The weekly Covid-19 infection survey published by the Office for National Statistics (ONS), which the Government said today said will continue in a “scaled down” form in the future, is the most reliable measure of the prevalence of coronavirus in the UK.

It is based on nose and throat swabs taken from a representative sample of tens of thousands of people in private households.

The same people are sampled regularly, whether or not they know they have Covid-19 or have separately reported a positive result.

The ONS then uses the results of the survey to estimate the percentage of people across the UK who are likely to test positive for Covid-19 at any one point in time, regardless of when they caught the virus, how many times they have had it and whether they have symptoms.

As such it is a far more accurate reflection of the scale of coronavirus in the UK than the number of cases published each day on the Government’s coronavirus dashboard, which include only those people who have reported themselves as testing positive for Covid-19.

The dashboard figures are therefore affected by how many people are coming forward for tests or who are taking a test because they know they have coronavirus symptoms.

The latest infection survey shows that around one in 20 people in private households in England had Covid-19 in the week to February 12, or 2.4 million people.

This is down from one in 19, or 2.8 million people, in the week to February 5.

Infections are still higher than they were before Christmas, when the estimate stood at one in 25, or two million people.

But levels have fallen since the peak of the Omicron wave at the start of the year, when a record one in 15 in England had Covid-19, the equivalent of 3.7 million people.

As well as measuring the overall prevalence of coronavirus, the ONS survey also estimates the number of new infections per day.

The difference between these estimates and the daily figures published on the Government’s dashboard has been getting wider in recent weeks.

An average of 99,900 cases of coronavirus per day were recorded from January 23-29, according to the dashboard.

But the true total was likely to be three-and-a-half times this figure, at 357,200 a day, according to the ONS.

The estimate for the week ending January 22 was nearly three times the dashboard average, while at the start of January it was around twice the number.

The changes announced by the Government on Monday to the rules concerning Covid-19 testing and self-isolating in England will further affect the daily dashboard figures.

With staff and students no longer encouraged to take twice-weekly asymptomatic tests, fully-vaccinated close contacts no longer asked to test daily for seven days, an end to routine contact tracing, plus no legal requirement to self-isolate following a positive test, fewer positive cases are likely to be detected and reported.

This means the number of cases announced each day will become steadily less representative of the scale of coronavirus in the population, particularly once free universal testing stops altogether on April 1.

Because of this, the ONS infection survey, even in “scaled down” form, will take on even greater importance in the future.

It can offer a far more accurate and comprehensive picture of the prevalence of coronavirus across the nations and regions of the UK, as well as the true number of infections, in turn helping to identify and track the arrival and spread of new variants.