directed FDA to come up with a system of active risk identification and analysis. As a result, the Sentinel and Biologics Effectiveness and Safety (BEST) programs now cover over 100 million people. “And that’s the number that we shoot for when we’re doing these types of safety evaluations,” said Anderson.
Considerations for FDA in COVID-19 vaccine monitoring include the need for rapid access to data for “near real-time surveillance,” said Anderson, adding that achieving this rapid cycle analysis approach requires a big data approach.
In terms of safety, FDA is planning to track between 10 and 20 outcomes of interest. “The question is, once we get these signals, how do we adjudicate them?” asked Anderson. He said that epidemiologic analyses are planned to be able to “resolve potential safety signals identified from near real-time surveillance…. We’ve got the capacity with these large databases to do that.”
The agency is also anticipating that case-control and cohort analyses can be used to probe further if safety signals are detected.
For effectiveness surveillance, said Anderson, the problem is that there may initially be limited information at the time of licensure on just how effective a vaccine is, and how durable vaccine response might be. “I just want to remind people that manufacturers have a part in this as well,” including pharmacovigilance and proposing their own postmarketing studies.
Along with CDC, FDA will be looking at general effectiveness, including probing data for effectiveness in certain subpopulations of interest. The agencies will also be looking to see how any authorized or approved vaccines fare in terms of durability of protection.
FDA will build on previous experience with influenza and herpes zoster vaccines in terms of effectiveness and durability, he said.
Ideally, the system should include large databases that cover tens of millions of individuals, so that very rare serious adverse events can be detected. The full spectrum of care from outpatient data to that from hospitalized patients should be included in safety surveillance. Ideally, records should have “significant clinical detail,” as is found in the medical record.
Several organizations have partnered with FDA in the BEST system, including IBM Watson, Humana, Optum, and other industry and academic partners. Getting at claims data gives large volumes of data, but there can be a lag of one to several months before data are accessible – not ideal for safety surveillance of a newly released vaccine, acknowledged Anderson. However, one data source alone, MarketScan, yields claims data for 254 million patients.
The FDA has had a vaccine safety partnership with the Centers for Medicare and Medicaid (CMS) since 2002, giving FDA access to data for about 55 million participants who are aged 65 years and older. Over 90% of the elderly in the US are Medicare recipients, so this is a representative database and not a mere sample, said Johnson.
“Many times, electronic health records provide a richer source of data than claims data,” said Anderson, noting that BEST has partnered with many large electronic health systems including Optum, which encompasses 105 million patients. Here, the lag is often quite brief – as little as a week in some instances. In some cases, health record data and claims data are linked, enriching data quality.
Final protocols and study reports will become available on the Bestinitiative.org website, said Anderson.
FDA