Former FDA Deputy Commissioner Forecasts the Future of Evidence Generation
Evidence generation has undeniably grown beyond traditional trials to include a number of other sources and continues to evolve. Amy Abernethy, former FDA deputy chief, sees a future in which therapies are approved earlier and continually evaluated with high-quality data, but it will be on industry to connect the dots and build a strong foundation to make this a reality, she says.
Abernethy, who pushed for the expansion of real-world data and evidence (RWD/RWE) capabilities during her FDA tenure and now serves as Verily’s chief medical officer and president of product development, sees the evidence-generation landscape shifting away from the traditional phase 1 to 4 clinical development journey and moving toward “continuous evaluation” across the lifecycles of drugs, devices, biologics and healthcare delivery interventions.
A primary example of this shift is the way industry pivoted to manage the COVID-19 pandemic and deliver vaccines at unprecedented speeds, Abernethy said during the Summit for Clinical Ops Executives (SCOPE) conference in Orlando. Reflecting the urgency of the situation, Pfizer/BioNTech’s and Moderna’s mRNA vaccines were first greenlit by the FDA under Emergency Use Authorization based on a smaller amount of data before receiving traditional approval. This demonstrates well the future she envisions, she said.
“These were shorter trials with a fewer number of patients. Then we watched the story of duration of effectiveness — do you need a booster? Can you mix and match? What about our children, myocarditis?” she said. “These questions played out through a combination of RWD studies, health system studies and more clinical trials. That’s how I think this world is going to play out.”
To Abernethy, the four guidances the FDA posted at the end of 2021 on assessing health records (EHR)/medical claims data, assessing registries, data standards in drug submissions containing data from RWD sources, and considerations for trials using RWD/RWE, are a strong jumping-off point for achieving modernized evidence generation. But industry will need to figure out the methods and capabilities, she said.
Building toward this future and achieving improved evidence generation, in her mind, requires three things of industry.
The first is the development of trial software that enables greater efficiency, flexibility and the delivery of high-quality datasets. For example, Abernethy believes software that allows for both centralization and decentralization of trials will be necessary, as well as the use of eConsent and patient-reported outcomes. “Being able to collect the right data at the right moment in time is very fundamental for this future,” she said.
Second, industry needs the capability to construct datasets that are longitudinal — that is, sets of data that are collected over time from the same participants. Abernethy believes these datasets will be “an underpinning of clinical research in the future.” To her, this will entail combining clinical trial data with additional datasets and building participant-centric registries that will gather high-quality data over time.
Abernethy sees clinical trials as snapshots in patients’ lives. “Most people enrolled in the trial have lived before the trial and hopefully live after,” she said. “If we think about the representation of our life journey through our data story, then we need to think about how we assemble datasets in a way that combines clinical trial data with the rest of the data in the ecosystem around individual patient stories that are then wrapped up in the context of cohorts.”
Abernethy gave an example: the need to evaluate across product lifecycles in medical devices/software as a medical device is now being seen elsewhere, most prominently in cell and gene therapies as well as oncology. Many cell and gene therapies will, per FDA guidance, require 15 years of longitudinal followup, creating a need to develop ways of getting that information from patients that provides strong evidence of safety and effectiveness without overly burdening them.
“In the context of modern evidence generation, we’re seeing the need for more evidence with greater depth, more generalizability across all populations, and the ability to play out stories across time,” she said. “I think time becomes one of the really important features in our evidence-generation story of the future.”
Lastly, with these longitudinal datasets, industry will need to think hard about how to tackle future novel study designs, the potential conduct of multiple analyses per datapoint, and the longitudinal evaluation and followup that will be necessary as part of the new world of evidence generation.
“We need to leverage new capabilities, data streams of all types — data that comes from imaging, digital sensors, tissue, the EHR, all these different data streams,” she said. “Molecular is not the only data stream, and we need to think about how we put all that together.”
Upcoming Events
-
23May
-
21Oct