This paper brings up to date the latest position on Covid-19 testing in the UK.
It sets out the different ways to test and the availability of the different types of tests.
The paper includes a model for mass testing that can evolve as testing capacity increases but importantly offers an operation that can be rolled out now. The authors believe mass testing, combined with other measures discussed in the Institute’s latest paper on exit strategies will enable us to exit the lockdown while keeping the level of infections under control.
Chapter 1
This paper is intended to follow our previous contributions to the testing and exit strategy debate in the UK.
Everyone we have interacted with, including those in government and Public Health England, have shown great willingness to improve the UK’s testing infrastructure.
They are all working extremely hard throughout a crisis unprecedented in scale and in scope. This paper is intended to stretch thinking and serve as a critical friend.
In particular, it points to a number of key points that government can address and take on board, all of which should be overseen by a new Minister for Testing[_]. Reporting directly into the prime minister in a structure that sits outside of Public Health England (PHE) and convening the best of health, industry, technology and science, this minister would absorb the following into her/his brief:
Scale antigen testing using every means possible. This includes taking up every offer of lab space[_], replacing proprietary reagents with white-label reagents where necessary and harnessing an ecosystem of antigen testing suppliers – including rapid antigen tests.
Recognise the importance of an antibody test. Administered often and at the right point in the disease cycle, this will be accurate and is a good tool in any event for understanding the spread of a disease. These tests can be used at home and millions can be frequently produced and distributed.
Don’t let perfection be the enemy of good enough. Modelling by Nobel prize-winning economist Paul Romer, detailed in this paper, shows that mass testing, based on tests less accurate than 98 per cent, can still be effective in a strategy of mass testing to end lockdown.
Partner British companies in the antibody and rapid antigen space with production lines. There are many promising tests out there which will only improve in accuracy – indeed, many already meet the criteria of a lower-accuracy test that, according to the excellent work of Romer[_], would still be of benefit. It’s important that production is considered now, as tests will need to be produced and distributed at scale.
Ambitiously recognise mass testing at scale involves a large proportion of population. The UK has struggled to ramp up its testing capacity but this should not drive its objectives: We need testing in the hundreds of thousands and, eventually, the millions. This is realistic, especially when the specification of tests are lowered and accounted for – as demonstrated by Romer’s modelling.
Administer both antibody and antigen tests. Separately, both tests have value. Together, they are invaluable. Mass testing should incorporate both, ideally at the same time for each patient. It’s important to note that the lack of availability for one type of test shouldn’t prevent the administering of another.
Contact trace online and offline. Digital contact-tracing efforts are going to be essential, and with the platform Apple and Google recently provided, the NHS app must be privacy-preserving. However, digital efforts are complementary and not a substitute for human expertise, and resources are therefore necessary to make these a key component to loosening restrictions.
Begin preparations for community testing now. The ability to test at scale is not just about the availability of tests. To put the infrastructure in place for mass testing, mobilisation needs to begin now.
Draw up immunity-certificate framework. Antibody tests can demonstrate immunity and the government should ready a framework to introduce immunity certificates. These may be time-limited and, for ethical considerations, may not be used to prioritise re-entry into the job market. However, they would be very important to easing the burden on testing capacity and, as part of a globally coordinated effort, may be part of the solution to lifting international travel restrictions.
Develop a clear exit strategy with mass testing at its heart. This should combine phased mass testing with other initiatives that will: 1. reduce reinfection rate, 2. ensure there’s always enough NHS capacity, and 3. manage burden on testing capacity. These initiatives include contact tracing, phased release of young people from lockdown and other measures referenced in our section on the "STIR" mass-testing strategy later in the paper.
Chapter 2
This section focuses on what we currently know about how infectious Covid-19 is, the fatality rate and the effect of Covid-19 and the lockdown on the delivery of other non-Covid health care. Building this understanding is vital in both identifying the right exit strategy from lockdown and establishing the right mass-testing regime.
Infectivity
A good starting point for understanding the infectiousness of Covid-19 is to look at the basic reproduction number or R0. This is an estimate of the average number of people an infected person will spread the virus to without intervention in a population where no one has immunity. Over time, as more people become infected and interventions occur to reduce the spread, the effective reproduction number, Rt, represents the actual infection rates. The purpose of suppression activities is to lower Rt to less than 1, which will cause the number of new cases to fall.
The R0 for Covid-19 has been estimated to be between 2 and 3.[_] However, the large number of asymptomatic undetected cases makes this estimate quite uncertain, and one recent paper estimated the R0 to be 5.7.[_] The R0 estimate matters, because it makes it possible to estimate what proportion of the population needs to be immune to achieve herd immunity. For example, if the R0 is 2.5, then herd immunity is achieved at 60 per cent, but if the R0 is 5, then herd immunity requires 80 per cent.[_]
All this underlies the need for comprehensive testing. This would allow a better understanding of how infectious the disease is, and how close we are to herd immunity.
Lethality
The lethality rate of Covid-19 is ideally recorded through the infection fatality rate – the proportion of deaths among infected individuals.
A simple estimate based on current numbers of confirmed cases and deaths would suggest a fatality rate of 6.5 per cent, however, this is not likely to be the infection fatality rate for three reasons. First, the number of confirmed cases is likely to be much smaller than the true number of cases, since those who become ill but are asymptomatic or have mild symptoms are often never tested. Second, deaths that occur outside of hospitals may be missed. And third, there is a lag of a few weeks between infection and death, so deaths from the newest cases have not yet been incorporated into the figures.
While the first of these factors would imply a lower death rate, and the latter two factors a higher one, the general consensus is that the number of missed cases is the biggest factor here, so the 6.5 per cent estimate is too high. Current estimates place the infection fatality rate at between 0.1 and 3 percent.[_]
It is useful to put the mortality estimates in context. We know that the infection mortality rate increases with age. Professor David Spiegelhalter found that by looking at infection mortality estimates from Imperial College[_], the risk of dying after being infected with Covid-19 is approximately the same as the risk of death from all causes (see the graph below).
Infection mortality rate of Covid-19 and annual mortality
Source: Winton Centre [_]
One way to interpret this graph is that a random person infected with Covid-19 undergoes the same risk of death in the next few weeks as they would have done in the whole year if they had not been infected.
This chart does not, however, reveal the effect of underlying health conditions on mortality risk. It is well understood that individuals with serious underlying health conditions (e.g. diabetes, cancer) are at much greater risk of death from Covid-19. While there are many cases of young healthy adults becoming very sick,[_] the most cost-effective mechanism to reduce deaths will be to reduce exposure by the elderly and the unwell to the virus.
Effects on the Delivery of Other Health Care
The spread of coronavirus has created obvious strain on health-care resources, which will likely contribute to excess deaths. In the week to 3 April, there were more than 16,000 deaths in England and Wales, while in a typical year there would only be around 10,000.[_] Around 2,500 of the 6,000 excess deaths were not due to Covid-19, suggesting the pandemic may already be indirectly causing deaths.
The government has focused on increasing the supply of labour and resources (e.g. bringing medical staff out of retirement and setting up emergency Covid-19 hospitals) and decreasing demand for non-urgent care (e.g. by cancelling elective procedures). The NHS has been helped by an almost 50 per cent reduction in A&E admissions, which is likely due to a mixture of individuals deliberately avoiding hospitals, and a possible reduction in accidents.[_]
Despite the increase in labour supply and possible reduction in accidents, the scale of the pandemic will inevitably mean that health-care resources are stretched more thinly between individuals. As a starting point, the NHS has very limited slack to accommodate these changes: The UK has historically low numbers of ICU beds compared to similar countries, and fewer doctors per person (2.8 per 1,000) than the average EU15 country (3.9 per thousand).[_]
While data on the current crisis is limited and descriptive, prior research suggests that reduction in health-care capacity leads to increases in deaths. A 2012 research paper found that reduction in nursing staff due to a strike led to in-hospital mortality increasing by almost 20 per cent.[_] Beyond the effects of cancelled procedures and lower-quality care, many individuals are likely to be avoiding the hospital altogether, possibly even in emergencies, further contributing to excess deaths.
Parents may also be less likely to vaccinate their children, which would create large risks from other infectious diseases such as measles. Measles is incredibly infectious; it has an R0 of 12-18, meaning that herd immunity requires a vaccination rate of 90-95 per cent.[_] Even small reductions in vaccination rates therefore can make the disease viable again.
The pandemic has also highlighted the serious risk of infection and death faced by medical staff who cannot self-isolate. This may reduce the attractiveness of working in medicine, which could lead to staff shortages or a need to increase compensation.
Chapter 3
Discussion on testing has permeated the public consciousness. This is good news. As we have argued since March, testing is key to exiting lockdown and mitigating the health and economic harms of Covid-19. The widespread discussion of testing has given rise to confusion around terminology and what testing is available, in what format and the roles different tests would play in any mass-testing regime.
Further, the unprecedented innovation in the biotechnology sector at home and abroad – a very welcome development – has see new developments, particularly in the field of rapid testing. It’s important to note that both the antigen and antibody tests have the potential to be delivered at speed, but currently this only really applies to the antibody test.
To avoid confusion, this paper will use the following terms:
Chapter 4
As of 9am on 20 April, the UK had conducted 501,379 tests for Covid-19, with 19,316 tests carried out on 19 April.
Of the 386,044 people tested, 124,743 were positive.[_]
Government Strategy
The government’s strategy on testing is set out in a paper published by the Department of Health on 4 April.[_]
This strategy aims for the UK to be conducting 100,000 Covid-19 tests per day.
To achieve this, the government has set out five pillars of work:
Boosting PCR swab testing. This testing would be conducted by PHE and NHS labs for patients and frontline NHS workers. The government target for this pillar is to reach 25,000 tests per day by mid- to late-April.
Creation of new swab testing capacity delivered by commercial partners. This pillar sees partnership with universities, research labs and private companies as a route to building a network of new labs and testing sites.
The government is working with several companies to bring online antibody tests.
The government is conducting a large-scale survey to work out what proportion of the population has had the virus. They are doing this using antibody tests operated by PHE at the Porton Down science campus.
Building a large Germany-style diagnostic industry in the UK.
As we have set out previously, we fully support the government’s strategy, particularly as a stepping stone to moving to mass testing.
Concerns have been raised, however, in recent weeks on both the capacity of scaling the PCR tests to the level required and whether antibody testing can be brought online in time.
Chapter 5
As we have set out previously, the antibody test is used to establish whether someone has had and is therefore immune (at least in the short term – see below) to the virus.
How Do Antibody Tests Work?
A typical antibody test would involve taking a small sample of a patient’s blood – for instance via a pin prick. The test looks for two types of antibody: IgG and IgM.
IgM (Immunoglobin M) are the first antibodies to be produced by the immune system. They have a half-life of around five days. IgM antibodies usually appear within five to seven days of infection and peak at around 21 days. Detection of these antibodies suggests the person has existent or a recent infection.
IgG (Immunoglobin G) antibodies are more numerous and can be detected around 10 to 14 days after infection. The presence of these antibodies indicates a person has recovered from the virus and is now immune.
The results using this test (example in image below) would indicate:
A positive for IgM would show someone has the virus or has recently had the virus.
IgM and IgG positives would indicate someone is within the first month of infection and immune.
Just IgG would indicate someone is immune and the infection occurred some weeks ago.
Types of Test
Rapid diagnostic test (RDT):
This is a simple positive/negative lateral flow assay test, like a pregnancy test kit, that can be done at home or at point of care. Typically, these tests have used a finger prick to produce a small blood sample but could also utilise saliva samples or nasal swabs.
These tests would take between 10 and 30 minutes.
Source: Center for Health Security [_]
Enzyme-linked immunosorbent assay (ELISA):
This is generally a lab-based test. It tests whole blood, plasma or serum samples. This method uses a plate that is coated with viral protein. Patient samples are then incubated with the protein and if the patients has antibodies to that protein, they bind together. The bound complex can be detected using another wash of antibodies that produces a colour readout.
Time to results is around one to five hours.
Source: Center for Health Security [_]
Neutralisation assay:
The neutralisation assay test uses patient antibodies to prevent the viral infection of cells and is conducted within a lab setting. The test utilises whole blood, serum or plasma from a patient. It shows if a person has active antibodies that are effective against the virus. The test relies on cell culture to allow Covid-19 growth. When the virus and cells are grown in decreasing concentrations of patient antibodies, it can show how many antibodies are in the patient serum that are capable of blocking virus replication.
The time to a result is between three and five days.
Source: Center for Health Security [_]
Chapter 6
Insight into potential long-term immunity comes from studies of the nearest relative coronaviruses, SARS and MERS. Considering the data from these two different viruses was cited as unlocking the “best guess” on Covid-19 immunity by Danny Altmann, Professor of Immunology at Imperial College London in a recent British Society for Immunology Webinar[_].
Altmann shared a long-term longitudinal study[_] of those infected by MERS in South Korea which shows a significant degree of immunity post-infection, lasting for up to a year albeit with some waning. They were tested with a plaque reduction neutralisation test (PRNT), considered to be the “gold standard” for detecting and measuring antibodies.
Interestingly, the study also showed a small number of people with relatively undetectable amounts of antibodies. In this instance, T-cells also provided a reliable guide to immunity and those with lower antibodies in the study had comparable T-cell counts compared to those with high antibody counts.
Elsewhere, a UCL research team looked at the historical patterns of three common coronaviruses[_]. Their results provide some evidence of immunity against reinfection by the same virus, as they did not identify any people who were reinjected by the same virus. Based on their simulations, if people had no immunity after being infected, the probability of zero reinfections by the same virus in their study sample was only 3.48 per cent, which they say suggests some immunity is likely.
‘Reinfection’ in South Korea
South Korea’s Centre for Disease Control and Prevention recently reported that 91 patients who had been infected with Covid-19 and then later tested negative had now tested positive again[_]. If these were genuine reinfections, they would cast doubt on the strength of the immunity the patients had developed. It’s possible that testing flaws may be to blame for this, and many scientists[_] believe it likely that these patients had a false negative test. Given that Covid-19 closely resembles the coronaviruses that cause SARS and, to a lesser extent, MERS, and the fact there are no reports of reinfections with the SARS virus, we are confident for the purposes of this paper that immunity will be conferred.
It can be reasonably inferred that exposure to and recovery from Covid-19 will provide some level of immunity. For the purposes of this paper, we have assumed that such immunity exists and it can be identified through the presence of antibodies.
There are questions around how long such immunity lasts and whether it is conferred in every case of Covid-19. For the purposes of this paper, we have made a confident assumption that there will be a level of immunity. Once more is understood about the virus, changes can be made to a mass-testing and tracing model, for example:
An expiry date on immunity certificates that matches the length of an immunity period following the contraction of Covid-19
Consecutive antibody tests before an immunity certificate is issued
Exploring development and use of T-cell tests
Ongoing but less-frequent antigen testing to ensure that reinfection has not occurred
Continuation of self-isolation if a person exhibits the symptoms of Covid-19
We watch with interest the work of the British Society for Immunology, whose expert group are currently collating what is known about the immunology of Covid-19 and developing immunology research priorities in response to the coronavirus outbreak[_]. The group will publish their outputs within approximately three weeks (May 2020), with the intention of urgently mobilising and coordinating the UK’s immunology research response to coronavirus (Covid-19).
Chapter 7
On 24 March the government indicated it had purchased 3.5 million antibody tests, saying they would be available “very soon”.
On 30 March there were indications it had made an agreement in principle to purchase 17.5 million antibody tests.
On 4 April Professor John Newton, National Testing Co-ordinator, said none of the tests they had looked at had proved viable, but that he was hopeful one would become available in months.
Last week the Medicines and Healthcare products Regulatory Agency (MHRA) published the specifications it is looking for from the antibody tests.
We understand the UK government is now sceptical of being able to secure a viable lateral flow antibody test to use for mass testing.
Instead they are actively investigating the potential of ramping up the ELISA testing PHE is doing via the Porton Down laboratory.
Specifications
The MHRA specifications for a point-of-care antibody test require 98 per cent sensitivity and specificity:
For antibody self-tests, MHRA requires sensitivity of 95 per cent and specificity of 98 per cent:
Case Study
Do Antibody Tests Work?
Discussions on the accuracy of antibody tests broadly fall into two categories:
The specificity and sensitivity of the tests
When they are used
Accuracy of the Test Itself
Sensitivity refers to how well the test works in identifying the IgM and IgG antibodies associated with the coronavirus. The higher the sensitivity score, the more likely the test is to detect these antibodies (and therefore avoid producing false negatives).
Specificity is the measure of how focused the particular test is on identifying the specific antibodies it is designed to detect (for coronavirus) and not other similar ones. A higher specificity score means the test is more accurate at specifically identifying Covid-19 antibodies.
There are legitimate steps that need to be taken to ensure a test meets the right sensitivity and specificity to be viable.
Timing of Testing
A further issue in terms of accuracy is the timing of when the test is used. Antibody tests are more effective, and accurate, when used post-infection. Where they are used around 14 days after infection – when IgG antibodies are detectable – they are most effective. So while the test must be effective in and of itself, it is also vital it is used at the right time.
Antibody Testing: What’s on the Market?
Appendix A includes a list compiled by Johns Hopkins of serology tests that have been approved for diagnostic use in other countries, as well as others under development.
Alongside this list we have also been in discussion with a range of companies that believe they are close to having a validated test (in the US), or have viable tests that have not yet been approved by PHE in the UK.
Examples of UK companies who claim to have viable tests are included below:
Biopanda: A Northern Irish biotech company that produces a rapid antibody test. The firm is selling its testing privately within the UK and has also previously despatched orders “throughout Europe and across the world.”
SureScreen: Private company based in Derby that says it has developed a rapid test that can reach results with 98 per cent accuracy. It is based on using a finger-prick blood sample and can produce a result in 10 minutes. SureScreen says its tests are being used by private buyers in the UK, Ireland, Germany, Kuwait, Netherlands, Oman, Spain, Switzerland, Turkey and the UAE.
In the US the Food & Drug Administration (FDA) has given Emergency Use Authorisations (UEAs) to three tests:
Source: FDA
This week Abbott launched a new test, without an EUA, which specifically looks for IgG antibodies, rather than both IgG and IgM. The company claims its test, when it is used at least two weeks after someone shows symptoms, has a sensitivity rating of 100 per cent and a specificity of 99.5 per cent.[_]
Chapter 8
The Two Tests Together = a 98.6% Detection Rate
The best test for an early infection is combining the antibody test and the PCR swab taken from the patient. Then we have a 98.6 per cent detection rate[_] within the first five-and-a-half days of infection. According to a study, the combined use of antigen and antibody testing improved identification of positivity through various phases of illness.
While antibody tests alone may not be enough to diagnose Covid-19, they can be a valuable diagnostic tool when combined with antigen tests. Owing to their scalability, antibody tests can tell us huge amounts about the spread and behaviour of Covid-19 and help us understand the immunity response to the virus.
Case Study
Reminder: What Does The Antibody Test Reveal and When?
A suitable antibody test for Covid-19 would reveal the presence (or not) of two antibodies:
Immunoglobulin G (IgG): This is the most common antibody. It’s in blood and other body fluids and protects against bacterial and viral infections. IgG takes approximately 14 days to appear but then remains, providing immunity.
Immunoglobulin M (IgM): Found mainly in blood and lymph fluid, this is the first antibody the body makes when it fights a new infection. It usually appears seven days post-infection and peaks at 21 days.
The following table shows the clinical significance of combining results from antigen tests (RT-qPCR) with the results from antibody tests. In this paper, we make the recommendation for mass testing to utilise both.
Source: Diazyme [_]
This table is based on the current knowledge about the rise and fall of Covid-19 antigens, IgM antibody and IgG antibody and the correlation of these level variations with the initial time of infection, onset of symptoms and recovery phase[_][_][_]. As shown in the graph below, serological tests are recommended to be used on patients at least three days after onset of symptoms or seven to 10 days after infection with the virus.
Variation of the levels of antigen and IgM and IgG antibodies after infection
Source: Biopanda Reagents [_]
Han Yan, a regulatory officer at antibody-test producer Biopanda, told Channel 4 that, “there is a window period where a person can be infected, and even show symptoms, but during which the antibody test will not come back positive. However, during this time, a PCR test should detect the presence of the virus.
“With all this in mind, we would never recommend our antibody test as a way of replacing RT-PCR for early-mid stage diagnosis of Covid-19 because of the risk of false negatives …
“So if the antibody test is being used to diagnose current infection, it must be used with RT-PCR.”[_]
The key conclusion is that the results of antigen and antibody tests do not necessarily need to agree. A disagreement between the two tests provides useful data and can often be traced to the after-infection time points at which the tests were performed. Overall, while antigen testing may be appropriate for the detection of the Covid-19 virus during the acute phase, antibody testing is appropriate during the chronic phase. Since the exact time of infection is often unknown, combining both tests improves the accuracy of the Covid-19 diagnosis.
Evidence we cite above from a study conducted in China, published in the New England Journal of Medicine, shows combining PCR and antibody testing dramatically improves the chances of achieving an accurate result.
The study looked at blood samples from 173 patients (in Shenzhen) and found that antibody tests began to give more reliable results than PCR testing after the first five-and-a-half days of the illness. By combining antibody testing with PCR testing, doctors were able to detect 98.6 per cent of coronavirus cases, as compared with 51.9 per cent by using PCR testing alone.
The study found that, “Serologic tests can improve early diagnosis of Covid-19. Because of the high false-negative rates with PCR, serologic tests will be a useful supplement to RNA detection.[_]
Based on what we know of the virus, there are limitations to the antibody test that mainly relate to the slow pace of the human antibody response to Covid-19. Several studies are ongoing but current science is not conclusive and it would appear antibodies may not be detectable before three days after onset of symptoms – or at least seven to 10 days after infection.
Chapter 9
We agree with Professor Neil Ferguson of Imperial College in calling for the government to have a “single-minded emphasis” on developing mass testing and tracing and tracking new cases.[_]
To achieve this, we propose a four-step model to mass testing – the STIR testing circuit – that can be scaled up as both testing capacity increases and more people become immune to the virus.
The Mass Testing Operation: How It Would Work
Source: TBI
The government’s objective should be to get as many patients as possible through testing on a regular basis and into the orange category (active socially and economically, not affected).
If a policy of herd immunity is pursued, the government may switch priority to focus on getting as many people as possible into the green category in the above chart (active socially and economically, immune).
The number of people who die from this disease will be mitigated by:Shielding of vulnerable groupsReduction in transmission (enhanced by contact-tracing, isolation and other measures including masks)Therapeutics, which may reduce the need for shielding once widely available
Administering both tests will become increasingly feasible, especially when rapid antigen tests become available. That’s why we recommend other measures, such as shielding, to reduce testing demand
Further, we wouldn’t limit antibody testing because of the availability of antigen testing – so people may receive the former while production of the latter is being scaled.
The Mass Testing Operation: In Depth on the Phases
Source: TBI
The “Joint European Roadmap towards lifting Covid-19 containment measures” makes clear how important this testing strategy is to easing restrictions.
As the document sets out, a key criterion for lifting the lockdown is: “Appropriate monitoring capacity, including large-scale testing capacity to detect and monitor the spread of the virus combined with contact tracing and possibilities to isolate people in case of reappearance and further spread of infections. Antibody detection capacities, when confirmed specifically for Covid-19, will provide complementary data on the share of the population that has successfully overcome the disease and eventually measure the acquired immunity.”[_]
Chapter 10
Economist Paul Romer’s modelling shows the importance of testing as part of a wider approach to handling the virus. (The coloured lines in the graphs below show the results from 50 runs of the model. The black line is the average at each date of these 50 runs.)
The Benefit of Testing to Suppress
Romer’s modelling shows that if testing is used to determine who needs to isolate, then the number of people needing to be confined is dramatically smaller.
Source: Paul Romer [_]
Accuracy
The modelling shows that these benefits are possible even with an imperfect test.
“How much difference does it make if the test used to send people into quarantine is bad? Not as much as you might think,” Romer has said.[_]
Source: Paul Romer. The coloured lines in the graphs show the results from 50 runs of the model. The black line is the average at each date of these 50 runs.
As Romer concludes, “The simulated data here contrast policies that isolate people who test positive using four different assumptions about the quality of the test. Even a very bad test cuts the fraction of the population who are ultimately infected almost in half. And when I say bad, I mean bad – an 80% false negative rate, which means that 4 out of 5 of people who are truly infectious will get a negative test result – i.e. a result saying that they are not infectious.”
Therefore, we believe that the pursuit of a “perfect” test is misguided for the purposes of informing strategy for exit from a lockdown. This should be distinguished from the use of such tests for clinical purposes, where the need for accuracy is much higher. The modelling above demonstrates that there is a clear net gain from using a test that successfully isolates even 60%+ of those with Covid-19 in society, and we are confident that tests with higher accuracies are currently on the market. The impact of lower accuracy is further mitigated by other measures including shielding, PPE (including masks) and staged exit from lockdown of key groups.
Chapter 11
As we have set out in a previous paper on contact tracing[_], tracing is going to be a key component of response and integral to any exit strategy, with measures adopted in this area needing to be both online and offline. However, there are a number of policy considerations that need to be worked through, in particular on digital tracing efforts, encompassing privacy, take-up and efficacy, even to the question of what defines close contact, both in terms of distance and duration.
Largely being explored for the first time, around the world a number of digital contact-tracing efforts are already underway. These use different approaches and different degrees of intrusiveness:
South Korea, for example, uses GPS and location tracking, having made significant legal changes around powers for data collection in a health crisis, following their experience with MERS.
Singapore’s open-source app, TraceTogether, uses Bluetooth technology, but has set out clear guardrails around privacy and time-limited collection of anonymised data.
The European Union has also developed a toolkit for member states in developing apps, with guidelines specifying that they need to be voluntary, approved by the national health authority, privacy-preserving (“personal data is securely encrypted”), and dismantled as soon as no longer needed.
Researchers and technology companies are also assisting with efforts, with groups such as Covid-Watch, MIT’s Private Kit: Safe Paths and DP3T working on privacy preserving approaches, as well as shared protocols. And Apple and Google recently announced a collaboration on interoperable, privacy-protecting APIs to enable the development of tracing apps. As well as requiring these to be used only on an opt-in basis, they will also limit certification to national health authorities in an attempt to limit the risk of disruptive or malicious “false flag” activities.
In the UK, the government has announced that it is developing an app, which is likely to be embedded within a strategy to loosen restrictions, although the details of this are yet to be released. Given the broader developments in this area, there are a number of areas and trade-offs the government will need to consider:
Privacy: Apple and Google have now provided the framework for which apps should be developed, but there are still some privacy questions about how positive identifiers are collected, and risks around tracking, including correlation attacks and deanonymisation, as well as broader cyber-security concerns. The government will need to be clear about what the architecture, access to data and purpose is, as well as how it is collected, how it is stored and retention.
Efficacy: The speed and timing at which this is deployed is key and must be in unison with other elements of the strategy and when the R0 number is low. There are some definitional issues that will also need to be worked through, as well as mechanisms to deal with false positives. One put forward by Covid Watch’s Tina White is confirmation from health-care providers provided via a separate app, but which also underlines the need for testing on scale.
Take-up: One of the issues with Singapore’s efforts has been a low level of take-up. Modelling from Oxford University suggests 56 per cent of the total population needs “to use the app to completely suppress the epidemic, if combined with ‘shielding’ of over 70s.”[_] The government are working with the Behavioural Insights Team on this issue, but public acceptance and take-up will be dependent on them being perceived to be effective and trustworthy and that they will be limited in duration. That Apple and Google have pushed it down to a platform layer helps. Government will build the app on top, pushing out targeted marketing and clearly explaining utility. Communications will therefore be key and working in unison with the health service and communities will be essential, with the app provided by the NHS, not Gov.uk. One other constraint, however, which also raises equity concerns, will be those without access to devices.
Online and Offline
What is clear however, is that digital methods should also be seen as complementary to other methods, not a substitute. Both online and offline should work together as part of a broader toolkit. This is something a senior official in the Singaporean Government, Jason Roy, has recently emphasised, saying that digital contact tracing should work alongside manual, with a “human-in-the-loop system” being “necessary to allow judgment to be applied, given the high likelihood of pre-symptomatic transmission of the SARS-CoV-2 virus.”[_]
This is particularly important to deal with issues around false positives and negatives. Manual tracing is labour intensive, but it is clear that human skills and expertise are a necessary part of the equation. In Wuhan, efforts allegedly included 1,800 teams of epidemiologists, each with a minimum of five people. As the UK launches its app, it is therefore crucial that digital efforts support human ones.
Chapter 12
The building blocks for a mass-testing regime are swabbing sites; the workforce to conduct swabs; tests and equipment; and coordination. The biggest constraint to scale lies in the availability of tests and protective equipment, on which there is substantial focus. The government should set out a strategy for mass testing so that plans can be activated immediately to mobilise the other components.
Swabbing sites: Establishing a network of swabbing sites is necessary to test the numbers of people required, while not putting health settings at risk. A mix of models for these sites are required to respond to different geographies and the needs of the population. Versions of all of these models are in operation or development in the UK but require immediate plans for scale. A number of drive-through sites are operating effectively for NHS staff and provide a comparatively high-volume method and should be a core part of the network of testing sites. For vulnerable people, motorbike couriers are in operation and are able to take swabs direct from people’s houses. In more rural areas, or low car-density locations, mobile-testing units may provide an appropriate option.
Workforce: Frontline health staff are focused on the care of patients; therefore, a temporary workforce needs to be identified and trained to conduct tests. The PCR test requires a practitioner to take a throat and nasal swab. The quality of these swabs will vary between practitioners. Locally based systems have already been activated using staff newly trained in the collection of these swabs. It will be important to monitor quality, but there is no evidence to date that this is not a viable route to scale.
Coordination: NHS Acute Trusts are currently coordinating the testing regime; this is entirely appropriate as testing priorities are focused around health needs. However, as emphasis shifts from testing for health purposes to testing a broader population, a new governance arrangement should be put in place. The aim should be to bring in other agencies who are better placed to coordinate a testing regime for a broader population, and provide the significant resource that will be required for contact tracing. This will enable acute trusts to focus on patient care.
Recommendations
To deliver mass testing in the community, planning to scale operations must commence immediately.
Central government should set out a framework for mass testing, indicating who should be prioritised and how regularly they should be tested. This should then be operationalised locally.
Responsibility for mobilising and administering the testing regime should broaden beyond NHS Acute Trusts.
Scaling Testing
Antigen Testing
Reaching the capacity to test hundreds of thousands – if not millions – of people on a regular basis is necessarily ambitious. The government has made very positive steps to reach such scale, including the organisation and appropriation of university labs to increase testing capacity.
Scale will be achieved through a combination of measures that include the following:
Managing the demand on antigen testing. This will be aided by shielding measures and immunity certificates, phasing the entry of citizens into the “STIR testing circuit”. See above.
Increasing lab capacity and efficacy.
Improving the collection, quality and delivery of samples.
Integrate collection and delivery of results through software solution.
How do we ramp up antigen testing?
Antibody Testing
Antibody tests are, feasibly, much easier to produce on mass scale. The roadblocks to achieving a mass rollout of these tests lie, at present, largely around the issue of test validation.
To scale up this capacity we therefore suggest the following:
PHE to set out the exact process for the validation process.
PHE to make available patient samples to ensure the private sector has full capacity to develop and validate their own tests. In particular PHE should ensure it makes available seroconversion panels (blood samples of patients positively confirmed as having the virus and then having recovered). These samples should be used to validate tests, ensuring they are carried out on blood containing IgM and/or IgG antibodies.
Government and PHE to publish robust modelling showing the levels of accuracy of home antibody tests that would enable a mass testing regime to be effective. We believe a lower threshold than 98 per cent will be workable for home testing.
Government and PHE to procure antibody testing kits from the widest possible network of suppliers – both small and large producers, as well as from domestic and international companies.
Chapter 13
There are currently no approved therapeutic drugs to treat Covid-19 in the UK. The need for effective treatment for patients suffering from Covid-19 is high, given the fact that one in seven patients hospitalised with the virus die and about 50 per cent of ICU patients eventually die. The UK controls the supply of drugs that appear to be relevant for managing Covid-19, many of which are also used to manage other diseases; they are being prescribed for trials and are discouraged for use outside such research.
The UK government, through its NIHR and UK Research and Innovation, has provided £20 million in funding “for research projects that will contribute to our understanding, diagnosis, prevention or management of coronavirus.”[_] The call for proposals covers two categories, one of which is vaccines and therapeutics. The fund closed on 13 February with results expected within 18 months. Given the overwhelming burden on hospitals in the UK (and elsewhere), development of treatments for Covid-19 is critical.
Regulatory authorities have fast-tracked the process for bringing drugs to market because of the scale of the pandemic. In conjunction with a nationwide testing program, therapeutics offer the hope of easing the burden on the country’s hospital system, hastening recovery from the virus, and allowing people to resume their lives, including returning to work.
Ongoing global research to develop treatments for the virus falls into three categories:
Drugs already approved and publicly available for treatment of other diseases that are being tested for treatment for Covid-19.
Unapproved drugs that have shown promise in animal studies with severe acute respiratory syndrome (SARS) and Middle East respiratory syndrome (MERS).
New research (pre-clinical trial) to develop new drugs and approaches. These will take the longest to develop.
The last category includes the use of antibodies, which are harvested from the blood plasma of people who have the virus. The challenge is that the process is slow as it requires donations of blood plasma, which must then be turned into a usable form that can be used as a therapy. Several SARS antibodies may hold some promise in treating coronavirus.
In the case of the UK, research funds are focused on supporting projects that show the potential for rapid development: "[r]e-purposing of existing therapeutics, e.g. proteases, helicases or entry inhibitors; development of mAbs or other biologics." [_]
The European Medicines Agency listed the following treatments that are currently undergoing clinical trials to assess safety and efficacy:
remdesivir (investigational)
lopinavir/ritonavir (currently authorised as an anti-HIV medicine)
chloroquine and hydroxychloroquine (currently authorised at national level as treatments against malaria and certain autoimmune diseases such as rheumatoid arthritis)
systemic interferons and in particular interferon beta (currently authorised to treat diseases such as multiple sclerosis)
monoclonal antibodies with activity against components of the immune system[_]
A full list of therapeutics under evaluation in the UK through its special fund follows at the end of this paper as Appendix B. Most trials are of existing, approved drugs.
Case Study
Hydroxychloroquine
Evidence suggests that usage of hydroxychloroquine is in widespread use beyond the United States. Sermo, a health-care data company, surveyed 5,000 physicians in 30 countries and found that 44 per cent prescribed the drug for Covid-19 patients; 38 per cent stated that it was making a positive difference.
Hospitals in the US are using hydroxychloroquine in combination with azithromycin, to treat coronavirus patients. Despite its efficacy in some cases, a small study (81 hospitalised patients) of the anti-malarial drug in Brazil was recently stopped because patients given the higher dosage (600 mg of chloroquine for 10 days) experienced heart arrhythmias; 11 patients eventually died. Results for the smaller dosage (450 milligrams of chloroquine twice a day for five days) could not be determined due to the small number of patients in the study.
As we have argued in previous papers on testing, we believe there is a vital role for therapeutics in navigating an end to lockdown and supporting a programme of mass testing.
By alleviating the symptoms of the virus, slowing its development, and giving the body time to recover, therapeutics offer a path to reduce demands on frontline health care and allow for medical interventions in the community.
Chapter 14
As this paper has shown, the only viable basis upon which to build an exit strategy from lockdown is mass testing.
Such an approach would see the UK government scale up antigen testing (both PCR and rapid tests) and roll out antibody testing.
While this remains the government’s stated objective, it will need the right structure and strategy.
Critically this must be driven by the appointment of a senior minister with sole responsibility for testing, reporting to the prime minister.
Mass testing would build on both the existing superlabs being created, the full capacity of smaller labs across the country, and scaled-up community testing (e.g. expanding drive-through testing and adding more community testing points like polling stations).
While questions have been raised about the accuracy of antibody testing, we understand that regular testing, using both PCR and antibody testing, is highly accurate – 98.6 per cent accuracy – and that antibody testing will become vital (through detection of IgG antibodies in particular) in identifying those now immune and who can therefore leave the mass-testing procedure.
The economic damage of lockdown (estimated at £2.4 billion per day) is not sustainable. But, equally, we cannot simply lift lockdown and risk a significant spike in cases of the virus.
A paper published last week by the Lancet argued that, “Close monitoring of the instantaneous effective reproduction number and real-time tuning of policy interventions to ensure a manageable second wave remains the over-riding public health priority … Early detection of cases is essential … [Significant levels] of testing should be maintained, if not increased, to monitor the real-time point prevalence of Covid-19, so that any possible reintroduction of infected cases could be swiftly identified and isolated, and their contacts traced and quarantined.”[_]
We fully support these conclusions and hope this paper, read alongside our colleagues’ paper on ways to end the lockdown, can act as a catalyst for ensuring the UK puts in place the right exit strategy, with the right architecture and capacity on mass testing.
Chapter 15
Information is taken from Johns Hopkins School for Public Health:
Tests that have been approved for research or surveillance purposes only
Tests that are still in development
[_]
Chapter 16
Table is taken from the website of the National Institute for Health Research:
[_]