The growth in cell and gene therapy development will force industry and health technology providers to “get smart” about combining prospective clinical trial data with real-world data sources and addressing the “data quality mismatches” that will inevitably result, said Amy Abernethy, Verily clinical studies platforms president.
Abernethy, the former US Food and Drug Administration principal deputy commissioner under commissioners Scott Gottlieb and Stephen Hahn, said the landscape of cell and gene therapies should be a “siren song” for the importance of collecting real data in real time. Speaking at an 11 September Friends of Cancer Research real-world evidence event, she named several factors to make the case:
- The length of follow-up required for a cell and gene therapy (15 years or more) will require sensors or wearables to collect data in real time.
- The typical child or adolescent receiving cell and gene therapy most likely has a serious disease at a point in life when they need flexibility in tracking outcomes.
- Informed consent will need to be reframed more as a dynamic “partnership” between patient and development that may change over time, given the length of follow-up.
- Cell and gene therapies will “force us in the real data space to do other things that we haven’t thought that much about yet,” like combining manufacturing data with real-world efficacy and safety data, Abernethy said.
During her more than two years at the FDA, Abernethy brought her health information technology experience (at Flatiron Health) into the agency’s slow, but evolving approach on ways to supplement randomized clinical trials with real-world data. While at the agency, she pushed FDA to lead on RWE to make drug developers more comfortable using the data. (Also see “Real-World Evidence Uptake Is “Silver Lining” Of COVID-19 Response, USFDA’s Abernethy Says” – Pink Sheet, 12 Oct, 2020.)
Abernethy has previously pointed to cell and gene therapy development as an opportunity to think differently about data collection in terms of a “parallel conversation” between the FDA and the Centers for Medicare and Medicaid Services on coverage post-approval. (Also see “Moving Beyond Aduhelm: Cell/Gene Therapy May Be ‘Poster Child’ For What Comes Next” – Pink Sheet, 25 Apr, 2022.)
Abernethy said it was time to start thinking about how to take existing, passively collected data like electronic health records and claims data and pair it with prospectively collected real-world data. That “doesn’t mean either one of them is the most appropriate or accurate data set,” she said. “They’re both going to be needed in order to solve these problems.”
“We’ve come a long way” in using real-world data, Abernethy argued: determining which sources can be leveraged and curated, how to use clinical trials to understand natural history, and beginning to simulate real-world control arms and “asking questions about what endpoints can we derive from many of these data sources.”
First, the technology must be available to link data sets in a longitudinal context to track events over time, she said.
“We also need the capabilities that allow us to deal with voluminous data,” which is where recent advances in artificial intelligence could help, Abernethy said. “The ability to leverage computation at scale in the cloud and collaborative research environments that are different” will be needed.
Another key consideration is the need for greater transparency in how real world data is used, which will require rethinking privacy and security considerations, as well as reframing informed consent into a “partnership” that can change over time, Abernethy said.
“There is an important aspect of consent in the new story of real-world data that we’re going to see more and more, because by definition, prospective data collection is going to be required as we move this real-world evidence story forward,” she said. “This is the moment to partner and give the person agency about what happens with their information and how you interact with them across time.”
Granting permissions in a real-world data environment will require a more “dynamic” partnership that can change throughout a relationship with a patient. They should be able to “pull back” their consent if their data was not used in the way they intended, as well as have the opportunity to say “yes” and move forward, Abernethy said.