This article from the UK outlines the pros and cons of using a serology test to test the general population for the purposes of Immunity Passport. Saying that there is merit in the test when used in conjunction with the next stage of trials for the vaccine.
First, it is worth briefly discussing some of the practical considerations relating to immunity passports. This discussion will not be exhaustive, but is intended as indicative of the key factors to bear in mind in considering the introduction of immunity passports.
A significant limitation on the introduction of immunity passports is the need for a sufficiently reliable rapid test for COVID-19 antibodies.22 Antibody, or serology, tests identify whether or not someone has antibodies to COVID-19, thus indicating whether or not they have previously been infected with the virus, and whether or not they are likely to mount an immune response preventing reinfection if they encounter the virus again.23 As mentioned, it is unclear the extent and duration of immunity infection and recovery from COVID-19 will result in. The WHO has repeatedly stated that there is no evidence of lasting immunity in those recovered from COVID-19.22 24 25 This is only true on a very restrictive understanding of what counts as ‘evidence’. The only way to establish with certainty that people are immune for 1 year, 10 years or their whole lives would be to wait that long after infection and test their immunity. But this is unhelpful in the short term, and there are other ways of making predictions about COVID-19 immunity.
Patients who have recovered from their COVID-19 illness have been found to have neutralising antibodies, which inhibit virus growth.26 Whether all illness results in sufficient levels of neutralising antibodies to prevent against reinfection is still under investigation. However, experience with other coronaviruses (including viruses that cause mild illness as well as more serious diseases like severe acute respiratory syndrome (SARS) and Middle East respiratory syndrome (MERS)) suggests that antibody responses are likely to persist for at least a year and protect against reinfection at least in the short term.27 Antibody responses in SARS and MERS waned after 2–3 years, which might suggest immune passports should be time limited. Reports of individuals becoming reinfected with COVID-19 are likely to be cases where the individual falsely tested negative in the setting of prolonged viral shedding.28 Assuming that antibodies do indicate solid immunity, the difficulty lies in correctly identifying those antibodies in a way that permits testing to be scaled up significantly, without exceeding a tolerable level of false positives/negatives.
Someone who is identified as having antibodies to a particular disease—in this case, COVID-19—is described as seropositive. Those without are seronegative. Antibodies might not be identifiable in blood tests until days or weeks after illness has resolved in the infected individual.23 29 There is therefore a delay between an individual being infected with COVID-19 and them testing seropositive. A key feature of diagnostic tests is their sensitivity and specificity. Sensitivity refers to the test’s capacity to correctly identify those who have antibodies as seropositive. Specificity refers to the test’s capacity to correctly identify those who lack antibodies as seronegative. A test that is very sensitive will have a low false negative rate: there will be few people who have antibodies whom the test erroneously identifies as seronegative. A test with very high specificity will have a low false positive rate: there will be few people whom the test identifies as seropositive who in fact lack antibodies.
It is particularly difficult to develop a serological test for COVID-19 that has very high sensitivity and specificity. The test relies on producing a protein unique to COVID-19 which antibodies will bind to (if they are present in the person’s blood). COVID-19 is a coronavirus, like many common cold-causing viruses. There is a risk that, if the protein used in antibody tests for COVID-19 is too similar to proteins present in other coronaviruses, many false positives will result because people will have antibodies from infections with other coronaviruses.23 In the context of immunity passports, a false negative will mean that someone who is immune to COVID-19 will need to continue observing lockdown requirements, while a false positive could be more disruptive, indicating that someone is protected from infection when they are not. Such an individual could receive an immunity passport while still at risk of contracting and spreading the virus.
Different tests vary in their accuracy and the quality of the evidence we have about their accuracy. In particular, sensitivity of tests in the first week or two after infection may be low.29 30 The numbers of false positives and false negatives such tests produce (and how disruptive these errors are) will depend significantly on the baseline rates of infection (ie, whether or not people commonly have antibodies for COVID-19 in their bloodstream). If the rates of seropositivity are low, many of those identified as being immune will be false positives. Seroprevalence will vary greatly: in some cities it may be as high as a fifth, though elsewhere it will be much lower.31 32 Similarly, among healthcare professionals, seroprevalence is likely to be higher than among the general population.33 Low specificity tests combined with low seroprevalence will result in high numbers of false positives, while high specificity and high seroprevalence will result in far fewer.
The first rapid serological test for COVID-19 approved by the US Food and Drug Administration was called Cellex, and has a sensitivity of 94% and specificity of 96%1.34 35 The quality of tests is continually improving however, for instance, Public Health England has now approved for use a test by Roche that has at least 99.8% specificity.36 To illustrate what this means for false positives/negatives, assume 3%–5% of the UK population (67 million) has been infected (as estimated by Neil Ferguson in April 2020).37 If everyone is either susceptible or immune this would equate to around 64.3 million susceptible and 2.68 million immune people in the UK. Using the (less accurate) Cellex test would correctly identify 94% of the 2.68 million immune people as seropositive, but would also incorrectly identify 2.6 million of those who are, in fact, seronegative, as being immune. In other words, 50% of those identified as immune would actually not be immune. If, however, the more accurate Roche test were used, on a population with higher baseline immunity (say, London: population 9 million, seroprevalence around 18%38) then false positives drop significantly (only 0.2% of those lacking antibodies would be incorrectly judged as seropositive, which in London would be 14 760 people).
How disruptive is it to identify non-immune individuals as being immune? Health experts have stated that an unreliable test is worse than no test at all.15 Whether or not this is true depends on how the test is used and the context in which it is deployed. For instance, false positives could be less damaging if they occur in young, healthy individuals who are less likely to suffer severe infection, if effective contact tracing could be upscaled and if the spread of infection of the virus can be maintained at a low rate through other interventions. Further, the damage caused by false positives is only meaningful when compared with the damage caused by alternative policies such as complete lockdown, with their attendant opportunity costs. It is therefore not necessarily the case that a test with an imperfect level of sensitivity and specificity is worse than no test. For example, if we compared a policy of releasing from lockdown all members of the community (as is occurring in many places), and a policy of selectively releasing those people who have apparent immunity, the latter would lead to much slower viral spread, even if the test had relatively low specificity.
If there are low levels of immunity in the general population, a test may need high specificity and sensitivity to be informative.2 However, in populations where baseline immunity is higher the positive predictive value (likelihood that someone is seropositive, given a positive test result) will be higher. Serological testing could therefore target populations with predicted high baseline immunity, those for whom lockdown is most personally and socially damaging (for instance, healthcare workers) and those who are less likely to suffer a severe infection if they did contract the virus after receiving a false positive result (younger people with no health vulnerabilities). We also need to be sufficiently confident that serology testing is telling us something of value: specifically, that people with antibodies to COVID-19 are no longer at risk of contracting the virus, and significantly less likely to infect others.
Further logistical difficulties arise concerning how people will be tested. It has been mooted that antibody testing could take place at home, using a small blood sample from a finger poke. Samples could then either be processed at home or sent to a lab for analysis.39 At-home testing reduces the transmission risks associated with people travelling outside their home, but increases the difficulty of ensuring that tests are performed correctly and that analysis and interpretation is also correct. Alongside such quality control problems, the incentives to acquire immunity passports mean people may try to cheat the test and fake positive results. If at-home testing is not possible then the costs of testing and the time taken to perform tests will be increased.
Another consideration is who should be prioritised for testing. As mentioned, this should in part be guided by the need to avoid harmful misdiagnoses. But, given limited testing capacity, it will also be necessary to focus testing on groups whose immunity status it will be most valuable to know. It is likely that testing will be prioritised to healthcare staff and other key workers. During April 2020 around 35 000 National Health Service (NHS) staff were reported off work because they or someone in their household had COVID-19 symptoms.40 Lack of staffing places a limitation on the capacity of the NHS to treat COVID-19 (and other) patients, and thus informs the extent to which lockdown measures are necessary. This capacity could be increased by facilitating immune staff to return to work. Similarly, other key workers such as care home staff, police, supermarket staff, transport workers and others whose continued work is essential to social functioning, would likely be prioritised for antibody testing.41 The public good created by enabling these groups to move around more freely justifies their position at the front of the queue for antibody testing and immunity passports.3 ,42 It is unclear how other groups should be prioritised relative to one another, and such decisions will need to be made if immunity passports are to be introduced.