The real death toll from coronavirus in the UK is already more than 40,000, a new analysis suggests – twice the total once described as “a good result”.

Ministers announced on Tuesday that there have been 17,337 fatalities, but this counts only those who have died in hospitals after testing positive for the virus.

Meanwhile, the office for national statistics (ONS) has found the number of registered deaths in the week ending 10 April was 75 per cent above normal in England and Wales, at 18,516.

Now the Financial Times has used that figure to calculate the likely number of “excess deaths” since the coronavirus struck the UK – concluding it could be as many as 41,000.

As 24 per cent of deaths normally occur in care homes, the analysis suggests that just under 11,000 more people than normal have died in residential care, probably from Covid-19.

Carl Henegan, professor of evidence based medicine at Oxford University, said: “I don’t think we’ve ever seen such a sharp upturn in deaths at that rate.”

He said the seasonal flu outbreak in 2017-18 may have killed 50,000 people in the UK, but those deaths “were spread out over many weeks”.

Last month, Stephen Powis, the medical director at NHS England, said 20,000 coronavirus deaths would be “a good result” – when an Imperial College study predicted as few as 5,700.

Excess deaths from all causes stand at 16,952 above the seasonal average across the UK since fatalities from Covid-19, the disease caused by the virus, began to mount in mid-March.

But, because of the lag in collating death registrations data, the ONS figures cover only the period to 10 April 10 are significantly out of date.

Assuming the relationship between hospital deaths and excess deaths has remained stable since, the FT’s estimate of total deaths from the virus by 21 April 21 is 41,102.

They are made up of almost 38,000 deaths in England and Wales, just under 3,000 in Scotland and just below 500 in Northern Ireland.

The ONS has described the number of recorded deaths as “unprecedented”, especially given the sunny and warm weather in the run-up to the Easter weekend.

The 18,516 fatalities in the week to 10 April compares with the most recent five-year average of 10,520 for the same week of the year.

David Spiegelhalter, the Winton professor of public understanding of risk at Cambridge university, backed the analysis as “the only unbiased comparison” given the problems measuring deaths and their causes.

He said coronavirus was not given as the cause on many of the death certificates, but was likely to be a direct or indirect factor.

Many doctors would initially have been reluctant to designate the virus as the cause on death certificates as it was a new disease and they could not have been certain.

Some of those who died from other causes may have been too scared to attend hospital, or did not want to be a burden on the health service – so they could be seen as possible indirect victims of the virus.