How the UK Tracks COVID-19 Now That Mass Testing Has Ended

TSBy Tayyab Sarwar3 September 2026Statistics & DataUK News

During the first two years of the pandemic, the UK had something close to a live readout of infection levels. Millions of tests were being reported daily, a large household survey was swabbing people at random and case numbers appeared on the news every evening. That system was expensive, so most of it was wound down. What replaced it is quieter, smaller and less precise, which is why the answer to "how much COVID is about right now?" is harder to give than it used to be.

Why case counts stopped being meaningful

Reported case numbers only ever counted people who tested and reported the result. Once free universal testing ended, that population shrank dramatically and became unrepresentative, weighted towards people unwell enough to seek care or working somewhere that still required testing. A rise or fall in reported cases after that point could reflect a change in testing behaviour rather than a change in infections. This is the single most important thing to understand when reading COVID-19 figures now. It is why our UK statistics page leans on other indicators.

Hospital admissions

Admissions data is now one of the main signals. It has the advantage of being collected consistently through routine NHS reporting rather than depending on the public choosing to test. It also has clear limitations. It lags infections by a week or more, since people are admitted after being unwell for some days. It also mixes together patients admitted because of COVID-19 with patients admitted for something else who happen to test positive, a distinction that matters more now that a large share of infections are mild.

Wastewater monitoring

People infected with SARS-CoV-2 shed fragments of the virus in their stool, whether or not they have symptoms and whether or not they ever take a test. Sampling untreated sewage at treatment works therefore gives a rough measure of how much virus is circulating in the population served by that site. It cannot be skewed by testing behaviour, it captures asymptomatic infections and it is comparatively cheap to run.

The trade-offs are real. Wastewater data gives trends rather than case numbers, so it can tell you that levels are rising in a region without telling you how many people are infected. Results are affected by rainfall diluting the sample, by how many people are actually in the catchment at the time and by the sampling schedule. UK wastewater programmes have been expanded, reduced then restructured several times since 2020, so coverage has not been constant. Read it as a directional indicator rather than a precise count.

Sentinel surveillance and swabbing schemes

Rather than testing everyone, sentinel surveillance tests a defined sample. In the UK this includes swabbing through selected GP practices and testing samples from patients admitted to hospital with respiratory illness. Because the same sources report continuously, the proportion of samples testing positive is a fair indicator of the trend even though the absolute numbers are small. The same infrastructure has long been used to track flu, which is one reason COVID-19 and flu are now reported together in the same respiratory surveillance publications.

Genomic sequencing

Sequencing a subset of positive samples is how new variants are spotted and tracked. The volume of sequencing has fallen a long way from its peak, which means new variants are typically identified somewhat later than they would have been in 2021 and rare lineages may be missed. It still works well enough to detect a variant that is growing quickly, which is how the lineages covered in our explainers on XFG and the BA.3.2 Cicada variant came to public attention.

What the different signals are good for

None of these measures answers every question, so it is worth knowing which one to reach for. Wastewater sampling responds earliest, because it picks up infection before people are unwell enough to reach a hospital, which makes it the better signal for spotting the start of a rise. Sentinel swabbing through GP practices sits somewhere in the middle, showing how much of the respiratory illness in the community is actually COVID-19 rather than something else. Hospital admissions arrive last but carry the most weight for planning, since they measure the burden that services actually have to absorb. Genomic sequencing answers a different question entirely, which is not how much virus there is but which version of it people are catching. Read together they give a reasonably coherent picture, though each one on its own is easy to over-interpret.

Where the published figures live

The main UK source is the national flu and COVID-19 surveillance reports published by the UK Health Security Agency, which bring the different data streams into one weekly document. Registered deaths involving COVID-19 are published separately by the Office for National Statistics. Reading the notes attached to these publications is genuinely worthwhile, because they state plainly what each measure does and does not capture.

How to read the numbers sensibly

A few habits help. Compare like with like over time rather than comparing a current figure against a 2021 peak produced by a completely different testing system. Treat direction of travel as more reliable than any single week's value. Expect revisions, since late reporting is normal. Be wary of any chart that puts current reported cases on the same axis as 2021 reported cases without explaining that the underlying collection method changed, because that comparison is close to meaningless.

What was lost, plus what it means for you

The household infection survey that ran during the pandemic was unusual internationally in that it tested a random sample of the population regardless of symptoms, which made it possible to estimate actual infection prevalence rather than infer it. Nothing currently running replaces that precision. The practical implication is modest for most people, since personal decisions rarely hinge on a precise prevalence estimate, but it does mean that headlines claiming an exact number of current infections are usually extrapolating further than the data supports.

A quick recap

Mass testing ended, so reported cases no longer track infections. What remains is a mixture of hospital admissions data, wastewater sampling, sentinel swabbing through GP practices and hospitals, plus reduced genomic sequencing. Together these give a reasonable read on whether levels are rising or falling, without giving precise infection counts. Trends are trustworthy, exact totals are not.

Frequently asked questions

Are COVID-19 cases still reported daily in the UK?

No. Surveillance data is published on a weekly cycle through the UKHSA respiratory reports rather than as a daily case dashboard.

Can wastewater testing tell how many people are infected?

Not directly. It shows how much viral material is present in the sewage from an area, which indicates whether infection levels are rising or falling rather than producing a case count.

Does wastewater monitoring identify individuals?

No. Samples are taken from combined sewage serving thousands of properties, so there is nothing in the data that could point to a household or a person.

Why do COVID-19 and flu appear in the same report?

They share surveillance infrastructure, they circulate in overlapping seasons and they place pressure on the same NHS services, so reporting them together is more useful than separating them.

Are new variants still being detected?

Yes, through genomic sequencing of a sample of positive tests, though at lower volume than during the pandemic, so detection of a new lineage tends to take longer.