False positive

Lateral Flow Tests Too Inaccurate for Mass Testing, Major Review Finds

Lateral flow tests (LFTs) are too inaccurate to be used in mass screening in schools and workplaces, a major review led by Jon Deeks, Professor of Biostatistics at the University of Birmingham, has concluded. The Telegraph has the details.

A Cochrane review of 64 studies found that asymptomatic testing only picked up around 50% of cases, and picked up far more false than real positives when infection rates were low in the community.

False results from lateral flow Covid tests may mean workplaces and schools will be needlessly forced to close even though the scheme was set up to keep them open, experts cautioned.

The group concluded that mass testing policies had been implemented “without any supporting real-world evidence”.

Jon Deeks, the review author and a Professor of Biostatistics at the University of Birmingham, said it was “shocking” that the Government had allowed the rollout of such a “large, invasive and expensive” testing programme without the data to back it up. 

“One of the issues which should have been picked up is these tests work a lot less well in people who are asymptomatic, and the Government has not been clear about that,” he said.

“It is being used in ways which actually risk people getting false reassurance when they miss cases, and getting false positives, and that has been a big problem in schools. There are instances when whole classes have been sent home – effectively, 100 people put into isolation.”

Prof Deeks added: “So I think there are big problems in the way that we’re using this. And when prevalence drops you get more false positives and fewer true positives, that’s a mathematical certainty, and there is a point where you should stop. We may already be at that point.

“When is it that we are actually going to be doing more harm than good? These testing policies have been implemented without any supporting real-world evidence.”

The review calculates that at current virus prevalence rates of 0.5%, around 72% of positives would be false.

In terms of not missing infections, the review found the tests correctly identified only 58% of those who were infected but had no symptoms, and missed 42%. It did not acknowledge, though, that where these asymptomatic infections are identified by PCR tests (as most of them are), the PCR positives are likely to be false or cold positives themselves.

The review found the tests to be more accurate in people with symptoms, estimating the LFTs would correctly identify around 72% of people with symptoms who had Covid, and miss 28%.

The researchers suggested the tests would be better used for symptomatic people, and that any positives should also be re-tested with a PCR test to reduce the chance of false positives (though they didn’t note that PCR tests are more sensitive than LFTs so are more not less likely to pick up cold positives).

Dr Susan Hopkins, PHE’s COVID-19 Strategic Response Director, defended the testing programme, essentially saying every little helps:

Rapid tests are effective at detecting Covid in people that are highly infectious, both with and without symptoms. Every day, rapid testing is helping us find cases of COVID-19 that we wouldn’t otherwise know about, breaking chains of transmission and potentially saving lives.

Mass screening of asymptomatic people to try to pick up invisible, isolated infections is an exercise fraught with pitfalls, as experts have been warning for months and Professor Deeks’ review now confirms. The Government needs to heed the lessons of this review and abandon its costly and invasive effort to make reopening society dependent on constant testing for a disease almost no one has. Ministers need to plot a much clearer and swifter path back to normal – a normal which doesn’t involve testing week in, week out for a disease for which you have no symptoms.

0.06% of School Covid Tests Positive Last Week

In the week that all students returned to school, only 0.06% of rapid Covid tests produced positive results. The TES has the story.

Only 0.06% of the rapid Covid tests carried out in schools in the week that all pupils returned were positive, new statistics reveal.

More than 4.5 million lateral flow tests were taken in schools and colleges in the week between March 4th and March 10th according to Department of Heath statistics. 

But they only picked up 2796 positive cases.

The question is, how many of these positive results really were positive? Earlier this month, Will Jones wrote here that “mass testing in schools was always a bad idea – cruel, intrusive, pointless. The high proportion of false positives, with their needless and disruptive requirement to self-isolate, only underlines that fact”. A report in the Telegraph also explained that infection levels are now so low that the majority of positive results could be false.

Biostaticians are concerned that infections in the community are now so low, that false positives and negatives are vastly outnumbering true cases, leading to real cases being missed and families needlessly being asked to isolate.

Pupils are currently being tested twice a week for coronavirus using lateral flow devices, but real-world data has shown they miss positive cases around 50% of the time.

Worth reading in full.

Stop Press: Figures show that children returning to school has not led to a rise in Covid infections. The Telegraph has the story.

A detailed analysis by Prof Jon Deeks, an expert in biostatistics at Birmingham University, has found that the number of positive cases among pupils are far lower than ministers expected.

“The big question is why is that happening? There are two explanations,” he said. “The first is that this test doesn’t work very well in children and it doesn’t detect the cases. And the other is that asymptomatic infection is much less common in children.

“If the reason is that there aren’t many cases, then this is good news. But if the reason is that the tests are missing cases this is bad news. And if we don’t know – that is very bad news.”

Worth reading in full.

Two Thirds of Covid Cases in Schoolchildren Are False Positives Say Experts

Mass testing in schools as children returned this week is resulting in hundreds of positives – but viral prevalence is now so low that the large majority of them are wrong and result in children and their contacts isolating needlessly. Sarah Knapton in the Telegraph explains.

Biostaticians are concerned that infections in the community are now so low, that false positives and negatives are vastly outnumbering true cases, leading to real cases being missed and families needlessly being asked to isolate.

Pupils are currently being tested twice a week for coronavirus using lateral flow devices, but real-world data has shown they miss positive cases around 50% of the time.

Similarly, although they pick up 99.9% of negative cases, meaning fewer than 0.1% will be false positives, the prevalence rate is now low enough that false positives will be making a significant contribution to the overall figure.

Previously, positive tests were confirmed using a more accurate polymerase chain reaction (PCR) lab test, but this week the Government admitted it had quietly scrapped the process at the end of January.

It means that thousands of pupils and their families are having to isolate needlessly, and missing more school after months away from classrooms.

The latest figures from NHS Test and Trace in secondary schools found 328 positives out of 663,332 tests between February 25th and March 5th. That’s 1 in 2,000 pupils, which is far lower than the 1 in 270 that the latest ONS survey suggests is the current community prevalence. Some are concerned this means large numbers of cases are being missed, though the PCR-based ONS survey picks up many of the non-infectious cold positives which the less sensitive lateral flow tests (LFTs) do not. There is no reason to regard missing cold positives as a problem, particularly given the personal and economic cost of isolating positive cases.

In primary schools, 613 positives were found from 721,546 tests, or about 85 in 100,000. This gives an overall school prevalence rate of about 0.067%.

At this rate (if it were the true prevalence), out of a million tests 670 should come back as true positives. However, with a 0.1% false positive rate (a conservative estimate for LFTs, which were found in a report to have a 0.32% false positive rate) there would also be 1,000 false positives. That means 1,000 out of 1,670 positives, or around 60%, would be false. Of course, the 670 positives per million from the last couple of weeks of testing include the false positives, so the actual proportion of false positives will be higher.

Professor Jon Deeks, a biostatistician from the University of Birmingham, told the Telegraph:

We would expect far more false positives than true positives amongst those testing positives in schools. There are many uncertainties but given the DHSC data it seems likely that over 70% of positive test results are false positives, potentially many more. Addition of a confirmatory PCR would add little cost and would most likely reduce false positives to 1 in 1,000,000. The refusal to confirm lateral flow results with PCR is at best perplexing, will make testing less attractive, and create harm by wrongly isolating individuals, families and other close contacts.

Confirmatory PCR testing for LFT positives is unlikely to help, however, in terms of weeding out cold positives, as the PCR tests are much more sensitive than the LFTs (picking up many more cold positives) and so are unlikely to give a negative where an LFT has come out as a cold positive.

There is also the question of how many of the tests have been incorrectly carried out, perhaps on purpose, if self-administered, to avoid having to isolate?

Mass testing in schools was always a bad idea – cruel, intrusive, pointless. The high proportion of false positives, with their needless and disruptive requirement to self-isolate, only underlines that fact.