A backlog in testing, leaves us in an incredibly dangerous position, says Christina Pagel

Confirmed cases of covid-19 in the UK were rising slowly over August: it took five weeks for daily average case numbers to double from 753 on 1 August to 1530 on 4 September. It took only nine days for confirmed cases to double again to 3050 by 13 September. Leicester was the first city in England to go into local “lockdown” on 29 June and now about 8 million people are living under some additional regional restrictions across the UK, with highest positivity rates in England. After months of slow decline, hospital admissions in England have doubled since the start of September. How worried should we be? 

While various reasons for the rise in confirmed cases have been mooted, the last week has seen the evidence all point in the same direction of increasing transmission.

According to NHS Test and Trace, the number of new people tested for covid-19 did not increase over August. A related measure of “daily tests processed” shows a similar story as do the rising positivity rates in Public Health England’s (PHE) weekly surveillance data. Rising case numbers cannot be explained by increased testing.

That said, official case numbers depend on who exactly is getting tested: what proportion of new tests is in people with new symptoms? How much is regular testing of asymptomatic care home staff and residents and other key workers? How much is testing concentrated in hot spots? Unfortunately, this information is not readily available, but there are two other studies that avoid test bias by testing a representative sample of the population. The Office of National Statistics (ONS) has been running an infection survey, reporting every week, regularly testing around 30,000 people for covid-19 regardless of symptoms. The Imperial College REACT study tests 150,000 people every month over a two week period with its most recent round of tests covering the end of August and start of September. 

After reporting increases in infection since early July lows, ONS did not see further increases over August, but was also at the limit of interpretability given low covid-19 prevalence and relatively small sample size. That said, the most recent report did find a distinct increase in new cases. The REACT study published its latest results on 11 September and these reflected the pattern seen in confirmed cases, Public Health England’s surveillance reports, and the ONS survey. REACT reported a sharp acceleration in the increase in cases between its previous round in July and the last two weeks, mostly in young adults, and estimated that the current reproductive ration was as high as 1.7 with a doubling time of 8 days. Unlike the government’s case data, REACT also published data on viral load as measured by PCR tests. Recent viral loads were higher than in June or July and on a par with those measured in May. Rising cases are therefore unlikely to be due to people recently recovered from covid-19 being counted as weak positives while no longer infectious. 

It is undeniable now that the UK is experiencing a significant acceleration in the spread of covid-19 in the community. In response, the government has instituted a “rule of six” in England with similar measures being instituted in the other nations.  

Stopping transmission requires stopping infectious and susceptible people from mixing. There are various ways of doing this: the best is contact tracing which identifies and isolates infectious people and their contacts, with minimal disruption to others. Done well, such as in South Korea and Germany, this can control and reverse new spikes without the need for severe restrictions (but alongside measures such as physical distancing and mask wearing). Mediocre contact tracing will not be enough to reverse rising cases, but will identify regions where cases are happening (through testing) and how (through tracing contacts), allowing further targeted measures, for example, restricting pub opening in a specific area. When contact tracing fails, blanket national measures are all that’s left without knowledge of exactly where or how transmission is happening. 

August saw a simultaneous increase in people mixing, encouraged by schemes such as “Eat out to help out”, and a weakening of our already mediocre test and trace system, mainly in England. There are currently large backlogs, many symptomatic people unable to get tested, and no resolution in sight. This leaves us in an incredibly dangerous situation: people who can’t get tested will be much less likely to isolate and their contacts will not be traced; backlogs mean that tracing the contacts we do know about will be much less effective due to the time delay, when contacts are already struggling to isolate; and it means that daily case numbers or weekly positivity rates will tell us little about what is actually happening. Covid-19 now has a chance to spread undetected again in our population, just as schools have reopened, as students return to universities, workers to offices, and as care homes remain vulnerable. 

However, we are still nowhere near the situation in March when likely more than 100,000 people a day were being infected and doubling time was four days. Hospitals are also now much better placed to treat covid-19 patients. Finally, we now know what works: a combination of physical distancing, hand washing, mask wearing, avoiding crowded indoor spaces and excellent contact tracing. If the government urgently fixes the testing system, improves contact tracing, and supports new cases and their contacts to isolate, while we all do our bit to protect ourselves and others, then there is still time to turn this around. 

Christina Pagel is Professor of Operational Research (a branch of applied mathematics) at University College London (UCL) and Director of the UCL Clinical Operational Research Unit.

Competing interests: CP is a member of Independent SAGE.