Last week, the UK government agreed a deal to acquire a large number of coronavirus antibody tests for NHS workers.
Antibody tests, which check whether you have ever had coronavirus, are useful for two main reasons:
- At an individual level, they can be used to identify people who have recovered from the virus (even if they never developed symptoms at the time) – and who may therefore be at lower risk for contracting or spreading the virus in future.
- At a population level, they can be used to estimate what proportion of a population has had coronavirus so far (as has recently been done in Spain, France, and New York).
What does a positive antibody test result imply?
It is true that no study has yet decisively demonstrated that people become immune to COVID-19 for a prolonged period after recovering from it. But if we had to make a prediction despite the current uncertainty, we might draw on research on SARS-CoV-1, which found that immunity to that virus lasted for at least a year, and on more recent research on COVID-19, which has found that people who get the virus tend to have immune responses which suggest subsequent immunity. The forecasters on the prediction website Metaculus have attempted to use evidence like this to formulate an explicit quantitative prediction – they think there is an 86% chance that people who recover from coronavirus will not be severely affected by it again in future.
So, if many more people in the UK are set to begin receiving antibody tests in coming weeks, we need to consider how to communicate their results in a way which best conveys our current state of knowledge – that having had coronavirus likely does reduce people’s risk of catching or spreading it in future, but may not eliminate the risk completely (not least because some antibody tests give unreliable results).
Results from an online experiment
To help address this issue, we ran an online experiment in April, involving 6,149 UK adults, to investigate how the framing of a positive antibody test result affects the public’s perceived risk and behaviour. The experiment was designed by Professor Dame Theresa Marteau, Dr Jo Waller and Dr James Rubin, and run and analysed by BIT.
Participants saw one of eight descriptions of the test result, shown below – these either varied the terms describing the test type (Immunity vs Antibody) and test result (Passport vs Certificate vs Result), or used a more descriptive framing (“a result showing a high level of antibodies” vs “a result showing they were low risk for catching coronavirus in the future”). The main hypothesis was that people would be more likely to think there was no risk of re-infection when the word “immunity” was used in the test description.
The key results, as shown in the graph below, show that only 2% of people thought they had ‘zero’ risk of catching coronavirus again when the type of test was framed as “antibody”, but this jumped to 6~10% when it was framed as “immunity”. Less than 1% of people thought they had zero risk when they read the more descriptive framings that said they had a “high level of antibodies” or “low risk for catching coronavirus again in future”.
We also investigated whether people might become less diligent in following other government guidance after receiving a positive test result. We found little evidence for this for hand-washing (only 4% said they would wash their hands less often if they got a positive test result), but 18% of people said they would meet others more often if they got a positive result – rising to 21% for people who got a result which used the word ‘immune’. These and other findings from this research will be more fully documented in forthcoming academic publications (1 and 2).
Even if people develop some immunity to coronavirus after recovering from it, it is still unclear how governments should use this information. It could be used to allow people to return to the workforce, or to see family and friends. But this may be unfair or even create perverse incentives where some people actively try to catch the virus in order to gain these extra freedoms. So far, the US CDC has said positive antibody test results should not be used to regulate admission to schools or workplaces; South Australia has gone a step further and banned the use of antibody tests by the public. Meanwhile China and India are taking a different approach by seeking to create individualised coronavirus risk scores based on people’s symptoms, health and travel history.
As far as we’re aware, no country has actually issued the kind of ‘immunity passports’ being discussed in the media. However, more than 200,000 people in the UK have now recovered from Covid-19 — and as these people begin receiving positive antibody test results, some may begin behaving as though they cannot catch the virus or spread it to others. The findings of this experiment suggest that avoiding the use of the word “immune” in these antibody test results will help prevent people behaving in these potentially risky ways.