To defeat COVID-19, the FDA needs a new testing paradigm

To safely reopen economies and prevent future outbreaks due to COVID-19, we need a paradigm shift in how we think about diagnostic testing. With traditional diagnostic testing for an infection, a patient presents with symptoms and a diagnostic test is performed to rule-in or out a particular diagnosis. With COVID-19, a significant percentage of those infected, and a disproportionate number of those who spread it, are completely asymptomatic.

Thus, to enable prompt identification and self-quarantine of those infected, and to make contact tracing possible, we need to test people who are asymptomatic, and we need to repeatedly test them since someone who is not infected one day may become infected the next and would never know it. Testing must also be rapid, with quick on-site results to enable a gating function; checking for infection before boarding a plane, visiting a nursing home, or entering the workplace, factory, or other sensitive sites. COVID-19 diagnostic testing must be fast, frequent and widespread.

In contrast, Americans today have to sit in lines for hours just to get a single test and then wait for days or even weeks to eventually get results, rendering their results more of a historical record than actionable information, leaving the infected individual to unknowingly spread the virus, and making contact tracing virtually impossible.

Simply put, if the goal is to prevent or manage a pandemic involving a highly infectious — but often asymptomatic — disease like COVID-19, then a highly sensitive test that is only administered once to a small percentage of the population with days to a week to deliver results is a failed test, no matter how accurate it may be.

So why don’t we have rapid, non-invasive, and frequent testing? It’s not for a lack of technology or innovation We at Columbia University Fertility Center and others have developed non-invasive diagnostic tests that can deliver results in 30 minutes in the field without the need for expensive equipment or highly skilled technicians.

The problem is with the way the FDA views testing. Currently, the “gold standard” diagnostic test for COVID-19 is RT-PCR, which requires sophisticated and expensive machinery and is typically run in high-complexity and severely backlogged reference laboratories and can take days to return results. Any newly developed test must essentially perform as well as the “gold standard.” In the FDA’s mind, having no test is better than having a slightly less accurate test. 

But a slightly less accurate test done repeatedly, and yielding results in a useful timeframe is far more beneficial than an accurate test done once. To take an extreme example, imagine a test that is sensitive enough to positively diagnose infected individuals 95 percent of the time, but is only done once, and another test that only picks up 50 percent of positive cases, but can be done repeatedly.

By the time you test the same person four more times, the chances that the 50 percent test picks up the virus is 97 percent, already better than the “more accurate” test. Add in some bonus facts — that the less sensitive test can deliver results immediately versus days or weeks, vastly more people can be tested, and testing is likely to be done closer to when the person contracted the infection — and suddenly the 50 percent sensitive test is doing a far better job at protecting the individual and society. Imagine what we can do with a 90 percent sensitive test.

COVID-19 is a very different type of virus, and thus the strategy to defeat it must also be different from what has been done in the past. What is needed now is a test that may be imperfect but can be repeatedly given to the majority of the population and deliver results immediately. Those tests exist. We just need the FDA to allow them to be used.

Zev Williams M.D., Ph.D., is the Wendy D. Havens Associate Professor of Women’s Health and chief of Reproductive Endocrinology and Infertility at Columbia University Irving Medical Center.