The Emperor has no clothes: A sober analysis of the Government response to Covid-19
Opinion piece by Eri Mountbatten-O’Malley (Edge Hill University)
The Government has been criticised for doing ‘too little, too late’. Proposals to suspend duties in the Care Act, 2014, have led Disability Rights to complain that there is ‘a real and present danger to the lives of Disabled people… effectively rolling back 30 years of progress for Disabled people’. Other regressive actions are taking place with citizens being tracked through their phones, fined for shopping for ‘inessential items’, and being watched by drones with even further restrictions promised if the public fail to adhere to Government stipulations. There is also an increasing concern about the rise of Police power with #PoliceState trending last week alongside #COVID1984. In some parts of the world citizens are being beaten with rods and in the Philippines there is live footage of President Rodrigo Duterte threatening citizens that he will order police to ‘shoot them dead’ if they fail to adhere to Government quarantine policy. Curfew is being enforced with tear gas in Kenya to disperse the crowds. The pattern is similar across the world.
The widespread perception of risk to life is of course driving public fears which has led to some of ‘the most momentous peace time restrictions on the liberty in peacetime, the Coronavirus Act 2020. Yet there appears to be widespread support for the quarantine measures. There is evidence in increases in social shaming as well and for example on twitter #COVIDIOTS is regularly trending. Some police forces are inundated with calls from diligent neighbourhood informants reminiscent of the breakdown of social fabric during long gone periods of autocratic European political history. The expansion of police and administrative powers is simply unprecedented and the social and economic fallout and ensuing human cost is unfathomable.
Although the Government must rightly do everything it can to protect the public, it must do so in ways that are proportionate to the risk. It must strike the right balance between respect for civil liberties and the legitimate aims for the protection of public health. Government justifications for the introduction of a wave of emergency powers however seems to have been predicated on misleading mortality statistics and poor methodological practices, contributing to what I term as a perfect sensationalist storm of error. This cannot be the basis for Government policy if we are to safeguard a healthy democracy.
A death, is a death, is a death… right?
The crude mortality rate in the UK (CMR) is the broadest measure of the total number of deaths from all causes in a given population, over a specific time period and is defined as total deaths per 1,000 population. Alternatively, the case fatality rate (CFR) is most often cited. Various subsets can be made using either age, gender or causal variables. In the context of an epidemic, the ‘case fatality rate’ (CFR) measures death rates in a population among diagnosed cases. This is important because CFR offers a picture of useful data for preparation of responsive services such as acute emergency care services. Importantly however, the infection fatality rate (IFR, also known as ‘true case fatality rate’, tCFR) is much more useful. This is because it helps to account for all cases of infection, including asymptomatic infections, in the wider population. The World Health Organisation (WHO) tends to publish its fatality estimates using CFR; however, as shown, this is method of measurement that targets those who seek emergency assistance and have been tested as a result of seeking emergency care. Thus, in countries where widespread testing is not instituted, such as the UK, CFR is misleading. This clearly amounts to a form of ‘selection bias’.
Unaccountable coding & risk projections
The two most influential studies in the UK to date, certainly in terms of Government policy, appear to be are Verity et all (2020), for China, and Ferguson et al (2020) from Imperial College. Verity et al use the broader CFR measure whereas Fergusson et all at Imperial College use IFR. However, Imperial use their own mathematical coding and modelling for estimating IFR. The trouble is that the ‘shocking’ Imperial College study released on 16th suggested that fatalities could be in the region of 500,000, in the UK and over 2,000,000 in the US. This study has almost single-handedly been responsible for the UK Government U-turn and a misleading characterisation of risk to public health. These figures have recently been revised down to around 20,000 of mitigated deaths. Understandably, experts keen to peer-review the coding have since raised a number of questions about the coding practices used to arrive at those figures. In response, Fergusson has tried to explain himself on twitter saying that he ‘wrote the code (thousands of lines of undocumented C) 13+ years ago to model flu pandemics’; this is far from satisfactory. Part of the problem is that the Imperial College model used projections from China and Italy to predict the rates of infection in the UK. Those are nations with significantly disparate social conditions to the UK. This was clearly a flawed standard to begin with.
Co-morbidities & post hoc fallacies
Relatedly, there is the issue of dubious death certification practices and new procedural guidance which seems to conflate COVID-19 deaths with other co-morbidities with a serious blurring of correlation and causation. There are a number of serious issues with both the revised methods of recording of COVID-19 deaths, as much as with the reporting of those fatalities. Both are having a misleading effect on the advertised numbers.
For example, guidance in the UK has recently been revised to account for COVID-19 as a notifiable disease. COVID-19 is now attributable as a ‘direct or underlying’ cause of death. However, in a note on the approach to mortality statistics, ONS stated that in publishing the figures, ‘it will not always be the main cause of death, but may be a contributory factor’ mentioned ‘somewhere on the death certificate’.
Further, Post hoc fallacies have been systematized as part of a wider certification policy-framework in numerous countries with revised reporting guidance in Italy, the US, UK and even Germany. Here in the UK, COVID-19 deaths are partly based on testing and the rather loose definition of a deceased person having ‘tested positive’ for the virus having then died irrespective of the actual cause of death. In an early report from the Italian National Institute of Health, one advisor to the Italian Government raised this as an issue there as well:
“The way in which we code deaths in our country is very generous in the sense that all the people who die in hospitals with the coronavirus are deemed to be dying of the coronavirus… On re-evaluation by the National Institute of Health, only 12 per cent of death certificates have shown a direct causality from coronavirus, while 88 per cent of patients who have died have at least one pre-morbidity – many had two or three,”
The other certification approach is based on symptoms. Guidance from NHS clarifies that if the test result has ‘not been received’ it would be satisfactory to give ‘COVID-19’ as the cause of death based on symptoms alone. However, we’ve also known for some time that COVID-19 is very similar to influenza so it’s unreliable to rely on presenting symptoms alone.
We therefore have a collection of errors from two extremes. On the one hand, the mere presence of COVID-19 through testing is enough to certify it as a cause of death (post hoc). On the other extreme, a clinical assessment based on symptoms is seen as sufficient ‘[w]ithout diagnostic proof’. The death certification and publication process is quite simply rife, layer upon layer, with confounding and misleading interpretation and representation of data.
COVID-19 in context
If we look at the Government published figures for COVID-19 as of 26th April the total is now 20,319. The seasonal flu deaths for a similar week period (using 5 year averages up to week 15) is 12,982. On the surface that seems like almost a doubling of fatalities, but once we factor that there is a two-week lag in flu statistics (only so far being published up to 10th April) compared with the figures for COVID-19 which are more or less live, we can estimate a further 4,000 in the flu back log. This would bring estimated flu deaths to at least 16,000 or more, not far off the published COVID-19 fatality figures. So, if we accept the published figures there is little difference between them over the first 15 week period of this year. In terms of overall deaths, these may well balance out with little or no excess deaths (excluding deaths indirectly related to the quarantine measures such as suicide). The BBC chart below helps to illustrate the anticipated risks overall a bit further as the pattern more or less tracks the same as if COVID-19 had not appeared:
Dr. William Schaffner, a vaccine expert at Vanderbilt University Medical Centre in Nashville, Tennessee (USA) has suggested that ‘When we think about the relative danger of this new coronavirus and influenza, there’s just no comparison… The risk is trivial’. It is suggested by at least one other a collaborative pre-review study out this week, including researchers from Stanford, suggests that the virus is likely ‘widespread’. Dr. Jay Bhattacharya, one of the researchers on that project, discussed the study in a recent interview where he suggests that the initial WHO forecast fatality rate of 1-3% was hugely out of step. Due to populational prevalence of COVID-19 (largely mild or asymptomatic), the actual fatality rates for COVID-19 is likely to be up to 10 times less severe than initially projected – closer to 0.1-0.3%. Even Fauci et all (2020) recently arrived at a similar conclusion:
“ …the overall clinical consequences of Covid-19 may ultimately be more akin to those of a severe seasonal influenza (which has a case fatality rate of approximately 0.1%) or a pandemic influenza”
In the coming weeks, all eyes will be on countries like Sweden who followed a policy comparable with the UK prior to the release of the Imperial College report on 16th March. So far, they have seen incomparably fewer deaths without any enforced quarantine. Indeed, some of the pre-review evidence from the notable Professor Wittkowski, former chief biostatistician and epidemiologist at Rockefeller University Hospital, suggests that our interventions may well have damaged our chances of reaching already herd immunity, increasing the likelihood of a second ‘rebound’. In the UK we likely peaked a week before the lockdown was even instituted, calling into question the whole efficacy of the policy. The list of dissidents against the status quo is growing day by day. Once we factor in the substantive issues raised here already, an honest assessment of the final fatality figures for COVID-19 deaths will likely be massively, not marginally, less fatal than flu.
The diminishing case for proportionality
Guidance from the WHO regulations (2005), specifically, Article 12 suggests that advice given to states should be based on ‘scientific principles’ regarding ‘assessment of the risk to human health’. As Benatar & Brock (2015: 93) aver, lockdowns should only be mandated ‘as a last resort’. Public health measures must be proportionate and should pay attention to the overall ‘net pay-off’ for mandated measures taking into account the potential harm on society. The maximal promotion of public health, no matter how conceived, should not become the ‘sole goal’ of an ethical public health policy. What this means is that for wide sweeping liberty-infringing measures to be ‘permissible’, the stakes need to be ‘very high’.
The problems I have raised here include the use of generalised CFRs infused with selection bias and flawed, unaccountable coding projections based on nations with considerably disparate socio-economic conditions. These issues have been compounded by government policies across the world which systematised post hoc fallacies. This is in defiance of well-established scientific, epidemiological and statistical best practice. Together these have contributed to creating a perfect sensationalist storm of error.
The Council of Europe advised that ‘the challenge for governments in this crisis is the ability to respond to this crisis effectively, whilst ensuring that the measures they take do not undermine our genuine long-term interest in safeguarding Europe’s founding values of democracy, rule of law and human rights’. Although, the quarantine measures seem to have met the ‘general interest of the community’ test early on, it is likely that this case is weaker by the day, particularly as the weight of evidence mounts and awareness increases against the prevalent (mis)conceptions of risk. We now have a reasonable understanding of this virus emerging from the crisis and we are slowly arriving at a consensus: that although particular groups are at risk, we are experiencing broadly similar levels of risk as standard influenza, with similar symptoms as well. I’d suggest that the stakes and the risk are therefore not at all high enough to warrant such sweeping Government measures. This is an issue that has also been recently raised by Francis Hoar QC who has undertaken a timely analysis of the relevant evidential grounds for lockdown and the ‘questionable’ scientific basis for lockdown. We need an honest assessment by the Government if we are to truly navigate out of this crisis.
The challenge for any government is to develop an approach to political decision-making that reflects an appropriate level of responsiveness to an emerging threat (apparent or actual), whilst also remaining sensitive to the need for proportionality under changing circumstances. This is a continual process. However, the UK Government response has been marked by a serious lack of transparency, regarding both the scientific basis for lockdown and the decision-making framework for easing of restrictions.
Decision-makers have the responsibility to respond to expert advice in the context of scientific best practice and legal principle with weighty (even paramount) consideration given for the very real needs and concerns of the populations they govern. If policy is based on scientific evidence, then as the evidence changes, we should see a change of both narrative and political decision-making. For some reason, this is not happening.
The consequences of prolonged lockdown are serious and I have absolutely no doubt that the human cost of these measures across the world will be so immense as to completely overshadow any perceived and debatable benefits of enforced lockdown. As Dr John Lee has suggested, ‘the moral debate is not lives vs money. It is lives vs lives’.
As a society, we must remain aware of the ever-present dangers and risks of being led by small groups of influential scientists. Laski once said, ‘we must ceaselessly remember that no body of experts is wise enough, or good enough to be charged with the destiny of mankind’. Scientific understanding is not the sole possession of a single individual or set of individuals; it is distributed, and its ‘locus is at the population level’. This ‘locus’ is a non-negotiable principle of any democracy worth its salt and we would do well to remember it.
If the evidence for risk to life is as problematic as it appears, it follows that the case for enforced quarantine and lockdown is increasingly hard to justify. If needlessly maintained, it could be noted as one of the greatest errors of political judgement and decision-making in modern history.
There is nothing virtuous in following a flawed narrative. It is deleterious to our national well-being and dangerous for the preservation of our British values, democratic norms and way of life. The question remains, who dares to doubt whether the emperor has any clothes?