Accelerating the fight against COVID-19 with real-world data on tests
Diagnostic Evidence Accelerator deploying real-world data against COVID-19
Almost every aspect of the fight against COVID-19 depends on the ability to determine whether an individual is or has been infected with SARS-CoV-2, but more than six months into the pandemic there is still uncertainty about the accuracy of widely used PCR-based diagnostics and serological antibody tests. The Reagan-Udall Foundation for the FDA and Friends of Cancer Research have launched an initiative, the COVID-19 Diagnostics Evidence Accelerator, that is intended to reduce that uncertainty by collecting and analyzing real-world data about test performance.
Knowledge about how tests actually perform, as opposed to data from in vitro studies that show how they could perform under idealized conditions, could be one of the keys to taming the pandemic.
“I have turned almost all of my attention at the moment to diagnostics because it is clear they are foundational” for the fight against COVID-19, Amy Abernethy, FDA’s principal deputy commissioner and acting CIO, told BioCentury.
Having accurate diagnostic tests and characterizing their performance is important not only for making treatment decisions for individual patients, but also for selecting patients to enroll in clinical trials and for understanding the effectiveness of treatments and vaccines, she said.
“This is not intended to be a COVID-only activity.”
If it is successful, the Diagnostics Evidence Accelerator could produce data that will solidify the foundation for reopening schools and businesses, developing therapeutic interventions and characterizing the efficacy of vaccines.
It could also create a template for the kind of diagnostic performance data-sharing that should be routine whether or not the world is scrambling to fight a pandemic.
The accelerator is working closely with FDA and will provide data the agency could use to make regulatory decisions, but it also plans to make the results of its work available more broadly. “Our goal is to get useful information about test performance out to the community,” Ellen Sigal, chair of Friends of Cancer and the Reagan-Udall Foundation, told BioCentury. “Whether regulators do anything with the information is up to them.”
Connecting the pipes
Abernethy said the accelerator is “creating handshake points” to facilitate data-sharing between testing labs and healthcare systems. The idea is to collect data on large, representative cohorts that will include the kind of COVID-19 tests individuals received and test results, correlated with patient data including prior exposure to SARS-CoV-2 and medical outcomes.
In the disjointed U.S. healthcare environment, collecting and reliably correlating these datasets is difficult because the companies and organizations responsible are not in the habit of speaking with each other, and often use inconsistent terminology.
The accelerator is “connecting up the data pipes so we can analyze the interactions among test manufacturers, test results and outcomes for patients, at scale,” said Abernethy.
Large laboratory companies Quest Diagnostics Inc. (NYSE:DGX) and Laboratory Corp. of America Holdings (NYSE:LH) are both participating in the accelerator.
Other participants include real-world evidence companies Aetion Inc., HealthVerity Inc. and IQVia Holdings Inc. (NYSE:IQV), insurance providers such as UnitedHealth Group Inc. (NYSE:UNH) and the University of California Health System, and biopharmas including the CSL Behring unit of CSL Ltd. (ASX:CSL), Eli Lilly and Co. (NYSE:LLY) and Roche (SIX:ROG; OTCQX:RHHBY).
First- and second-order questions
As with many other aspects of COVID-19 medical countermeasures development, the development of real-world data about tests is moving collaboratively and quickly, with organizations that usually compete cooperating to make steps in parallel that ordinarily would be conducted in sequence.
“We are simultaneously answering questions while building underlying datasets and capabilities,” Abernethy said.
The accelerator is starting with “first-order,” or more immediate, questions that will provide valuable insights and also create the capacity to answer important second-order questions, she said.
On June 25, accelerator participants agreed on the its initial first-order question: “Among people who have positive PCR tests for COVID-19 and then have a subsequent serology test, how likely is the antibody test to be positive and when does it become positive?”
Answering that question will provide important information about the sensitivity -- but not the specificity -- of antibody tests. The process will take several weeks and require that the accelerator develop capabilities to query relevant datasets that will be needed to tackle related second-order questions such as whether past exposure to SARS-CoV2 confers immunity, and if so, for how long.
“A critical question is at a population level: What do we know about the development of immunity -- the prevalence and characteristics of the population who have had SARS-CoV2 and whether they can they get it again,” Abernethy said.
Understanding reinfection rates is critically important for reopening businesses.
“The performance of the tests is important for understanding immunity and prevalence. If you don’t understand the performance of tests, your understanding of immunity can be pretty erroneous,” Abernethy said.
Although a variety of approaches are being used to understand immunity, the accelerator could, Abernethy said, “provide more precise, scaled information and start to layer on more confidence” and certainty about immune responses.
The accelerator is racing against the pandemic, but the systems it is creating could outlive COVID-19.
“This is not intended to be a COVID-only activity,” said Abernethy, who added that the systems and connections that the accelerator is creating could be broadly applied to understanding real-world performance of diagnostics in the future. “In the past, the pipes weren’t connected up correctly. We will learn what are the motivators for maintaining those kinds of connections” so they can persist.