Parting of the waves

Gaps between confirmed cases and PCR-positive estimates grow.

Anthony B. Masters


On 1st April 2022, free Covid-19 testing for the general public in England ended. In Wales, some free testing capacity continues to the end of June. For people in Scotland, free rapid tests will not be available from 18th April.

Under a renewed definition, UK confirmed case figures are counts of infection episodes. For the United Kingdom except Wales, episodes are positive tests 90 days or more apart. In Wales, that gap in time is 42 days. There are differences between public health agencies in counting rapid positive tests too. That means agencies include SARS-CoV-2 reinfections.

Testing eligibility and capacity affects confirmed case counts. At the start, testing may have only been available for those going to hospital. Testing capacity expanded, with home kits and rapid lateral flow tests being attainable. On 4th and 5th January 2022, there were over 2m reported viral tests in the UK.

Now, governments are reducing that testing capacity. Withdrawing free tests may mean more people do not get tested. That lowers case ascertainment rates.

How can we discern trends, when testing limits the sight of surveillance? The Office for National Statistics runs a random sample of private households. That survey is not influenced by changes in surveillance testing for the public.

There are three main limitations with this survey. First, the survey measures the positive rate among those tested. People can stay positive with PCR tests for some time. Like a bath filling and emptying water, the crest in positive rates is after the crest of infections.

Second, the pool of potential respondents are private households. Institutional housing, like care homes, are not part of this estimate.

Third, random probability surveys are expensive. The ONS reported annual costs of about £390m. There are many sources of uncertainty, including in swab tests and non-response errors.

(Image: R Pubs)

There is a parting of the waves. Whilst confirmed cases in Wales fell, high positive rates in random samples persisted. In 3rd to 9th April, around one in 13 Welsh people (6.8% to 8.6%) would test positive. Infections continue in Wales, but public health surveillance catches fewer cases.

The R code is available on GitHub and R Pubs.



Anthony B. Masters

This blog looks at the use of statistics in Britain and beyond. It is written by RSS Statistical Ambassador and Chartered Statistician @anthonybmasters.