Public Health England’s counts of new daily coronavirus cases still includes 30,000 people who were counted twice before July, MailOnline can reveal.
Although the total number of people infected has been adjusted down to 313,798, as of yesterday, the sum of all the daily increases amounts to 344,384.
When the Government split its statistics into ‘pillar one’ and ‘pillar two’ tests, 30,000 Covid-19 patients were found to have been counted twice and the overall tally had to be redone.
Although the grand total was changed six weeks ago, officials have been unable to show exactly where and when the duplicate tests happened.
It is the latest in a string of issues with Government data about the pandemic, with it emerging that Public Health England was recording anyone who ever tested positive as a victim, even if they died in a car crash months after recovering.
There are now five separate measures for fatalities across the UK and experts have dubbed the handling of official statistics ‘confusing’.
And today it emerged 1.3million counted tests have been removed from the official government total because they were counted twice.
Case figures are just an ‘estimate’ of the real prevalence of the virus, accounting for around a third of the real number of infections because thousands of patients will never develop any symptoms.
But the numbers are still critical for understanding the progress of the outbreak and letting experts analyse the size of the outbreak.
And the daily updates released by the Government will become ever less accurate as the number of people catching the disease gets lower because inevitable false positive results will skew the number to look higher than it really is, experts say.
Public Health England has not adjusted the number of people being diagnosed each day and has shaved off 30,000 positive tests that were counted twice, but it is not clear when or where the people were double-counted
Professor Carl Heneghan, an evidence-based medicine expert at the University of Oxford, has followed the Government’s data closely throughout the epidemic.
He said the issue appears to have come from when the Department of Heath split testing into pillars one and two.
Pillar one refers to tests done in hospitals and medical facilities while pillar two is members of the public who are tested in drive-through, walk-in or home tests.
Professor Heneghan told MailOnline: ‘There is seemingly a problem when you start to introduce pillar one and pillar two tests – they seemed to be double counting tests.
‘Somebody would have a pillar two test and then gone into hospital and had a pillar one test, and they thought it was two people.’
He said it was unsurprising that data errors were creeping and that some allowances should be made because of a difficult situation, but that it is ‘vital’ that numbers are correct.
Professor Heneghan said: ‘If the number of cases is wrong, the case fatality rate and everything gets skewed.
‘It is vital they’re correct but, to be honest, it doesn’t surprise me there have been areas where you’ve had discrepancies that need to be corrected.’
He added: ‘It does concern me and I think it’s important that data and epidemiology is transparent and it’s clear that [decisions] actually are based on up to date info.
‘What we’re interested in is understanding trends, and information has to be correct for that.’
Numbers of people diagnosed with the coronavirus are updated each day by the Department of Health and Public Health England.
While the Department of Health is responsible for organising the tests, carrying them out and reporting the results back, PHE controls the statistics.
On July 2 officials consolidated their counts of the cases and as a result the total number of people diagnosed dropped by 31,388, from 315,145 to 283,757.
This shows more than 31,000 people had been counted twice in the Government’s testing system up to the end of June.
But despite the total number being updated, the daily numbers were never changed.
The data still shows that May 1 had the highest number of cases of any day in the epidemic so far, with 6,201, and that 33,760 people were diagnosed in that week, from April 27 to May 3.
The daily new numbers of cases in England add up to a total 344,384, with a peak of 6,201 people getting diagnosed on May 1
But the total number of cases was adjusted on July 2, where a drop of more than 30,000 cases can be seen. No further details about the decline have been offered
But the true number of cases diagnosed in that week cannot be confirmed because hundreds or thousands of those people could have been counted twice.
Dr Simon Clarke, a cellular microbiologist at the University of Reading, said the official data was ‘a bit of mess’.
‘My understanding is it’s still basically an estimate,’ he said. ‘It also needs to be remembered that the diagnoses is way below what the [Office for National Statistics] thinks the real numbers of cases are…
‘I think it is confusing for people. A lot of people think the numbers they see are it – that that is how many cases there are – but it’s not.’
The discrepancy is just one in a string of issues with the Government’s coronavirus data.
Officials this week confirmed they were changing the way they count the number of people who are dying because PHE had been including anyone who died of any cause after a positive test for Covid-19.
Even someone who tested positive in March, then recovered and died after getting hit by a bus in August would have been included in the Covid-19 death toll.
Now the Department of Health is including only people who die within 28 days of their diagnosis, in an attempt to reduce the number of people included wrongly.
NHS England and the devolved governments in Wales, Scotland and Northern Ireland have been using the 28-day cutoff throughout the outbreak.
As a result of this change the official death count dropped by 5,000 this Tuesday, from 46,706 to 41,329.
PHE will still publish weekly stats showing how many people died within 60 days of a positive test, to try and include people who have long stays in hospital or long-lasting Covid-19 illness.
And it emerged today that the Department of Health has wiped 1.3million swab tests off its count of the number of tests done – a staggering 10 per cent of the total.
The Government has repeatedly boasted about the number of tests it is capable of doing – with capacity now reportedly at around 340,000 per day – but it has now emerged one in 10 tests have been counted twice, The Guardian reported.
The double-counted tests had been done between May 14 and August 12, but the total number of tests ‘made available’ before and after the adjustment is unclear.
The Government describes tests as ‘made available’ because it includes tests sent to people’s homes that are counted whether someone takes the test or not.
It says a total 13,785,297 tests had been completed by yesterday, Wednesday August 12.
In a statement on the website on Wednesday the Department said: ‘An adjustment of -1,308,071 has been made to the historic data for the “tests made available” metric.
‘The adjustments have been made as a result of more accurate data collection and reporting processes recently being adopted within pillar 2.’
The error reportedly came to light in July but was not corrected until this week.