But more rapid tests are under development that could help get the epidemic under control, even though they aren’t as accurate as the ones that take longer. Some are like home pregnancy tests that patients can perform on themselves, get results in several minutes and repeat often. Others, including two that recently received emergency authorization from the Food and Drug Administration, require small devices that can be used in places where we know spread is happening, such as nursing homes and food-processing plants, and in settings where we are concerned that transmission could happen, like schools and workplaces.
It didn’t have to take eight months, more than 5 million cases and nearly 170,000 deaths from the pandemic in the United States for this idea to take hold. Because of what we learned fighting previous outbreaks of Ebola and Zika, we have been advocating for widespread screening with rapid tests since January. Modeling of rapid testing that we did during the West African Ebola epidemic in 2014 showed that if we could find 60 percent of new Ebola cases within a day of patients becoming infectious, we would immediately halt the epidemic. Another study suggested that up to a third of all Ebola infections could have been averted had rapid tests been used then. The risks that an individual test might give the “wrong” result can be far outweighed by the benefit that faster, wider and more frequent testing brings: It gives public health officials the chance to figure out where and how the virus is spreading in close to real time.
One of the keys to stopping any epidemic where no vaccine is available is minimizing the “window of transmission” between the moment someone becomes infectious and when they are isolated and no longer able to infect others. If this is done so each infected person, on average, infects one person or fewer, the epidemic will contract and eventually die out. With the coronavirus, this is challenging because people who become symptomatic are thought to be most infectious before their symptoms start, and up to 40 percent of patients or more never develop any symptoms — yet they may be just as infectious as those who do.
With transmission widespread and uncontrolled in many places in the United States, there is no way to know whether someone may be infectious without testing people frequently, even when they don’t show symptoms. Testing is not only a means of diagnosis for an individual — it is a critical lever for preventing spread to others, and so far, it’s been badly underused in the response to the pandemic here.
But polymerase chain reaction (PCR), the laboratory test currently used to diagnose covid-19, has not yet been scaled to return results fast enough to screen a large percentage of the population every day or even every few days. Due to shortages of reagents and machines, PCR capacity hasn’t been able to match the sheer volume and speed needed. Because it can only be done by specialized labs, PCR testing also requires samples to be brought to these labs, which is logistically difficult to coordinate for millions of people a day.
We ran into similar challenges during recent Ebola and Zika epidemics. During the early period of the West African Ebola epidemic, there was no way to know without testing who among the thousands of people with symptoms like fever and body aches actually had Ebola, as opposed to other common illnesses in the area, such as malaria. At the time, PCR tests could not be scaled to test enough people each day. So many who were sick with Ebola remained in their communities and continued to infect others. When Zika, a virus that causes birth defects in pregnancy, was spreading in 2016, Florida was unable to scale PCR enough to test all pregnant women — only about 2 percent of the population — even once during their pregnancies, much less other people in their communities. In both instances, we proposed using rapid tests to search out who in the community was infected.
Now, with covid-19, the United States faces a similar challenge. With more widespread transmission, we need to serially screen an even larger fraction of the population to find and stop spread. As with Ebola and Zika, this is proving difficult to do with PCR. Current PCR testing for the coronavirus is struggling to keep pace with patients presenting with concerning symptoms, let alone asymptomatic cases. As a result, we are likely detecting as few as 10 percent of all infections. PCR has also been too slow to return results: By the time someone has symptoms, they have likely been infectious for a couple of days, and then results take days more (and sometimes weeks) to come back. Even if these throughput and turnaround times could be improved — for example, with pooled testing and other process innovations — PCR would still require millions of samples to be somehow collected and brought to centralized labs on a regular basis. At best, it would still take hours to get results.
This is where, as others have also recently noted, rapid tests could be transformative. One version of these tests for the coronavirus require people to spit on a strip of paper that turns a different color several minutes later if infection is detected. These tests can be done by people on their own at home. Because they only require paper strips that can be produced at relatively low cost, they can be used frequently — even daily — by all people in areas with covid-19 transmission. Another set of rapid tests are run off small devices that likely cannot be decentralized to each home, but could be used for frequent on-the-spot screening in schools and high-risk venues, such as food-processing plants and other workplaces, that continue to see large outbreaks.
The caveat is that rapid tests may be more prone to giving the wrong result — a false-negative (where the test says you don’t have covid-19 when you do) or a false-positive (where the test says you have it when you don’t). Understandably, this has given regulators and public health authorities pause, particularly regarding the paper strip tests, which can miss as many as half of infections.
But this was the same hang-up that prevented the deployment of Ebola and Zika rapid tests in past outbreaks. It’s true that rapid tests may not be as accurate for definitive diagnosis at an individual level, but they’re still a critical tool for transmission control at a population level. For transmission control, reducing the infectiousness-to-isolation period is paramount, and factors such as speed, frequency and the ability to be decentralized and scaled to screen large numbers of people (e.g., population coverage) are also important.
In effect, when testing is limited as it is now, up to 90 percent of infections go undetected and are effectively “false-negatives” themselves because they are never tested. And when transmission becomes uncontrolled and we have to resort to lockdowns, the entire population is treated as if they are potentially “positive.” Though rapid tests may have more false-negative results than PCR, they would enable screening of people who would otherwise never be tested, or who would only be tested after days of being infectious and exposing others. The higher false-negative rate could be offset by frequent testing, so an infection missed one day may be caught the next. There is also reason to believe that people who are most capable of infecting others — including superspreaders who seem to account for the majority of transmission — harbor larger loads of virus and are less likely to test false-negative.
False-positive results can be addressed by confirming all positive results with PCR, as long as the number of false-positives generated by widespread screening is within the scope of available PCR testing. Depending on how high the false-positive rate is, a test used on tens of millions of people in the context of overall low disease prevalence could generate hundreds of thousands or more of false-positive results a day. In this scenario, PCR capacity would struggle to keep up. Those who screen positive would need to be isolated and confirmatory PCR testing prioritized for people requiring hospitalization where certain diagnoses might affect treatment decisions. Requiring so many people to presumptively isolate or, after, act as if they are now immune would be problematic. But this can be mitigated by prioritizing rapid tests for which this is less of an issue and effectively engaging the public on understanding these implications. In the big picture, isolating even tens of thousands of people is a lesser evil than allowing the epidemic to persist or requiring lockdowns that affect millions. (New testing platforms, such as one that uses CRISPR technology, are in the works and could enable rapid testing that is highly accurate without the need for confirmation with PCR.)
The development and approval of new rapid tests need to be fast-tracked, and measures like the Defense Production Act should be used to accelerate production. As these tests become available, whether for home self-screening or at the workplace, they should be targeted to where transmission is greatest, such as among essential workers, communities of color and counties with high-transmission rates. Local epidemiological data should also be used to determine other location-specific groups who are at higher risk for infection.
In an ideal world, if everyone across the country used rapid self-screening tests every day for two weeks, we could identify and stop most transmission chains. Even if this could happen within the next two to three months in conjunction with preventive measures, such as more protective masks, better ventilation indoors and ongoing social distancing, we could substantially control the epidemic before the winter.
Ultimately, to stop the epidemic, we don’t need to find every infection. If enough transmission chains are detected and stopped, then the epidemic will begin to shrink. In combination with masking and social distancing, this is achievable — and it’s also imperative. We know what to do. Now we need to do it.
The need for rapid testing was clear months ago, and we missed the opportunity to implement this approach sooner, but it is still not too late. Going forward, widespread rapid tests should become the go-to strategy for uncontrolled infectious-disease epidemics when no vaccine is available. The same “warp speed” push that’s accelerating vaccine development should also help encourage rapid testing platforms that can be configured to new pathogens as they arise so that we are always a step ahead of future threats.