The UK is about to enter its third week of lockdown, with the government poised to extend the measures beyond next week. But while the worst of the coronavirus pandemic is still ahead of us, ministers have long been touting antibody tests as one possible route back to normality.
In theory, a coronavirus antibody test would be able to detect whether someone had previously been exposed to the virus that causes Covid-19. Although we don’t yet know whether people who have had coronavirus have long-term immunity against the disease, this is the case with lots of other viruses, and that’s why ministers hope that antibody tests could be vital for knowing which NHS workers and key staff are still vulnerable to the disease.
The UK has flirted with the idea of antibody tests for some time, but as the days have turned into weeks the government’s initial gung-ho attitude towards the tests has been undermined by the tricky reality of developing and manufacturing these tests. On March 19, prime minister Boris Johnson claimed that “antibody tests are coming soon” – but no timetable was given. Later, this figure was upped to 17.5 million with still no sense of a date when they’d be used. At a meeting before MPs on the science and technology select committee it emerged that none of the tests currently being considered by the government are good enough to rely on and that reliable tests won’t be around until May. So what’s the hold up?
When people talk about Covid-19 tests, they roughly split into two categories. One detects for the presence of the virus and can only tell you if a person is currently infected with Covid-19. These are called polymerase chain reaction (PCR) or antigen tests and it’s these tests that are behind the daily figures of confirmed coronavirus cases you see in the news. Such tests rely on a nose swab to pick up virus cells that are then detected in the test.
An antibody test is a blood test, most likely a finger prick kit similar to home STD tests, and is carried out after you’ve been ill. When your body is exposed to a foreign pathogen, it produces antibodies that help fight the invader next time it enters your body before it has a chance to make you ill. Antibodies are a little like your body’s immune memory, although some of those memories stick around for a lifetime while others fade after a few months, which is why you can suffer from the same cold over and over again.
But it takes time for your body’s antibody response to kick in. That’s why antibody tests are best carried out at least 28 days after infection, and won’t show whether someone currently has the disease. “With the PCR test, you had a relatively narrow window – a couple of weeks, probably – where you can go back retrospectively to someone and use and find out whether they’ve had the virus,” says Andrew Preston, a reader in bacterial pathogenesis at Bath University. “The antibodies will hang around for months, and in some cases, possibly years, so you've got a much longer window.”
Though we don’t yet know the specifics of this particular coronavirus these antibodies could last in the body for a lifetime, rendering their host immune. “In 99 per cent, if not more, of infectious diseases if you’ve had the pathogen and you’ve got antibodies, that suggests that you’ve mounted an immune response,” says Preston. “I think we’d be pretty unlucky, in terms of Covid-19, if that didn’t mean you had a period where you were immune to another episode of disease. That doesn’t mean then that you’re not possibly going to pick up the virus again, but we’d suspect if you did, any symptoms you have even if you hadn’t first time round would be remarkably more mild.”
The theoretical benefits of this insight are huge. A clean bill of immunity would reassure key workers that they weren’t at a risk of infection – hospital staff, of course, but also bus drivers, supermarket attendants, and employees of infrastructure and utilities providers.
On a larger scale, we would get a clearer picture of how many people had been infected, particularly among those who were asymptomatic. From this, we would work out a clearer mortality rate. Theoretically, immunity passports could be issued on the basis of this positive test, and we could return, in Matt Hancock’s words, to “normal life”. “Even if it’s at the end of April you could, in theory, retrospectively test everybody in the country to know exactly how many people have had it at what stage we’re at,” says Preston. “And that then obviously has huge implications for these models and discussions of herd immunity.”
But things are not that simple. The hold-up isn’t that there aren’t any tests: Singapore, China, and other countries have conducted limited numbers of antibody testing, while American companies like Biomerica and Chembio Diagnostics are selling antibody tests outside the US. You can buy one right now, should you so choose.
The issue that these existing tests are unacceptably unreliable. The best tests we have at the moment are reportedly 90 per cent accurate. This might sound high, but testing is a little more complicated than that. There are two really important measures when it comes to testing: sensitivity and specificity. Sensitivity refers to how well a test correctly identifies those who have had Covid-19: the true positive rate. While the specificity refers to how well the test correctly identifies those who haven’t had a disease: the true negative rate.
While a test that is 90 per cent sensitive and 90 per cent specific may sound like it has a small margin of error, the maths turns out to be quite different. False positives are particularly grievous when we don’t know how many people in a given population have had Covid-19. “They just don’t pass the muster”, says Robert West, a professor of health psychology at University College London.
Let’s say, explains West, that the test correctly picks up 95 per cent of people who are positive and have had Covid-19 and correctly detects 95 per cent of people who are antibody negative and have never had the disease. In a scenario where five per cent of people have had Covid-19, in a population of 1,000 people, 50 will be antibody positive. If the test had perfect sensitivity it would pick up every one of those 50 people. But our test is only 95 per cent sensitivity, so it’ll actually detect 48 people (technically 47.5, but you can’t test half a person) as having had the disease. Two people who have had the disease will actually be told that they haven’t had it.
Now let’s turn to specificity. Remember, in our fictitious population 950 people are really antibody negative – they’ve never had Covid-19. If our test is 95 per cent specific, then of those 950 people who really are antibody negative, 903 of them will get a correct negative test. That leaves 47 people who will receive a positive test when they’re really negative – people who could think that they can return to work with no risk of contracting Covid-19 when they’re actually as at-risk as any of the other 950 people without antibodies against the virus.
“Of the people who test positive 47 will actually be positive but 47 will actually be negative – 50 per cent false positives,” says West. “And the reason is that there are so many negative people even a small error rate yields a large number of false positives.” In other words, because we think that only a very small percentage of people have ever had coronavirus, we need tests to be very specific otherwise they will end up overestimating the number of people who are immune to Covid-19.
It’s only when the prevalence goes up that you can have a high confidence that a positive test result is real. “Another thing to bear in mind is that one can also add other information to change the base rate,” he says “For example, whether or not the person has had symptoms of Covid-19 and whether they are in a geographical hotspot.”
Couple this with other blind spots – like not knowing how long immunity for this particular virus lasts – and it isn’t just hyperbole when officials say that a poorly performing test is far more dangerous than no test at all. “You’d be sending people into the line of fire with a false premise of being protected, and it also completely messes up our understanding of where we’re at with the pandemic,” says Preston.
Even operating under the assumption that the UK procures an acceptably accurate test, the technical challenges of rolling them out to the public are big, too. “Because they started so late, the UK government is at the back of the queue for getting some of the reagents and other materials that are needed,” says Susan Michie, a professor of health psychology at University College London who is also a member of the scientific group advising the government’s response to Covid-19. (This should come as no surprise, as the UK has flopped repeatedly at upping its testing rates.)
A related point, says Michie, is that so much public health service infrastructure and staffing have been cut over the last ten years. “There’s not been the lab potential for R&D and technological developments, that there could and should be in terms of having additional capacity,” she says. “Running a health service and running things like Public Health England and the testing laboratories on a kind of just in time principle, it’s been shown to be totally inadequate. Just as we don’t have capacity within the health service for clinical cases, so we don’t have capacity within the lab.”
The third problem is a mix of the two above. When the government realised that capacity and reagents were lacking, they should have sought to galvanise universities and the private and public sector, says Michie. And this should have occurred months ago. “It’s just like with PPE and ventilators and the rest of it – dragging heels, not anticipating properly, what problems and what capacity issues they’re likely to be, and dealing with it,” she says.
This lack of infrastructure capacity even plays into the reliability of results as you scale the tests up, explains West. “You have to not only be confident that in the perfect condition for the test they will operate as required, but also that when you scale it up and put it out into all sorts of different labs, with standards and different levels of training, that it will stand up,” he says.
If it’s a home test kit, there are going to be numerous issues with this too. Tests done in hospital pathology departments are read by professionals. Tests done at home may require the person to self-certify. “I’m still nervous about giving those out for home use,” says Martin Hibberd, a professor at the London School of Hygiene and Tropical Medicine. “We’d be taking people at their word of what the result means. Pregnancy tests are famously easy to read, but people still misread them.”
Validating these tests could be a nightmare. “Different tests will probably have different sensitivities,” says Preston. “So does that mean we need to validate a range of tests to give us the ability to churn out millions of these things? Or do we need to find one test and then worry about how many of them we can get hold of?”
The hype around antibody tests is understandable. It’s evidently because no government has, as of yet, devised a clear exit strategy. But while these tests may well play a part in this escape, they’re not there yet. And misplacing our faith in the tests could have disastrous consequences. “The concern is that you get a test – and the pressure from the business sector to get the economy going means it’s seen as a silver bullet, restrictions are eased, it doesn’t pan out as expected, and then we’re in a bigger mess,” says Mackie.
Will Bedingfield is a staff writer for WIRED. He tweets from @WillBedingfield
😓 How did coronavirus start and what happens next?
❓ The UK's job retention furlough scheme, explained
💲 Can Universal Basic Income help fight coronavirus?
🎲 Best video and board games for self-isolating couples
👉 Follow WIRED on Twitter, Instagram, Facebook and LinkedIn
This article was originally published by WIRED UK
