When is a screening test not a screening test?

I’ve never seen the madness of what we’re doing (with these lateral flow tests). They’re not ready, they’re not fit for purpose. I’d sooner rather them hang these tests on a Christmas tree in Trafalgar Square, that would be better.

Jon Deeks, FMed Sci. Twitter, 21.12.2020.

The COVID-19 pandemic of 2019-2020 took governments around the world by surprise, not least that in the UK. ¹ Nevertheless, and despite all the failings of political and practical public health action of the early months, the international academic community and industry, particularly bio pharma and med-tech, were galvanised into action. The manifestations of this have been found not only in the remarkable success of the programmes to develop vaccines against COVID-19 but also in the speedy development of reverse transcription polymerase chain reaction (PCR) and rapid lateral flow device (LFD) tests for SARS-CoV-2 nuclei acid fragments and antigen, respectively. ²

However the deployment of these tests has proved highly controversial, with public health academics in particular falling out over the appropriateness of the use of LFD for mass asymptomatic population testing, the controversy having been brought about by a government-sponsored initiative in Liverpool in November 2020, branded as SMART (Systematic Meaningful Asymptomatic Repeated Testing).^3–5

The Liverpool pilot included a comparison of LFD versus PCR testing in 5869 asymptomatic adults between 6 and 29 November 2020. Excluding void results, the LFD testing showed a sensitivity relative to PCR of 40.0% and a specificity of 99.9%. This would be a direct comparison as a clinical test but is not as a public health test for infectious individuals, as PCR remains positive in the post-infectious period, often for weeks. However, popular misinterpretation of these statistics as a clinical comparison with a ‘gold standard’ led to vehement criticism of the Liverpool trial, while the Liverpool team has continued to defend the approach as a valuable tool in the fight against the COVID.

Critics of LFD for mass asymptomatic population testing held the so-called gold standard criteria for evaluating screening programmes first enunciated by Wilson and Jungner in 1968. ⁶ The starting point for Wilson and Jungner’s contribution in the post-war technological age and the beckoning frontiers of genetics was the growing momentum to adopting screening approaches to a widening agenda of non-communicable conditions.

The central idea of early disease detection and treatment was seen to be bringing to treatment those with previously undetected disease while avoiding harm to those not in need of treatment. The 10 classic Wilson and Jungner screening criteria for assessing the utility of a screening test encompassed the significance of the disease as a health problem; the acceptability and availability of effective treatment that was cost effective; the availability of a suitable test that was acceptable to the public and capable of recognising disease at a latent or early symptomatic stage with the natural history and progression of undiagnosed disease being well understood; that there should be an agreed policy on whom to treat as patients; and that case-finding should be a continuing process rather than a once-and-for-all project.

At the time of Wilson and Jungner’s publication, infectious disease, outbreak control and the prospects of a novel pandemic had all but disappeared from the public health radar.⁷ Since 1968, efforts have been made to revisit the screening criteria they espoused, and it seems unlikely that they expected their offering to be set in stone for eternity. In particular, it has been argued that

although the evidence for disease/condition and test/intervention principles may best be assessed by clinical and epidemiological experts, assessing the evidence for program/system principles would require a more diverse set of experts and stakeholders, including health service programme managers, policy analysts, information system specialists, health economists, ethicists and members of both average and high risk population groups. ⁸

In addition it has been argued that many of the emerging criteria reflect broader trends that have recently shaped both Western medicine and society more generally, including consumerism and the shift away from paternalism towards informed choice and the accountability of decision takers. In the row over LFD testing, Nichol and colleagues at the University of Portsmouth have quoted Sayre’s law, that ‘the intensity and bitterness of academic disputes is inversely proportional to their importance’ but that this is an exception.^9,10 Certainly, this seems to be an example of where ‘the person who frames the question determines the range of solutions’. ¹¹

Much of the criticism of the use of asymptomatic population testing for COVID-19 using LFD comes from those who frame them as being part of a screening programme.^12–15 Wilson and Jungner’s criteria are used as a paper tiger by citing issues of sensitivity and specificity as being in breach of the golden rules of screening. In falling into this trap, critics appear to be adopting an individualistic approach more appropriate to the health prospects of individuals at risk from genetic or acquired non-communicable disease rather than the traditional population concern of public health practitioners in the face of a devastating infectious disease pandemic. With COVID-19, as with influenza, tuberculosis, syphilis and cholera before it, the priority is to locate the infectious agent as precisely as possible in time, place and person, and in preventing its replication and contagion reduce the population burden of death and long-term morbidity.

Rather than seeing the rapid finding of asymptomatic cases of COVID-19 through LFD testing as a process of screening, it is more apt to see it as part of a process of enhanced surveillance. As such, despite shortcomings, it can enable public health teams to monitor the geographic and demographic spread of the virus and focus all the public health tools, environmental, social, hygienic and therapeutic at containment and suppression. In this ambition, public motivation and engagement are critical to achieve regular testing, something that was well demonstrated in Liverpool where LFD acquired the characteristics of a democratic ‘People’s Test’, empowering ordinary people to believe that they had a part to play in their own COVID destiny. Although the UK was an early adopter of Lateral Flow methodology, Denmark, Germany and other countries are now rolling them out.

Criticisms of the approach based on unequal social uptake, flawed testing technique, gaming of results and the failure to self-isolate cannot be laid at the door of the test itself. Rather, questions of the political level of commitment to adequate levels of resourcing to local public health teams for labour-intensive public health epidemiology contact tracing and intervention together with adequate social and financial support for self-isolation is what should be in the sights of critics. Clear and accurate communication with the public about how to interpret test results is important given the chance of missing people who are positive for the virus, even at high viral loads.¹⁶

Nor has LFD technology stood still. Experience in the UK, not least in Liverpool, confirms that serial asymptomatic testing with LFD works when implemented properly and can become a key factor in the management of COVID-19 globally. The accuracy of these tests is now quite well defined, picking up 8 out of 10 with a high viral load and a false positive rate of 1 in 1000.¹⁶ Furthermore, the potential adoption of Artificial Intelligence-augmented uses of LFD holds out the prospect of improved end-to-end testing accuracy and utility and reduced falsification.

We urgently need a practical consensus on the utility of programmes of LFD testing and commitment to their implementation as an important part of an integrated approach to COVID-19 surveillance, suppression and treatment in which PCR will be seen as partner not competitor as part of a total package. A test is not primarily a screening test when it is an essential tool for epidemiological surveillance and intervention. Hanging millions of LFD on a Christmas tree in Trafalgar Square may not be the wisest move.