Accelerating the collection of real-world data about COVID-19

BioCentury is providing this story for free given the urgent need for information about the COVID-19 crisis. For more analysis, sign up for our daily email

If there are silver linings to the COVID-19 crisis, one is the way groups and individuals who often have adversarial or competitive relationships are joining together to fight a common threat. Another is the acceleration of science and medicine that comes from overcoming barriers that have slowed progress.

The COVID-19 Evidence Accelerator, an initiative of the Reagan-Udall Foundation for the FDA and Friends of Cancer Research, embodies both of these trends.

Its overall goal is to create ways to learn from the COVID-19 patient experience and reinvest that knowledge in ways that benefit future patients.

“The Reagan-Udall Foundation and Friends of Cancer Research have created a portal and process to acquire and aggregate the avalanche of real-world data about COVID-19 so as to facilitate and accelerate the FDA’s ability to rapidly and efficiently analyze and act on that data,” former FDA Commissioner Andrew von Eschenbach told BioCentury. von Eschenbach is president of Samaritan Health Initiatives Inc. and a member of the Reagan-Udall Foundation’s board of directors.

The Accelerator is bringing multiple teams together, working in parallel to gather real-world data that could answer specific questions.

But the real benefits are likely to come not from the answers but from the steps that must be taken to ask the questions and then make sense of the answers.

That process could produce invaluable knowledge about the natural history of COVID-19, shape the way real-world data is collected in the future, and help regulators determine what kinds of decisions such data can reliably support.

The Evidence Accelerator

The Accelerator’s call to action requires health systems, data companies, academics and others to get to work and quickly come to agreement on topics that outside a pandemic pressure cooker could take months or years.

They are identifying core data elements, identifying those that are relevant to specific questions, and agreeing on uniform methods for collecting and reporting data.

The Accelerator is not creating a massive data pool; there is no data transfer involved. Rather, its participants are formulating strategies that allow them to query their own data sources and obtain results that can be compared with those obtained by other groups.

The distributed approach has the advantage of making it possible to cross-check results obtained by different groups and proceed without having to put in place infrastructure to transfer and combine data.

Reagan-Udall and Friends said they cannot yet identify the participants because commitments have not been finalized, but they told BioCentury the Veterans Administration, Duke University, Yale University, the Sutter Health system, Novartis AG (NYSE:NVS; SIX:NOVN) and a number of other organizations are involved.

Reagan-Udall was created by Congress to serve as an interface between FDA and the private sector, while Friends, which is providing the Accelerator with expertise on real-world data, has a track record of advancing oncology by pushing regulators, academics and drug companies to come together to solve practical problems at the intersection of regulation and science.

The real benefits are likely to come not from the answers but from the steps that must be taken to ask the questions and then make sense of the answers.

Hydroxychloroquine in the real world

The first question the Accelerator is posing involves hydroxychloroquine, which is being widely used off-label and outside of COVID-19 trials.

Hydroxychloroquine and chloroquine have an received Emergency Use Authorization (EUA) from FDA for hospitalized COVID-19 patients.

Guidance from NIH recommends that because of the lack of data on efficacy and known toxicities, both drugs should only be used to treat COVID-19 in a clinical trial.

FDA released a warning on April 24 about the potential for the hydroxychloroquine and chloroquine to cause cardiac toxicity, which the agency said could be worsened by the addition of azithromycin (see “FDA Warns About Dangers of Hydroxychloroquine, Chloroquine for COVID-19”).

About 70% of hospitalized patients, however, are receiving one of the drugs outside of a clinical trial, often in combination with azithromycin.

Accelerator participants are being asked to provide data on several parameters for hospitalized patients who are receiving hydroxychloroquine with or without azithromycin, and for matched control groups.

Parameters include: characterization of patient populations in the treated and control groups; stage of disease when patients are treated; doses and whether patients are receiving monotherapy or co-prescriptions; safety signals, including in subpopulations and control groups; and comparative effectiveness vs. control groups.

Reagan-Udall and Friends staff expect four or five groups to respond in parallel to the question.

Creating a synthetic control arm

Even if the Accelerator does not obtain actionable data about hydroxychloroquine, the process could be valuable for two reasons.

First, it is forcing a diverse group of stakeholders to agree on a common set of tools, ranging from query vocabulary to methods of selecting control groups, for gathering real-world data.

Second, in the course of gathering data about the outcomes of patients who received hydroxychloroquine, Accelerator participants will obtain natural history data from patients who haven’t received the drug.

This data could be used to create synthetic control arms (see “Broadening Role for External Control Arms”).

There is a dearth of knowledge about the natural history of COVID-19, and the history is likely changing over time, so there is a need to continuously reapproximate control arms.

In particular due to the continual updating, those datasets could provide answers to critically important questions, including signals about the efficacy of elements of standard care, how the populations that are hospitalized change over time or across regions, and differences in subpopulations.

“If there is an efficient way to acquire and aggregate that data and then be able to present that in a way that the FDA can efficiently analyze and act on that data, then they can make decisions that inform the American people as to what’s working or not, for whom it’s working in what setting and at what dose,” von Eschenbach told BioCentury.

Lab meetings

The Accelerator is also providing a venue for participants to share results and generate new hypotheses via weekly lab meetings.

The first meeting featured presentations about how real-world data is being used to gather actionable data about COVID-19.

Atul Butte, director of the Bakar Computational Health Sciences Institute at the University of California San Francisco, discussed how the University of California is extracting data about COVID-19 across six medical schools and health systems, and gave some idea of the complexities.

For example, according to a summary posted by Reagan-Udall, Butte reported that within the UC system, there are 24 different ways to order a a COVID-19 test, and at “UCSF, it takes 300 lines of code to determine whether or not a patient is on a ventilator.”

One thing the UC system has learned from real-world data, Butte said, is that “end stage renal disease is a preliminary strong predictor of which SARS-CoV-2 positive patients will be admitted to the hospital.”

Patrick Ryan reported at the initial lab meeting on the COVID-19 activities of the Observational Health Data Sciences and Informatics (OHDSI), a collaborative, open-source network that draws on about 600 million patient records from 18 countries .

Ryan is VP for observational health data analytics at the Janssen Pharmaceutical unit of Johnson & Johnson (NYSE:JNJ) and assistant professor at Columbia University

Data from OHDSI is the “largest study ever on the safety of hydroxychloroquine using historical data aggregated across 14 partners,” according to a summary of Ryan’s lab talk.

The study concluded that hydroxychloroquine “appears safe in short-term use for rheumatoid arthritis, but long-term use may be associated with increased cardiovascular mortality.” It also found that hydroxychloroquine plus azithromycin “increases 30-day risk of heart failure and cardiovascular mortality.”

Further analysis of the coronavirus crisis can be found at

How to gain access

Continue reading with a
two-week free trial.