In the middle of March, when the pandemic was still in its early phases in most countries, the director of the World Health Organization, Tedros Adhanom Ghebreyesus, very clearly stated what would be the key element in the fight against the pandemic: “Test, test, test.”
Almost a year later, with the impressive progress in developing vaccines against COVID-19, part of the focus is changing. Countries are in the middle of a race to vaccinate the population, a race that is becoming more urgent in a context in which the virus is mutating, with new more contagious variants emerging in the United Kingdom, South Africa, and Brazil.
But there is a problem: There is insufficient vaccine, and it will take a long time to produce enough doses to reach the much-sought-after herd immunity. In a context of supply restrictions, the only option is to prioritize. While this is true in developed countries with more access to the vaccine, it is even more important in developing economies that do not have the negotiating clout to be able to purchase a large amount of vaccine in a short period of time.
Therefore, countries are using different prioritization strategies to allocate scarce vaccine supplies in order to reduce the number of deaths to a minimum and reach herd immunity as soon as possible. These strategies generally focus on the type of activity individuals perform (prioritizing healthcare personnel, elderly home care workers, and other essential activities); on older people, and on people with pre-existing conditions that increase the risk of death should they become infected. But could the “test, test, test” strategy play some role in this prioritization?
A recent article in Science suggests it can. More precisely, the article argues that the use of antibody tests (also called serological tests) could be a key complement to a prioritization strategy for allocating vaccinations.
The reason is clear: Just as the vaccine provides immunity, so does exposure to the virus. Although there is some uncertainty surrounding how long exactly the immunity lasts, the expectation is that those who have been vaccinated and those who have already been exposed have a much lower probability of contracting the virus and of spreading it. This is precisely how vaccination and exposure of a significant portion of society to the virus confers herd immunity.
Given the scarcity of vaccine, it does not make sense to allocate it to individuals who already have immunity. Doing so would be a waste of resources, as a vaccination would have no effect: neither on the objective of protecting the individual in question nor on the objective of speeding the transition to collective immunity.
This leads to the following proposal: Perform serological tests on individuals from priority groups whose turn it is to receive the vaccine. If they do not have antibodies, they can proceed with the vaccination. If they do have antibodies, they will be vaccinated later, and the doses can be allocated to other individuals of the same priority level.
How important is this? How could including this test as a requirement contribute to accelerating collective immunity? The answer is: It depends. Let’s illustrate this with a numerical example. Imagine a country where 10% of the population has been exposed to the virus and therefore has immunity. If prior exposure does not depend on age or occupation such that everyone has the same probability of having been infected, the use of serological tests prior to vaccination would save 10% of the vaccine. If there were no restrictions to the supply of vaccine, this would simply mean less expenditure on vaccine and on the corresponding resources necessary (human resources, logistics, etc.) to administer it. This doesn’t seem that important.
However, given the restrictions to the vaccine supply, the cost of not doing it is much higher. It means taking longer (10% longer, if vaccine administration is linear) to cross the threshold of herd immunity. It means more infections, more hospitalizations, and more deaths. It may also mean longer confinement, more days without in-person classes, the closure of more businesses, and significant harm to economies.
Now let’s imagine a country where the infection rate is 30%. In the absence of a serological test, three of every 10 people vaccinated would already have immunity. The benefits of testing prior to vaccination triple! It is for this reason that the study in Science concludes that “preferentially vaccinating seronegative individuals yielded large additional reductions in cumulative incidence and mortality in locations with higher seroprevalence and modest reductions in locations with low seroprevalence.”
Are there any reasons to not do this?
One possible objection is that although it is a good idea to de-prioritize those who have already been infected, the test is unnecessary. It should be enough to de-prioritize individuals who have had confirmed cases. This would provide many of the same benefits without any of the costs associated with deploying the test.
But there is a problem with this argument: Confirmed cases are only a fraction of the real number of infections, even in countries like the United States that have been able to deploy a relatively large number of tests (more than 900,000 per million residents, with a positivity rate of 8.5%). For example, the COVID-19–Projections model estimates that in the United States, for each confirmed case, a total of 2.8 individuals have had the virus. This would suggest that using confirmed cases would only produce one-thirds of the benefits.
It would be even worse in the countries of Latin America and the Caribbean, where test deployment has been much lower. For example, Brazil has done 135,000 tests per million residents, with positivity rates above 30%. In Mexico, the number of tests per million is even lower, with positivity rates close to 40%. Under these conditions, the degree to which the number of confirmed cases underestimates the prevalence of the virus could be much greater. In fact, toward the end of July, an infectious disease specialist from the Johns Hopkins School of Public Health estimated that Mexico could have had more than 7 million cases, although official counts place the number at under half a million.
A second objection has to do with the trustworthiness of antibody tests. Clearly, to capture all the benefits of this proposal, antibody tests must be trustworthy. There are two types of testing errors: false negatives and false positives. False negatives are not a big problem, as they simply mean that someone who should not be vaccinated ends up getting a dose. Unless the percentage of false negatives is very high, the proposal would still be effective, although less so.
The problem of false positives is more serious, as it would mean not vaccinating individuals who do need to be vaccinated. When the tests are designed, there is generally a trade-off between sensitivity (the test’s capacity to correctly identify individuals who have had the illness) and specificity (the capacity to correctly identify individuals who have not been exposed to the virus). For the purposes of this proposal, it is important to use tests with a high rate of specificity (to avoid false positives) although they will be a little less sensitive. When they were first introduced to the market, COVID-19 antibody tests had accuracy problems. However, this article in The Lancet Infectious Diseases suggests that several of the most commonly used tests have high rates of reliability, with accuracy above 98%.
A third objection has to do with the fact that being exposed to the virus may not provide full immunity, particularly from the new variants of the virus. Thus, individuals with antibodies may need the vaccination to prevent reinfection. There is much we do not know about the new variants and the possibility of reinfection. One study of more than 20,000 hospital workers in the United Kingdom suggests that reinfection is rare. Of the more than 6,600 participants with antibodies or positive PCR tests (the positive group) 42 cases were found of “possible” reinfection and 2 cases of “probable” reinfection. In contrast, in the more than 14,000 participants in the negative group, 409 infections were detected. The study’s conclusion is that having been infected reduces the probability of reinfection by 83%. The study is particularly relevant because it uses data through the end of November, when a significant portion of infections in the United Kingdom were already coming from the new strain. On the other hand, the new outbreak of infections in Manaus, where it was thought that herd immunity had been reached, is raising concerns.
Unfortunately, the most important decisions when it comes to controlling the pandemic have to be made despite uncertainty. It is crucial to follow the science, but science still does not have all the answers, and in prioritizing vaccinations, we can’t wait for them. The decisions have to be made based on the information available now. And that information suggests that at a minimum, infection is much less likely in individuals who have already been exposed. A good way to think about this issue is with the following hypothetical case: imagine your parents are in their 70s and both of them are in good health, but one of them has had COVID-19. If you can only vaccinate one of them, which will it be? The answer is clear.
The final objection is that the cost of deploying and administering the serological tests would be too high, particularly in countries with fiscal restrictions like those of the region.
This argument is not sound. Serological tests are faster, cheaper, and easier to deploy than PCR diagnostic tests. Many antibody tests provide results in 15 minutes. Although there is no data on the prices of these tests, the WHO reference for purchasing rapid antibody tests—which likely have a similar cost—is US$5 per test. In comparison, according to this article, the U.S. is paying US$19.50 per dose for the BioNTech-Pfizer vaccine and US$15 per dose for the Moderna vaccine (although the Oxford-AstraZeneca vaccine is cheaper because the company has opted to sell it at cost). All of this would suggest that adding the antibody test to the vaccination process should not be too complicated or too expensive.
It is true that so far, most of the countries of Latin America and the Caribbean have not tested enough. The message “test, test, test” therefore remains extremely relevant. If combining testing with the vaccination strategy will help save lives, achieve herd immunity faster, and reopen economies more safely, the question should not be whether countries can afford to do it. It should be whether they can afford not to.