Dear Members of Parliament,
I am writing to you to express my grave concerns regarding the Government’s policies in respect of coronavirus. I am an airline pilot with a major UK operator. As such I am used to processing information and analysing it logically and am not prone to hysteria. Before I was an airline pilot I qualified with a PhD in Applied Mathematics and Statistics. I spent nearly a decade as a researcher and lecturer in The Division of Epidemiology and Biostatistics at The University of Leeds and have published numerous papers in internationally refereed journals on these subjects. Much of this letter is based upon my experience in this field, in which I am qualified to hold an opinion, and the remainder I present as an extremely concerned citizen. The analyses I present in the appendices to this letter are either my own or from trusted academic sources.
The most important message to convey is that I strongly believe the Government’s response has been, and continues to be, disproportionate to the true threat posed by this virus. While this was, perhaps, understandable in March when less was known, the policies that are still in place, which are both economically and societally ruinous, are now much less credible. The Government appears to be locked into the single objective of dealing with this one virus at the expense of a myriad public health issues, many of which are exacerbated by the current COVID-centric policy choices.
The first recorded outbreak of the virus in the spring teaches us that the health impact of the virus was, in terms of clinical impact, akin to a severe influenza season. Indeed Dr Anthony Fauci said in the New England Journal of Medicine in February that the “clinical consequences of Covid-19 may ultimately be more akin to those of a severe seasonal influenza”. The data both in the UK and worldwide have borne this out. The mortality burden of COVID-19 in the UK has been similar to the relatively severe 2018/19, 1998/1999 and 1999/2000 influenza seasons, and significantly lower than the 1968 H3N2 influenza pandemic which killed approximately 80,000 people in the UK. These outbreaks were as severe, if not more so, than the current COVID epidemic and yet the country was not closed down risking economic ruin and serious long-term public health consequences.
Appendix A deals with the statistics of the present epidemic and its associated mortality burden. It demonstrates that the mortality burden is felt substantially by the over 70s and even then the majority of those individuals (over 90%) have one or more co-morbidities so that recovery from any respiratory illness is likely to be compromised. This defines a clear vulnerable group who may need to take some extra precautions during the epidemic. It is not the function of the state to force them to do so however: they are old enough and wise enough to make their own decisions if provided with an objective assessment of the true risks they face. In this regard the Government has completely failed in its duty to its citizens by instilling fear rather than providing rational and proportionate public health messaging.
SAGE minutes from March 22nd on ways to increase adherence to social distancing contains the following: “[t]he perceived level of personal threat needs to be increased among those who are complacent, using hard-hitting emotional messaging”. The proposed means of achieving this include advice such as “use media to increase sense of personal threat” and “consider use of social disapproval for failure to comply”. Since the spring the media been keen to maintain the hard-hitting emotive narrative of a deadly virus that is dangerous to everyone, which is simply not supported the data. Rather than providing simple, effective and proportionate public health education the Government and media opted for a policy of “psychological warfare” against our citizens.
In terms of the wider public health effects, during the lockdown there was a 50% reduction in A&E presentations for heart attacks according to the British Medical Journal's Open Heart. A similar figure applies to strokes. Both of these conditions have poor outcomes unless treated in a clinical setting. The lack of A&E presentations is thus a serious concern because it points to significant excess mortality for those conditions.
Cancer diagnosis and treatments were, and continue to be, severely disrupted with Cancer Research UK estimating that on average 2,500 cancers were going undiagnosed each week during the lockdown. A study conducted by DATA-CAN, the Health Care Research Hub (HDR UK) for Cancer, estimates that the number of excess cancer deaths attributable to the disruption of cancer care could be as many as 35,000. The additional mortality burden from just these three conditions is likely to be broadly similar to that of COVID-19 and those additional deaths are attributable to the Government’s policies which have clearly dissuaded, and continue to dissuade, contact with the NHS. Recently released SAGE minutes estimated that the indirect mortality burden caused by the coronavirus measures could be as high as 75,000.
All of this is without considering the long-term damage done to children’s education, widespread mental health issues from loneliness and reduced social interactions, failed businesses and rapidly-increasing unemployment despite the Treasury’s Furlough and Job Retention Schemes, which are merely delaying the problem because the markets supporting those jobs are collapsing.
The origin of the first of many U-turns from a policy of “herd immunity” to more stringent measures seems to be Imperial College’s Report 9 based on the modelling of Neil Fergusson. Appendix B deals with some aspects of the credibility of this model and its assumptions. Although framed as a “reasonable worst case scenario”, it was evident to other eminent scientists that the scenario therein was most unlikely to be a credible one. Professor Johan Giesecke (former Swedish State Epidemiologist and current member of the World Health Organisation’s Strategic and Technical Advisory Group for Infectious Hazards) was one such doubter as he made clear in a media interview with Freddie Sayers on 17th April. Professor Michael Levitt, a Nobel Prize winner, also predicted that Imperial’s modelling would over-predict deaths by at least a factor of ten. Professor Sunetra Gupta of Oxford University doubted the Imperial model and its high death toll was credible. Last month John Ioannidis, Stanford professor and the most referenced scientist in the world, released a preprint paper demonstrating that lockdowns in Europe had little to no effect and that the Imperial modelling is not robust and built on faulty assumptions.
These people are not crackpots and their predictions of a sharp rise of cases for a month, followed by a long tail of two or three months, have been correct from the outset because they predicted SARS CoV-2 would behave like every other respiratory disease in the northern hemisphere. This has happened irrespective of the severity of lockdown with Sweden being the oft-quoted comparator. I would encourage you to seek out their various interviews, both video and written, because their narrative is different and, importantly, they have been correct in their predictions.
From the beginning the Government claimed to be “following the science”, giving the impression of a definitive path, when in fact the science was far from a settled matter and there were alternative viewpoints which should have been considered and debated. Aside from the occasional disquiet among members of SAGE, there are no signs that the Government seriously sought any alternative viewpoints, and understanding why they did not will be a key matter for the inevitable Public Inquiry.
More recently, a letter to the UK Government from a panel of experts led by Professors Carl Heneghan, Sunetra Gupta and Karol Sikora, and another signed by 66 General Practitioners led by Dr Ellie Cannon, highlighted the additional public health effects of the Government’s continued COIVD-centric approach, including physical and mental health effects. Abroad a similar letter from Belgian GPs to their Government and The Great Barrington Declaration and Petition are examples demonstrating that scientists and doctors with “dissenting views” are, rightly, becoming more vocal and insisting on a meaningful scientific debate.
Turning to the matter of a vaccine it should be noted that there hasn’t been a successful vaccine for a coronavirus in humans or animals. The flu vaccine, for example, has failed to reach 50% efficacy in seven of the last ten years. There have been recent murmurings of the possibility of compelling, either by mandate or “social coercion”, the population to take this vaccine once it is available. It is extremely important to be aware that forced or coerced vaccination would contravene the Nuremberg Code, which the UK was instrumental in creating, protecting any person from medical interventions without informed consent. There will be many people who are well-informed of the true risks of this disease who will exercise their right to deny consent for an expedited, essentially experimental, vaccine, the longer-term health effects of which are by definition unknown. As an example the Swine Flu vaccine was found to cause narcolepsy in some individuals.
It does seem that the Government’s policy endpoint is presently mass-vaccination supplemented by mass-testing until the vaccine is available. While I am not implying the pharmaceutical industry is behind a conspiracy it is nevertheless an industry which has a history of exploiting public health scares for its own profits. Following the 2009/2010 H1N1 pandemic the European Council launched an investigation into the influence of the pharmaceutical industry on the WHO and the global swine flu campaign. This was seen as a step towards improving transparency of what it called “the Golden Triangle of corruption” between the WHO, the pharmaceutical industry and academic scientists. The Parliamentary Assembly of the European Council adopted Resolution 1749 (2010) calling for more transparency and changes to the handling of future pandemics. A selection of the salient points, which are instantly recognisable and applicable to the present COVID “pandemic”, are reproduced in Appendix C. Regrettably little appears to have been learnt in the decade since this resolution was adopted.
Prior to a potential vaccine the Government appears to be relying on mass-testing as a means to identify viral outbreaks; the so-called “Operation Moonshot”. This is also a poorly-thought-out policy with a questionable scientific basis. Firstly there are doubts about the diagnostic ability of the PCR technique: PCR tests were never intended for diagnosis, and in particular for SARS-Cov2 there exists no “gold standard” against which to assess their diagnostic capability. Appendix D deals with the significant issue of false positives, which is a substantial flaw of mass-testing schemes. Briefly, throughout the summer the virus prevalence was low according to The Office for National Statistics. In these circumstances the testing produces mostly false positive results (around 90% of positive tests are false positives), even at the apparently-low 0.8% false positive rate quoted by the Health Secretary. Belgium, for example, terminated its community testing programme in September in part due to these issues.
Through the summer months the COVID mortality rate of hospitalised patients fell to the normal background hospital mortality rate of around 1.7% and the strong correlation between hospitalised cases and deaths, which was present in the spring, completely decoupled. This is suggestive of the majority of summer positive results being false positives. This is bad news for two reasons: first uninfected individuals have been labelled as “cases” on the basis of a false positive test and second, the high proportion of false positives obscures any real increase in the viral prevalence should it occur. Essentially the Government was “working in the dark” and basing its summer and autumn policies on flawed tests and data.
Surveillance data from NHS Triage, The ONS population sampling campaign and the UCL Zoe App all show a plateau or even a reduction in the number of actual symptomatic cases in the last fortnight. The official testing data however, still shows a significant rise. As one of several examples of data-illiteracy in Public Health England, every positive test is reported—even repeated positives—but individuals with negative results are only reported once even if they test negative multiple times. This failure to accurately capture repeated testing is artificially inflating the proportion of positive tests.
You may also recall that PHE were recording deaths of persons who tested positive at any time previously. An individual with a positive test who made a full recovery, but subsequently died months later of an unrelated cause, was considered a COVID death. Fortunately this bizarre practice was highlighted and stopped, but it seriously calls into question the veracity of the Government statistics and official data, which have consistently appeared to inflate the magnitude of the COVID issue. These are not isolated examples, but are, perhaps, the most well known.
Appendix E deals with the present trajectory of the so-called “second wave”. On the basis of the available data it is not appropriate to call the trajectory a second wave, and I believe it was irresponsible of the Prime Minister to declare to the media that we “are in the second wave”. I also believe it was highly questionable that the Chief Scientific Advisor presented to the public essentially a guess as to case numbers at various time points which, although he claimed was “not a prediction”, nevertheless has the capacity to reinforce fear in members of the public who are less well informed or rely on the Government and mainstream media for their information.
The data are once again misrepresenting the true situation because they are reported on the basis of absolute numbers rather than per 1,000 tests or a similar standardising measure. The increases in tests performed, coupled to the false positive issue and incorrect reporting of multiple testing mentioned earlier, can account for much of the apparent-rise in “cases”. Appendix E shows some of these data presented in a correctly standardised format based on work by Professor Norman Fenton, Professor of Risk and Probability, at Queen Mary University London.
Much of the “second wave theory” is built on the Spanish Flu pandemic which was over a century ago. It is also widely believed that the apparently-larger second wave of Spanish flu was in fact a completely different pathogen and so the comparison is probably flawed. The most likely outcome for SARS-CoV2 is a seasonal endemic respiratory virus; the same as most of the pandemic-causing viruses of the past, including the 1968 H3N2 virus, which remains endemic and in circulation today.
I have concerns over the way that the Government imposes it’s ad-hoc policies through the Public Health Act (1984). This act gives Ministers the same powers as Magistrates and allows the confinement of infected individuals for the prevention of infection or contamination. I, and many other people more schooled in the law than me, believe the Government is overstepping its authority under this act. Retired Supreme Court Justice Lord Jonathan Sumption has been vocal on this issue and the Government’s avoidance of scrutiny through the use of this act. The act of parliament that does confer the rights to take some of the steps the Government has is the Civil Contingencies Act, but measures taken under this act are for a very limited duration and subject to significant parliamentary scrutiny. This is a Government that seems to prefer to avoid scrutiny and debate, for example the attempted prorogation of parliament in the final act of the Brexit debate. I believe at this stage wider scrutiny of all the Government’s policies and scientific data would be appropriate. Some progress was made in this regard by the recent actions of Sir Graham Brady, but I still feel that the concessions made by the Government did not go far enough.
Much has been made of the “Swedish model” for handling the virus and I have always believed their approach was significantly more sensible. Appendix F looks at the Swedish epidemic which should teach us, among other things, that lockdowns do almost nothing to prevent the spread of disease epidemics. Swedish education has been significantly less disrupted, the Swedish economy is less impacted and importantly their policies have the substantial benefit of being sustainable over a longer period of time. Video from Sweden at the height of the epidemic showed life continuing largely as normal: people were shopping, visiting restaurants and bars and were not wearing masks. Sweden’s state epidemiologist Anders Tegnell realised at an early stage that it might be necessary for society to live with this virus, as we have many before, over an extended period of time while natural immunity was established. Fortunately natural immunity may not be as far away as first believed—see Appendix F. This will of course negate the need for widespread vaccination outside the most vulnerable and any kind of digital tracing or “health passes”. Doubtless the pharmaceutical and biotech companies won’t agree with this and will lobby for their adoption, but I am certain we have progressed to a stage of this epidemic where these measures are unnecessary and can be written off as the hideous apparatus of a police state: the total antithesis of a civilised and free western society.
I believe this is the most important single issue for our country in our lifetime. If you are accepting the Government’s narrative at face value I would ask that you use the information in this letter and the appendices to consider an alternative view which has over recent weeks and months gathered momentum among the scientific community and public alike. The Government’s ad-hoc policies have been confusing, largely inconsistent and have prevented businesses and individuals from forming coherent plans. The quarantine of those arriving from abroad, for example, is based on a completely arbitrary unscientific threshold of 20 “cases” per 100,000 and is destroying the travel industry, a vital economic engine, en masse. The Government is undeniably “fiddling while Rome burns”.
Holding the British public and British businesses in the current state of purgatory is not a viable long-term strategy because it inflicts incredible economic and social damage. I am determined that the the post-Brexit dream of “Global Britain” does not die in this economic and socially-destructive nightmare before it begins. As the divisions in Parliament become more apparent in the coming days I would ask that you use the information I have provided to assist in deciding where to place your support. The stakes are extremely high: what happens in the next few weeks and months will determine our country’s path and prosperity for decades.
APPENDIX A. UK MORTALITY SITUATION FROM ONS DATA
It is a simple matter to verify which age groups are most affected by mortality from COVID. By plotting the UK mortality data up to and including the week ending 18th September (week 38) reveals that:
- excess deaths are occurring in all age-groups over 45
- A substantial proportion of excess deaths due to COVID and non-COVID causes are in the over 70s.
Note that over 90% of COVID deaths are to those with a pre-existing medical condition, most notably heart disease and diabetes
Fig. A1. Age distributions of COVID and non-COVID deaths for females and males.
Using the Office for National Statistics Deaths registered weekly in England and Wales for 2020 to Week 38 (ending September 18th) gives Fig A2.
Through the peak of the pandemic mortality was above the five-year average for 12 weeks, producing an excess mortality burden of approximately four weeks of normal-rate mortality.
Between the week ending 26th June and the week ending 14th August mortality had returned to below average levels. The heatwave in late July and early August accounted for some above average mortality showing in the data throughout the month of August as explained in the ONS’s Statistical Bulletin for the week ending 14th August 2020. (https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/bulletins/deathsregisteredweeklyinenglandandwalesprovisional/weekending14august2020). At the beginning of September mortality is approximately average, or very slightly above, mostly not driven by COVID which accounts for only around 1.5% of all deaths in England and Wales.
Fig. A2. Deaths in England and Wales by week of death occurrence compared to five year average mortality.
Focussing on the peak of the epidemic between late March and early May there is significant above average mortality from non-COVID causes—that is those deaths where COVID is not mentioned on the death certificate. Concentrating just on the peak for the time being the chart below shows the excess deaths that did not mention COVID on the death certificate.
Throughout the peak of mortality the deaths from non-COVID causes were above the five year average by around 2500 each week for five weeks. A total of 13,237 excess deaths not attributed to COVID in six weeks through the peak of the epidemic (Fig. A3).
The media often sought to attribute these deaths to “undiagnosed” or “unreported” COVID, but we can dispel this theory by looking at the population fatality risk profiles of the COVID and non-COVID deaths.
Fig. A3. Non-COVID excess deaths above five year average through the peak of the epidemic.
Fig. A4. Population risk profile for COVID deaths on a linear scale (left) and logarithmic scale (right).
Fig. A5. Population risk profile for non-COVID deaths on a linear scale (left) and logarithmic scale (right).
Fig. A4 shows the population fatality risk for COVID deaths. These plots show that the risk of death from COVID is extremely low until around age 70 and the risk grows approximately in proportion with an individuals background mortality risk.
There is an important difference between males and females COVID mortality risks: for a given age the risk of death for a male is approximately twice that of a female.
Fig. A5 shows the risk profile for the non-COVID excess deaths. The population fatality risk does not differ markedly between males and females for non-COVID excess deaths. This tells us that the majority of these "extra" deaths are not due to under-reporting of COVID: if they were they would share a similar risk profile to the COVID deaths. They are in essence “collateral damage” caused by the disruption of the health service throughout the peak in April and May.
Fig A6. shows the same ONS data from the point of view of place of death. We see that the outbreaks in care homes drove much of the mortality in the peak. Most of the non-COVID mortality through the epidemic peak was driven by care homes and private homes rather than hospitals.
The most recent data throughout August shows a serious malfunction of the NHS even now so far after the epidemic peak. Hospital deaths remain appreciably below average while home deaths are significantly above average.
Fig. A6. COVID and non-COVID deaths by place of death.
Fig. A7. COVID and non-COVID deaths by place of death for the 12 weeks to 18th September 2020.
Focussing on the last 12 weeks highlights the scale of the problem. Home deaths remain further above the five year average than hospital deaths are below. This highlights that the current policies are feeding both excess mortality and mortality exportation from hospitals to private homes.
- Deaths registered weekly in England and Wales, Office for National Statistics. Available from: https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/bulletins/deathsregisteredweeklyinenglandandwalesprovisional/latest
APPENDIX B. CREDIBILITY OF THE IMPERIAL COLLEGE MODEL
The change of policy that occurred in March from what was termed “herd immunity” to the lockdown approach appeared outwardly to be the result of Imperial College Report 9 . This report, which it should be noted at the time was not peer reviewed for publication in any scientific journal, seemed to form the basis of the first of many Government U-turns in the coronavirus saga.
Firstly, to use an unreviewed piece of work as a basis for policy is highly questionable. This is work whose assumptions and data have not been tested or scrutinised by others in the field. The model’s computer code had not been inspected in detail by experts. It took some weeks for Imperial to provide their computer code for examination on the publicly available “github” site, and even then not before it was modified so that the Report 9 results could not be identically reproduced. The peer review process, while far from perfect, is still the accepted standard for the assessment of scientific work.
Fergusson et al, suggested that the UK would suffer 510,000 deaths without immediate non-pharmaceutical intervention (lockdown). Although presented as a reasonable worst case scenario it was clear to some, even at the time, this was not remotely credible.
Nobel Laureate Professor Michael Levitt FRS was able to accurately predict the courses of the COVID epidemics in numerous countries including the UK and China using a data-led approach only. This data-led approach demonstrated as far back as March that the epidemic would amount to four normal-weeks-worth of excess deaths and not the 510,000 claimed by Imperial. The UK death toll of just over 45,000 is approximately equivalent to about four weeks of normal death.
The Imperial model also included a number of highly questionable assumptions, the most important of which was the use of 0.9% infection fatality rate—the rate of death among anyone infected (including mild or negligible symptoms), which differs from the case fatality rate—the rate of death among those identified medically as cases through symptoms and testing. It’s almost always the case the that the estimation of infection and case fatality rates during epidemics is on the high side. For example the lowest infection fatality rate calculated during the swine flu was in fact five times higher than the ultimately-accepted value post-epidemic.
The Imperial Paper used the work of Verity et al  (another paper which has not been peer reviewed) to determine the infection fatality rate to be in the 1% region, but even at the time this paper was released the it was already apparent that infection fatality rate was somewhat lower than 1%, particularly in individuals of economically-productive age. Indeed the Verity paper estimated the infection fatality rate for under 60 years of age to be 0.15% and for those aged 60 and over to be 3.3%. Even these estimates are subject to enormous uncertainty because of how the authors estimated the infection prevalence.
The infection prevalence in Verity et al, and therefore ultimately the Imperial model, was estimated using 689 individuals who were repatriated from Wuhan on flights over a three day period that tested all individuals (regardless of symptoms) for infection immediately upon arrival. Testing following this repatriation yielded 6 positive individuals. So the Verity et al infection fatality rate is substantially based on a point estimate of prevalence from just six individuals.
As far back as March the Centre for Evidence Based Medicine were estimating the infection fatality rate as 0.1% to 0.4% (between 0.5% and 0.8% lower than Imperial). In an Editorial in the New England Journal of Medicine (February 28th), Dr Anthony Fauci said
“that the overall clinical consequences of Covid-19 may ultimately be more akin to those of a severe seasonal influenza”
Another paper , published in pre-print on March 9th, gave the estimated infection fatality rate of the Diamond Princess cruise ship as 0.5%. It should be noted that the Diamond Princess was as close to a closed experiment as we have in a relatively infectious environment with a cohort biased towards the older, more vulnerable group, and even then the infection fatality rate was estimated at only 0.5%.
Iceland, benefitting from a relatively small population, had tested a significant proportion of its inhabitants during the epidemic and had an infection fatality rate of approximately 0.2%. Work by Hendrick Streeck, director of the Institute of virology and HIV Research at the University Bonn, yielded an infection fatality rate of around 0.26% , while later, in the summer, the American Centre for Disease Control and Prevention (CDC) also updated it’s estimate of the infection fatality rate estimate to at most 0.25%.
All of this suggests that Fauci was correct in his February editorial in saying that the impact of COVID was likely to be similar to that of a severe seasonal influenza (infection fatality rate approximately 0.1% to 0.2%). Unfortunately there was a mistake in Fauci’s editorial which appears to have gone unnoticed by the peer reviewers in which he said that the case fatality rate for influenza was 0.1% as opposed to the correct statement that the infection fatality rate of seasonal influenza is 0.1%. This caused confusion in some media outlets but it is clear that Fauci meant “the infection fatality rate is 0.1%” because he had additionally commented that the impact of COVID would be similar to that of seasonal influenza.
Using an infection fatality rate of 0.9% for modelling which would be used to steer national policies seems difficult to justify because it is based on data that is subject to huge uncertainty, especially given that many institutions, expert bodies and real-world observed data were already suggesting the infection fatality rate was far lower. This highlights also why “worst case scenarios”, even reasonable ones, should not be communicated to the public: they have a tendency to cause unnecessary alarm. The Government and its advisors need to retain the flexibility to revise their estimates in line with more recent or more complete data without becoming prisoners of an over-pessimistic scenario, based on potentially unreliable data and assumptions, that is already in the public domain.
Last month John Ioannidis, Stanford professor and the most referenced scientist in the world, released a preprint paper demonstrating that lockdowns in Europe had little to no effect and that the Imperial modelling is not robust and built on faulty assumptions . While not reaching the extreme of denying lockdown has any value, they are nevertheless certain the benefits are marginal, potentially not worth the collateral public health damage and their paper culminates in the following observation:
“Given the analyses that we have performed using the three models that the Imperial College team has developed, one cannot exclude that the attribution of benefit to complete lockdown is a modelling artefact.”
Imperial’s pre-COVID modelling track record is abysmal and certainly not one to be proud of. It should also have called into question his credibility as a Government advisor at an early stage although, alas, it did not. His previous grossly-exaggerated predictions for previous disease epidemics include:
- Foot and Mouth: 50,000 predicted deaths; actual number circa 200 (250 times over-estimate).
- Bird Flu: 200,000,000 predicted deaths; actual number circa 400 (500,000 times over-estimate).
- Swine Flu: 65,000 predicted deaths; actual 457 deaths (150 times over estimate).
Imperial’s predictions for the number of COVID deaths in the first UK epidemic peak were an over-estimate by a factor of 12. Imperial’s modelling doesn’t have a great record unless your metric is gross pessimism.
While the Government’s initial reliance on the unpublished Report 9 may have arisen from a perceived need for urgency at the start of the epidemic, the continued use of Imperial’s modelling should have subsided in favour of a data-driven approach and more up-to-date information as the epidemic developed through the spring and early summer.
- Ferguson N et al. Impact of non-pharmaceutical interventions (NPIs) to reduce COVID-19 mortality and healthcare demand. Imperial College London (16-03-2020), doi: https://doi.org/10.25561/77482.
- Verity R, Okell LC, Dorigatti I, et al. Estimates of the severity of COVID-19 disease. medRxiv 2020; Available from https://www.medrxiv.org/content/10.1101/2020.03.09.20033357v1.
- Fauci A et al. COVID-19—Navigating the Uncharted. Editorial, New England Journal of Medicine.
- Russell, T et al. Estimating the infection and case fatality ratio for COVID-19 using age-adjusted data from the outbreak on the Diamond Princess cruise ship. Eurosurveillance doi: 10.2807/1560-7917.ES.2020.25.12.2000256
- Streeck H, et al. Infection fatality rate of SARS-CoV-2 infection in a German community with a super-spreading event. https://doi.org/10.1101/2020.05.04.20090076
- Chin V, Ioannides et al. Effects of non-pharmaceutical interventions on COVID-19: A Tale of Three Models. https://doi.org/10.1101/2020.07.22.20160341
APPENDIX C. THE 2010 H1N1 “PANDEMIC” AND RESOLUTION 1749 (2010) OF THE EC PARLIAMENTARY ASSEMBLY
After the 2010 H1N1 (“Swine Flu”) pandemic the European Parliament passed a resolution calling for an inquiry into the role of the pharmaceutical industry on the global Swine Flu campaign, focussing especially on the industry’s influence on the WHO. A selection of the salient points, which are instantly recognisable and applicable to the present COVID “pandemic” are:
- It [The EC Parliamentary Assembly] is particularly troubled by some of the consequences of decisions taken and advice given relating to the pandemic leading to the distortion of priorities of public health services across Europe, the waste of large sums of public money and unjustified fears about the health risks faced by the European Public at large.
- The Assembly notes that grave shortcomings have been identified regarding the transparency of decision-making processes relating to the pandemic which have generated concerns about the possible influence of the pharmaceutical industry on the major decisions relating to the pandemic.
- In Recommendation 1908 (2010) on lobbying in a democratic society (European Code of Conduct on Lobbying), the assembly noted that unregulated or secret lobbying may be a danger and undermine democratic principles and good governance.
- Furthermore, the Assembly regrets that WHO has not moved swiftly to revise or re-evaluate its position on the pandemic and the real health risks involved, despite the overwhelming evidence that the seriousness of the pandemic was vastly overestimated by WHO at the outset. In addition, the Assembly regrets the highly defensive stance taken by WHO, whether in terms of being unwilling to accept that a change in the definition of a pandemic was made, or an unwillingness to revise its prognosis on the pandemic.
- [R]evising and updating existing guidelines on working with the private sector or preparing such guidelines where they are lacking, in order to ensure that:
- a wide range of expertise and opinions is taken into account, including contrary views of individual experts and opinions of non-governmental organisations;
- declarations of interest of experts involved are made public without exception;
- closely collaborating with the media in order to avoid sensationalism and scaremongering in the public health domain;
- The Assembly also calls on member states to:
- develop systems of safeguards against undue influence by vested interests if they have not yet done so
- ensure that the private sector does not gain undue profit from public health scares and that it is not allowed to absolve itself of liabilities with a view to privatising profits whilst sharing the risks.
While there are people who might seek to label as “conspiracy theorists” those who might question the involvement of the pharmaceutical industry, there is clear history in this regard and a sufficient level of suspicion for the EC to investigate and adopt a resolution.
- Handling of the H1N1 pandemic: more transparency needed. Resolution 1749 (2010) of the Parliamentary Assembly of the Council of Europe.
APPENDIX D. FALSE POSITIVES AND MASS-TESTING IN LOW PREVALENCE
Throughout the summer months the Government has ramped up testing for COVID-19, but once again, this was entirely the wrong strategy and, fundamentally, a waste of resources.
No test gives a 100% accurate result; tests need to be evaluated to determine their sensitivity and specificity, ideally by comparison with a “gold standard”. The lack of such a clear-cut “gold-standard” for COVID-19 testing makes evaluation of test accuracy challenging.
Any medical test accuracy should be separated into two components: sensitivity and specificity.
- The sensitivity is the proportion of positives that are correctly identified: that is those who get a positive test result and really do have the disease (true positives).
- The specificity measures the proportion of negatives that are correctly identified: those who get a negative test result and really do not have the disease (true negatives).
There has been concern that PCR testing is not appropriate for this kind of virus. The PCR process works by amplifying even the smallest amount of viral RNA, whether it is viable infectious virus or not. Once an infected person has successfully recovered from infection they will be ‘shedding’ viral RNA for many weeks. This RNA is non-viable and could not lead to infection in another person but will still be detected if the cycle threshold (Ct) of the test is set too high.
Research published in the Journal of Clinical Infectious Diseases  suggests that infectious virus is not present for Ct > 24, whereas the UK has been using cycle thresholds in excess of 40 throughout the summer. This means that many of the positive PCR results obtained will be die to non-infectious viral shedding and pose no danger to other individuals. In the research mentioned, Ct < 24 led to a test specificity of approximately 97% (that is a 3% false positive rate).
A recent paper in the British Medical Journal based on systematic reviews of current COVID tests  suggests realistic test sensitivity and specificity values of 80% and 95% respectively (5% false positive rate).
The false positive rate is exceptionally important, particularly in times of low prevalence, such as during the summer months. Under normal medical protocols a single test result is insufficient justification for a diagnosis. For example, for breast cancer screening, if the only known information was a positive mammogram, there is still a 91% chance the patient does not have breast cancer. In order to make a definitive diagnosis the clinician needs additional information or symptoms to be present.
For some reason with COVID however, we are justifying cases on the basis of a single positive PCR test result. I am not sure why or how we have allowed this contortion of normal medical diagnostic procedure for this virus, but it has serious consequences and may be leading to poor decisions on the basis of an exaggerated sense of risk.
During the peak of the epidemic we tested mainly those patients who were hospitalised and were performing around 15,000 tests each day. Allied to symptoms it is likely that those testing positive really did have COVID present whether they contracted it in the community and were hospitalised or they were admitted to hospital COVID-free and contracted it in hospital (nonsocomial infection).
During the summer months since the peak of the epidemic the virus prevalence in the community has been low—estimated by ONS in July to be around 1 in 2,300 persons and in September to be 1 in 1,400 (95% credible interval: 1 in 1,900 to 1 in 1,000).
Although the Health Secretary suggested that only those with symptoms should take tests it is clear from reports of door-to-door testing in, for example, Birmingham, that this has not been the case. As an example consider 10,000 random individuals presenting for a test. Assuming the worst case of 1 in 1000 individuals with the disease. Of the 10,000 individuals and 1 in 1000 prevalence we expect that 10 individuals presenting to truly have COVID.
Using the BMJ’s figures of 80% sensitivity and 95% specificity yields 8 true positives, 2 false negatives and 500 false positives (5% of the 9,990 true negatives). Of the 508 positive results only 8 (1.4%) are correct and 500 (98.6%) are false positives. The false positives completely swamp the true prevalence. To use an engineering analogy, there is lots of noise and very little signal.
This is not only problematic from the point of view of incorrectly diagnosing individuals but also masking any real increase in virus prevalence should it occur. The more tests are performed the worse the false positive situation becomes in low prevalence.
For example the Government’s “Operation Moonshot” seeks to test 10 million individuals each day. Clearly the majority of these individuals will not have COVID and on average at 1 in 1,000 prevalence (the upper end of the current ONS estimate) we would expect that 10,000 would have virus present. This is before we concern ourselves with whether they would have symptoms or not (the vast majority of infections showing no or only mild symptoms).
The Health Secretary recent stated on Talk Radio that he believed the false positive rate was 0.8%, which is the lowest value presently accepted in the literature. Bearing in mind the BMJ review , I suspect this is optimistic, but let’s take him at his word. Then assuming we detected all of the 10,000 infected persons correctly, Moonshot would produce 799,200 false positives each day (89% of all positive tests are false positives). Although those individuals with real infection (symptoms) would be preferentially tested, the vast number of false positives produced from such a community mass-testing programme would swamp the true positives many times over.
Even in the most extreme case of the virus vanishing off the face of the planet, Moonshot would still be producing circa 800,000 false positives every day with no virus present.
This also conveniently highlights one reason why zero COVID, which is the preferred position of some individuals, makes no sense. Even with no virus present testing would return a significant number of false positives. Because any death within 28 days of positive test is attributed to COVID, this means that we would continue to incorrectly ascribe deaths as COVID based on false positive test results even if the virus had “vanished” long ago. In other words, we can never reach zero by testing and attributing deaths on the basis of test results alone.
Clearly mass testing, although suggested by the WHO and being current Government policy, is highly problematic. Recently Belgium to scrap its community testing scheme (equivalent of UK Pillar 2 tests) altogether and I believe they are right to do this.
- Bullard J, Dust K, Funk D et al. Predicting infectious SARS-CoV-2 from diagnostic samples. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America 2020
- Watson J, Whiting P, Bush J. Interpreting a covid-19 test result. BMJ 2020;369:m1808. https://www.bmj.com/content/bmj/369/bmj.m1808.full.pdf
APPENDIX E. THE UK “SECOND WAVE”
The current “second wave” needs putting in context. On 21st September Patrick Vallance showed a graph to the nation of “cases” doubling every seven days. He stated that this “wasn’t a prediction”.
Either he was presenting total guesswork and nonsense, in which case he is not a credible scientific advisor and should resign, or this was a prediction and therefore I, and other citizens, are entitled to measure him against his presentation.
Fig. E1. Vallance projection with actual cases by reporting date and 7-day average. Source: Centre for Evidence Based Medicine, University of Oxford.
As a statistician I was horrified by the presentation of such a graph to the general public. This was one of the worst examples of statistical reporting I have ever seen, which is an achievement given the consistently-poor quality of the Government’s statistics and their presentation throughout.
First, the presentation of raw numbers is an appalling mistake because it does not account for the vast increase in testing between the spring and now. In the spring, there were around 15,000 tests a day. This has reached in excess of 200,000 each day by the autumn. Even a statistics undergraduate should be aware of the need to standardise raw numbers by the number of tests performed and yet, inexplicably, the Government and its advisors are not. At the very least this is grossly incompetent.
Using the simple plot of daily “number of cases” (which actually means “number testing positive”) is an inappropriate way to monitor the COVID trend. This is because much of the increase in “cases” is explained by an increase in the number of people tested, and that many of the “positives” are either false positives or people who are asymptomatic (or who will only ever have mild symptoms). A better, but still simple, alternative plot is daily “cases per 1000 people tested”.
Fig. E2. New COVID positive tests per 1,000 tests. Source: Professor Norman Fenton, Queen Mary University London via http://probabilityandlaw.blogspot.com
In April we were only testing those hospitalised with severe symptoms; this is generally true and in fact the testing ‘strategy’—i.e. who primarily gets tested has changed several times. The proportion of positives among those tested in April was obviously higher than now. It is also the case however, that the proportion of people now testing positive who will be totally unaffected by the virus (whether they have it or not) is also much higher.
That is why we need to distinguish between mild, severe and asymptomatic cases. We really need to see the plot of severe cases, and hospitalisations is the best approximation we have for that.
However, that is also somewhat compromised because we are now entering the normal flu season when hospital admission inevitably rise significantly—but it is the best proxy that we have in publicly available data.
Fig. E3 shows the raw number of COVID-19 patients admitted to hospital in recent weeks in context with the original peak in the spring. Even this plot, which is incorrectly not standardised by persons tested, shows that at present the increase in the number of individuals testing positive is not exponential as suggested by the Government on numerous occasions.
Fig. E3. New COVID hospital admissions in raw numbers. Source: UK Coronavirus Dashboard.
However, even if we ignore a major problem with the data for Wales, where suspected COVID cases are included in the hospitalisations, which artificially inflates admissions data, there is still a problem in using this plot to monitor the progress of COVID. To see why, here is what the website says about the England data:
"data include people admitted to hospital who tested positive for COVID-19 in the 14 days prior to admission, and those who tested positive in hospital after admission. Inpatients diagnosed with COVID-19 after admission are reported as being admitted on the day prior to their diagnosis"
In other words, we have no idea how many of the 'COVID hospital admissions' were people actually admitted because of COVID. A person entering hospital for, say, cancer treatment who recently tested positive for COVID will be officially classified as a COVID case. The same is true of those entering hospital for any treatment who have not previously been tested positive for COVID but test positive at some time during their stay. We have to therefore assume that - as in the public generally - a proportion of the “COVID hospital admissions” are people who either:
- don't have COVID;
- are asymptomatic (or who will only ever have mild symptoms).
It also means that, as with “COVID cases”, most of the recent increase in “COVID hospital admissions” may be explained by the general increase in number of people being tested.
Unfortunately, data are not provided for number of people being tested before and during UK hospital admissions. It is not however, unreasonable to assume the number is roughly proportional to the total number of people being tested in the UK. So it makes sense to plot the daily hospital “COVID admissions per 1000 people tested” in preference to simple the COVID admissions—see Figs. E4 and E5.
Fig. E4. New COVID hospital admissions per 1,000 tests since 22nd April 2020 when testing data was first available.
Fig. E5 New COVID hospital admissions per 1,000 tests since 1st July 2020.
Plotting the current infection and hospital admissions data per 1,000 tests performed shows not only that current infection levels are negligible compared to the spring season, but also that the hyped talk of a “second wave” by the Prime Minister and press is completely unjustified.
The notion that “cases”, as determined by positive PCR test, are increasing exponentially (or indeed rapidly) is also unjustified. Emotive language as used by the Prime Minister, for example “we are in a second wave” damage his credibility and the credibility of his policies, which to be honest is tenuous at best.
On 3rd October the news media, namely the BBC, reported that “the UK had hit 10,000 new cases for the first time since mass testing began”. This headline is highly misleading because the news media, and often Government, quote their “case” statistics by date of reporting and not by the data that the specimen was taken. Due to technical issues there were delays reporting the data meaning that ‘case numbers’ will be inflated while the backlog is cleared. The simple solution is to report tests by specimen date, and presenting the 3rd October data by specimen date (Fig. E6) shows that the alarm caused by the media is unjustified because the 10,000 plus “cases” were distributed over as many as 23 previous days.
Fig. E6 New COVID positive tests presented by specimen date.
Basic statistical errors, such as those described in this Appendix which have been consistently made in the media coverage and Government briefings can be at best described as grossly incompetent. They also do not give any confidence that the Government is basing decisions on credible scientific data and analysis undermining the Government’s policies and overall credibility.
APPENDIX F. HERD IMMUNITY & SWEDEN
The often quoted theoretical herd immunity threshold of 60%, calculated naively as 1 - 1/R0, is applicable to random vaccination or infection. Real disease epidemics don’t work in this way because they are naturally selective towards the vulnerable group, which in this case is well-defined and relatively small. Several research groups have estimated the herd immunity threshold significantly lower than 60%, and 20% seems to be the common estimate, for example [1,2] although others still argue it may be slightly higher .
Even if this is not the case there are many papers now which demonstrate immunological cross-reactivity with other coronaviruses, for example [4-7], which means that many more people have some form of immunity than initially believed. The proportion of individuals who haven’t been exposed to SARS-CoV2 who have such T-cell immunity could be, according to a number of studies, as high as 60% [4,6]. This shouldn’t come as a great surprise to anyone who was in the field of epidemiology in 2010 as a similar large-scale “immunological head start” was demonstrated for Swine Flu .
In terms of the Swedish epidemic, despite the UK lockdown the net mortality burden for Sweden was not significantly different from the UK: approximate overall mortality rate of 0.06% of the population. Fig. F1 rescales the epidemic curves for Sweden and England and Wales such that the epidemic peaks for both countries are at the value 100. The epidemic curves are almost identically shaped despite the significant differences between their national strategies.
Fig F1. Covid deaths for Sweden and England and Wales renormalised to 0 to 100 scale for both countries.
If lockdowns were really an effective virus suppression strategy, one would expect that the shapes of these curves would not be the same. Recent work and data point to lockdowns being of highly questionable efficacy for virus control [8-10], though undoubtedly they cause harm due to disruption of the health services and societal disruption.
The detractors of the Swedish approach often compare Sweden with the other Nordic nations who, on the face of it, achieved lower rates of COVID mortality with more stringent policies. One cannot view a disease epidemic in isolation and the most likely explanation for this observation comes from each nations previous flu season severity.
Sweden and the UK both had relatively mild 2019/20 flu seasons which left additional vulnerable individuals when SARS-CoV2 began to circulate early in 2020; the so-called “dry tinder effect” . In contrast the other Nordic nations had comparatively normal flu seasons. There is a negative correlation between flu season intensity of the previous two flu season and COVID-19 death rates in 32 European nations , meaning that those countries with lower mortality in their previous flu season have suffered a demonstrably greater mortality impact from COVID-19 ( r squared of 0.396, significant at the 1% level on a 1-tailed test).
Fig F2. Relationship between CVOID-19 mortality impact and mortality impact for previous influenza seasons for European nations. Source: .
- Aguas R, et al. Herd immunity thresholds for SARS-CoV-2 estimated from unfolding epidemics. https://doi.org/10.1101/2020.07.23.20160762
- Gomes M, et al. Individual variation in susceptibility or exposure to SARS-CoV-2 lowers the herd immunity threshold. https://doi.org/10.1101/2020.04.27.20081893
- Fontanet A & Cauchemez S. COVID-19 herd immunity: where are we? Nature Reviews: Immunology.
- Grifoni D et al. Targets of T Cell Responses to SARS-CoV-2 Coronavirus in Humans with COVID-19 Disease and Unexposed Individuals. Cell, Volume 181.
- Doshi P. Covid-19: Do many people have pre-existing immunity? BMJ 370 https://doi.org/10.1136/bmj.m3563
- Hicklin T. Immune cells for common cold may recognize SARS-CoV-2. National Institutes of Health: Research Matters. https://www.nih.gov/news-events/nih-research-matters/immune-cells-common-cold-may-recognize-sars-cov-2
- Sekine T et al. Robust T Cell Immunity in Convalescent Individuals with Asymptomatic or Mild COVID-19. Cell, Volume 183.
- Colombo M et al. Trajectory of COVID-19 epidemic in Europe. https://doi.org/10.1101/2020.09.26.20202267
- Atkeson A et al. Four Stylized Facts About COVID-19. National Bureau of Economic Research, USA. https://www.nber.org/papers/w27719.pdf
- Wood S. Did COVID-19 infections decline before UK lockdown? Available at https://arxiv.org/abs/2005.02090v5
- Klein D et al. 16 Possible Factors for Sweden’s High Covid Death Rate among the Nordics. Social Science Research Network. Available at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3674138#
- Hope C. COVID-19 Death Rate is Higher in European Countries with a Low Flu Season Intensity Since 2018. Cambridge Judge Business School Working Paper 03/20