Ask the Accelerators is a series highlighting the perspectives and work of organizations and individuals participating in the COVID-19 Evidence Accelerator. Organized by the Reagan-Udall Foundation for the U.S. Food and Drug Administration (FDA) and Friends of Cancer Research, the COVID-19 Evidence Accelerator joins leaders in the health care data and analytics space to explore how real-world data (RWD) can inform key COVID-19 research questions. 

Diagnostic testing is a critical piece of the COVID-19 puzzle—not only to diagnose those presenting symptoms and provide the appropriate treatment, but also to identify asymptomatic individuals in an effort to curb the spread of disease.

Additionally, as COVID-19 vaccines become available, neutralizing antibody tests will be key to developing a vaccine distribution strategy by illuminating who already has antibodies, and who needs a vaccine first.

On the leading edge of the effort to develop COVID-19 diagnostics is William Morice, II, M.D., Ph.D., Chair of Laboratory Medicine and Pathology at Mayo Clinic and President of Mayo Clinic Laboratories. Since becoming one of the first labs to join the U.S. response to the pandemic, Dr. Morice has led Mayo Clinic Laboratoriess through several major milestones, including developing and releasing some of the first serologic and neutralizing antibody tests for COVID-19. 

Here, Dr. Morice shares what’s top of mind as he and Mayo Clinic continue to iterate on diagnostic testing for COVID-19, including the potential applications and limitations of real-world evidence (RWE) in supporting this effort. 

Responses have been edited for clarity and length.

Q: Tell us about the work your team is doing around diagnostics for COVID-19. Have you hit any major milestones, or come across any key learnings? 
A:
As part of the integrated practice at Mayo Clinic, we provide all of the diagnostics needed to support the diagnosis and management of patients with COVID-19. That includes molecular testing to detect the virus itself, serologic testing to detect whether someone has been exposed and had an immune response, and neutralizing antibody testing to see if they have seroconverted—meaning those antibodies are bioactive against the virus and detectable in someone’s blood. It also includes all of the other diagnostics used to support the care of the hospitalized patients: cytokine storm assays, and other routine assays used in the management of patients with complex care. 

In terms of major milestones, we were one of the labs called upon by Vice President Pence’s White House team to join the early response to the pandemic. We had a laboratory-developed COVID-19 test introduced in the early days of March, and, by the end of March, we had high-throughput diagnostic processors from Roche Diagnostics operating to increase the speed at which we could process COVID-19 samples on Mayo Clinic’s Rochester and Florida campuses. 

In early April, our serologic test became available as one of the first in the country; we were also one of the first in the country to have a high-throughput neutralizing antibody assay available in early June. 

Mayo Clinic Labs was also one of the foundational partners for the Minnesota partnership on COVID-19 testing, along with the University of Minnesota and Governor Walz. 

Q: How has your work changed or developed as you've learned more about COVID-19? Has RWE played a role in improving your understanding of the disease, and the tests that you need to detect it?
A:
Our work has evolved as the national response to the pandemic has evolved. For the first time, we’re seeing the need to use testing not only for diagnosis, but also for screening and societal management—which is key to managing the pandemic itself. COVID-19 has posed the challenge of being most transmissible before patients are symptomatic, and also of identifying the best specimen to detect asymptomatic individuals. 

One of the first ways we applied real-world evidence to this work was in trying to understand the performance of tests designed for clinical diagnostic use. Real-world evidence helped us understand their strengths and limitations in helping manage the pandemic for both ill and asymptomatic individuals. 

Q: What challenges have you encountered while you've been working with COVID-19 real-world data (RWD), and how have you overcome them? 
A:
The biggest challenges have been with the availability and the interconnectedness of data—both horizontally in an instance of care and longitudinally across an episode of care. In order to understand who is actually COVID-19-positive, you need not only the test data, but also the outcomes—for example, longitudinal data around serial testing and clinical data about contact tracing. There's a lot of information peripheral to the test that's required to understand the significance of test performance, and whether it performed as intended.

To overcome those challenges, we’ve leveraged Mayo Clinic’s integrated practice and integrated medical record to mine our data in an effort to understand the performance of tests.

We’ve also been participating in the COVID-19 Evidence Accelerator, and other initiatives on medical devices and real-world evidence. It's a logical leaping-off point to extend that work into diagnostics, as it requires input from both the clinical lab and the care providers to understand the most meaningful data to accumulate, especially because of challenges with connecting data.

Q: What's keeping you up at night with regards to COVID-19 diagnostics? What's the next set of questions you're hoping to answer, and how can RWE inform the work ahead?
A:
Who is infected and spreading the virus, and how can we leverage all the available tools to help control the spread? That’s the key question that is keeping me up at night, and even once we have a vaccine, its answer will be critical to managing the spread of COVID-19. We're still looking at each tool individually, but I think we can use real-world evidence to gain a more comprehensive view of how the tools, in totality, are answering critical questions.

Now, I’m thinking about how we can access good data to help us understand how tests are performing. We still don't have great information, for instance, on serologic testing: Does it play a role in detecting those who are infected but have negative molecular results? What's the role of antigen testing, which will be less sensitive, but lower cost? We need real-world evidence to answer these questions. 

The other thing that will be keeping me up at night is the question of who needs vaccines. If someone has been infected, how much does their immune response protect them? 

Take convalescent plasma as an example; we have real-world data to support the findings of the Expanded Access Program, but because people participated at many smaller medical centers that were overwhelmed with patients, it was difficult for providers to manually record the data. So now, while we saw convalescent plasma had some impact on patients’ disease, it's difficult to parse out the true effect—and the best ways to predict its effects in additional patients. 

This highlights the limitations and challenges of working with real-world evidence—it’s not the limitations of the treatments that create challenges for researchers, rather, it’s the limitations of the available data. Although it may be easier to collect real-world data than run a randomized controlled trial, especially now as programs studying antibody therapies struggle to enroll patients, my other major concern as we look ahead to vaccines and therapies is whether we have robust enough real-world evidence infrastructures to draw meaningful conclusions about the effectiveness of COVID-19 interventions, and who needs vaccines most. 

Q: What kinds of RWD can we extract from laboratory tests and other data sources to innovate in the diagnostics space? Are there any data elements that are particularly or uniquely important to informing COVID-19 diagnostic development?
A:
In terms of what to extract from the laboratory, methodologic data is important to understanding the performance of different assays. These technologies are often expensive, but we don’t know the relative value of one compared to others because they’re often tested in isolation, or in very controlled settings. And with diagnostics for COVID-19 and otherwise, it’s key to make sure we have good information flowing out from the lab, not only about the results of the tests but about how the results were created.

In particular, COVID-19 has highlighted the importance of longitudinal clinical data. As we assess the concept of repeat infection, or determine how often we need to test someone, we need longitudinal data to answer our questions. 

Q: What's most exciting for you as you look ahead to future opportunities for RWD and RWE to support diagnostic testing, even beyond COVID-19?
A:
COVID-19 has given us a roadmap for how we will assess the clinical laboratory going forward. Before the pandemic, the medical community struggled with finding the best way to determine the effectiveness of clinical diagnostics in supporting care—a challenge which highlighted the fact that we didn't have a holistic view of how data created from the clinical laboratory impacted care, either in a positive or negative way. We can get this information from real-world evidence, and link outcomes to testing. 

Ultimately, we will leave this pandemic—which is exciting in and of itself—and when we do, we will have a more durable health care infrastructure to allow us to build off of this momentum and tackle other challenges in health care, in the U.S. and globally.