The government continues to face criticism over the rate of testing for coronavirus in the UK.

On 18 March, the Department of Health announced it planned to increase capacity to 25,000 a day by the middle of this month. On Tuesday, the government said it wouldn’t meet this deadline, and now expects to reach the target at the end of April.

As FactCheck reported this week, there’s also been confusion from ministers over how many tests are currently happening.

And amid all this, there’s a further question that experts have raised in recent days: how reliable are the results from the current coronavirus tests? We’re going to look at what we know, and what we’ve yet to discover.

Limits of the current tests

Dr James Gill, a locum GP and Honorary Clinical Lecturer at Warwick Medical School, explains: “Currently there are two tests in use, a PCR [polymerase chain reaction] lab test looking directly for the virus, and the newer antibody test which is looking for evidence that the body has been exposed and reacted to the virus”.

The people who show up in the government’s official figures for coronavirus testing have had the PCR laboratory test.

Anthony Woodcock, who’s a Professor of Respiratory Medicine at the University of Manchester, describes this type of test as the “gold standard” – but he and Dr Gill both point out some of the problems that come with it.

For one thing, PCR tests are labour-intensive and take some time to complete – which, as Professor Woodcock says, can leave patients “sat on wards without a firm position on their infectivity for 24 hours”.

And it’s not just waiting for the results. Both point to the issue of “false negatives” – where the test says a patient has not got coronavirus when in fact they have.

Professor Woodcock and Dr Gill both estimate that the rate of false negatives could be around 30 per cent.

Professor Woodcock suggests that these could be caused by “low levels of the virus in the early stages of the disease” or “especially poor technique in sampling or conducting the test”.

He describes the dangers of these “many” false negatives: “potentially infectious patients could be nursed in COVID negative areas, or sent home thinking they are non-infectious”.

Dr Gill says that as the UK’s lab testing rates expand, there will be more automation in the process, which should in theory reduce the chance of errors and incorrect results.

Patients are being tested more than once

Some patients are receiving more than one coronavirus test in the course of a single day.

We can see this in figures tweeted out by Public Health England earlier in the week that show some 8,278 tests were carried out on one particular day on just 4,908 individuals.

The latest stats show 10,657 tests were carried out yesterday on 7,771 patients.

We don’t know exactly why each of these patients would have had the number of tests they did.

But as Professor Woodcock points out, if the current tests can only detect the presence of the virus 70 per cent of the time, “we have a lot of patients waiting for a repeat virus testing/diagnosis”.

In other words, it’s possible that some of those repeat tests are happening to confirm what doctors believe to be a case of coronavirus that just hasn’t been picked up by the first or second tests.

A spokesperson for Public Health England told FactCheck: “‘It would be unfair to say that the test is not very reliable, there will be numerous reasons patients in inpatient settings will need to be retested throughout their patient journey.”

What about other types of test?

Professor Woodcock says the current process of PCR testing will “continue to be needed”, but that in the next two to four weeks, we could move towards “point of care” testing. This could see patients getting results at their bedsides within 15 minutes, rather than waiting hours or days for central labs to report back.

For that to work, PCR machines would probably need to be moved into hospital emergency departments, he says – “however, local PCR machines are not yet widely available”.

Colin Butter, Associate Professor of Bioveterinary Science at the University of Lincoln says the PCR test was the “correct first step” in terms of the world’s response to the coronavirus outbreak, but that “a much better and faster test would be one that detects a viral protein”.

This, he says, involves making a “monoclonal antibody against the relevant protein” – a process that he estimates would take about two or three months. Once developed, you would not need any specialist machinery and the results could be available within 15 minutes.

So far we’ve looked at tests that try to establish whether a patient currently has the virus. There is another category of test – sometimes called an “antibody” or a “serology” test – which can reveal if the person being tested has ever had coronavirus.

Eleanor Riley, Professor of Immunology and Infectious Disease at the University of Edinburgh, explains that these tests “use viral proteins as ‘glue’ to trap the antibodies present in serum”.

Professor Woodcock says that among the range of antibody tests, “the ‘stick’ test will be invaluable” to help with fast diagnosis, and to allow healthcare staff to return to work if they can confirm they’ve already had the virus.

The UK government has ordered millions of these tests – though there’s some uncertainty around exactly when they could be rolled out.