This article is a part of a series in which leading commentators explore innovative scientific, medical and epidemiological approaches to the Covid-19 pandemic
The writer is an English statistician and epidemiologist at the London School of Hygiene and Tropical Medicine
Boris Johnson, the UK prime minister, is right that antibody tests that identify people who have recovered from Covid-19 and can thus return to work will be a game-changer for medical staff dealing with the care crisis. But of itself, detecting who is immune won’t end lockdown — either in the UK or anywhere else.
Most people aren’t immune because they haven’t been infected yet. One alternative is to wait for a vaccine, which will take at least a year, and continue with a severe lockdown. Alternatively, countries can gradually relax lockdowns and let the pandemic run its course slowly, which China and Taiwan have done and the UK seems to be planning.
This “flattens the curve” and prevents overload of health services. But by the time herd immunity is developed, at least one in 300 people are likely to die (about 200,000 in the UK). A far better approach would be weekly testing of whole populations, with immediate quarantine of infectious people and their contacts. This would allow the lockdown to be released while protecting lives and controlling the epidemic. Many say it cannot be done. But Britain may be one of the few countries that could do it immediately.
Every household would receive and return self-sample nasal and throat swabs every week in barcoded tubes pre-printed with their names and date of birth. These self-administered tests would result in some inaccuracies. But this would not matter. The tests are not for the benefit of infected patients. They are to quarantine infected households earlier and so reduce the infection rate in the whole population.
To test everyone weekly in the UK, about 10m tests per day would be needed, a hundred times the government’s eventual target and 600 times more than Germany is now doing. This seemingly impossible feat may be possible if every polymerase chain reaction machine in the country is diverted to 24-hour testing.
PCR is an amazing technology that locates a particular coronavirus gene sequence and creates multiple copies that can then be easily detected. They are inexpensive. One 96-sample PCR machine, which costs well under £1,000, can test almost 5,000 people a week if it runs continuously.
Globally, that translates into ending economic disaster for a capital outlay of about £1 a head: 20p for PCR machines and 80p for reagent production and other facilities. The UK, with a population of 66m, needs about 14,000 machines.
My hunch is that the UK already has that capacity. Every biology department in every university has dozens of PCRs; the genome project at the Wellcome Sanger Institute has thousands. Most academic labs have machines that can make the PCR primers that the test requires. The limit then would be mass production of other reagents. As a first step, the government should ask every biotech firm what they can make.
Recruiting every little lab in Britain with a PCR machine to do Covid-19 testing would see old PCR machines chugging alongside modern automated systems. If processing 10m tests a day can’t be achieved immediately, the programme should start in selected cities — and then expand as capacity increases.
If successful, it would be a wonderful achievement. Stopping the epidemic and the lockdown would save about 200,000 lives. That is 1 per cent of about a third of the population — the eventual number predicted to become infected under current policy.
It would also save the UK billions from otherwise lost economic activity, at little cost. I believe it promises to be at least as good, if not a better, system than South Korea’s, which has been praised for its use of intensive contact tracing, often via cell phone tracking, to control the epidemic.
In Daegu, where the virus in that country originated, lockdown reduced the daily number of new cases from about 500 at the beginning of March to about 20 last week. But an upward trend may now be emerging elsewhere in South Korea. By contrast, weekly testing of whole populations would detect new infections more rapidly and completely, and then be extended to biweekly and then monthly testing as virus prevalence falls. It would also be less intrusive than contact tracing, and guard civil liberties better.
Of course, there are practical problems, but they are manageable. Labs would have to be run by staff supplemented by competent students and technicians diverted from current duties, most of which are suspended anyway. Quality control would be achieved by including positive and negative controls among samples. Incompetent labs and companies that produced useless kits would thereby be discovered in the first round of testing and excluded.
As well as saving lives and money, this collaboration would also have lasting benefit. Once established, it could be revived at short notice in the next epidemic of a new virus. It should be established for that reason alone, and would cost almost nothing to maintain in readiness.
The only questions are how soon university and commercial labs can switch to doing Covid-19 tests, and how soon British biotech companies can provide the kits. Patrick Vallance, the chief scientific adviser, should issue that public call now.
Letter in response to this article:
Copyright The Financial Times Limited . All rights reserved. Please don't copy articles from FT.com and redistribute by email or post to the web.