By Gordon Woo for
COVID-19 Actuaries Response Group – Learn. Share. Educate. Influence. COVID-ARG.COM
Learning from pandemic experience is a continuous adaptive process, rather like the virus itself. Given that there has not been a severe pandemic in living memory, public health policy can be ‘exploratory’, with the need for course corrections as circumstances require. Looking back at the challenges of 2020, we can draw constructive lessons on improving public health policy in 2021. These can be split into three areas: following the science; following the rules; and following the virus.
Following the science
Not long after the 2009 infIuenza pandemic, I served on the Blackett Committee on extreme risks under the chairmanship of Sir John Beddington, the government Chief Scientific Advisor. The Blackett Committee advised on the government risk matrix, which classifies risks according to impact severity and likelihood. Pandemic risk stood clearly at the top of this risk matrix.
The current Chief Scientific Advisor, Sir Patrick Vallance, relies on the SAGE Committee, and related sub-groups, to provide guidance on scientific issues concerning the pandemic. There is considerable uncertainty associated with any hazard, and such uncertainty leads to differences in expert judgement. This is reflected in the establishment of an alternative SAGE Committee, under the leadership of Sir David King, a former Chief Scientific Advisor to the Government.
Managing a crisis involving a new external hazard is an exercise in decision-making under substantial uncertainty. There is rarely enough data to provide full scientific support for the decisions. This is clearly demonstrated in the formal review of the COVIDSIM model developed at Imperial College, under the leadership of Neil Ferguson. This elaborate epidemiological model has over 900 input parameters, and the uncertainty in the code was found by the reviewers to be substantial.
Inevitably, at any moment, there is a spectrum of decisions that could reasonably be made. This can make it harder to initiate necessary policy changes, which are open to criticism as U-turns. Furthermore, any assertion to be following the science belies the diversity in scientific expert judgement.
Rather than group meetings amongst scientists, which are all the more difficult and fraught for being held online, there are formal methods for the elicitation of expert judgement in a risk context, which take account of the different capabilities individual scientists have in characterising uncertainty and the state of knowledge. The value of a formal elicitation procedure was highlighted in the run-up to the November 2020 lockdown. A number of alternative epidemiological projections from the modelling subgroup SPI-M were presented to the public, without interrogating the underlying assumptions.
Public health decisions in an ongoing crisis often need to be made before enough scientific evidence is accrued to achieve scientific consensus. In Asia, the wearing of face masks has been quite common since the 1950s, as a matter of public etiquette, and increased after SARS in 2003. Both UK and USA have been slow in recommending the wearing of face masks, even though this is a simple inexpensive non-pharmaceutical intervention. On 4 March 2020, England’s Chief Medical Officer, Chris Whitty, advised that wearing a face mask had almost no effect on reducing the risk of contracting the illness.
This advice seemed questionable, and has since been reversed. Inevitably, there is a time lag for rigorous scientific studies to be undertaken. On 19 January 2021, one year after the first COVID-19 case was confirmed in USA, a study was published showing that a 10% increase in mask-wearing was associated with increased transmission control. This demonstrates that the precautionary principle may need to be reversed depending on the context – rather than ‘prove this measure works’, we would have ‘we assume this measure may work, and it is highly unlikely to be harmful’.
Governments should be guided by science, but may not necessarily be following the science in the fullest sense of the term. It was former Prime Minister, Tony Blair, who first suggested deferring the date of a second vaccine jab so as to increase rapidly the proportion of the elderly population who would have the protection of one jab given limited vaccine resources. This is a reasonable public health policy, which has been implemented amidst some scientific misgivings, notably from Dr. Anthony Fauci, the leading American expert. He was adamant that the USA would follow Pfizer and BioNtech’s guidance to administer the second dose of its vaccine three weeks after the first. Ideally, there should be formal pharmaceutical trials of the consequences of delaying a second jab, but such trials would take many months. In 2021, public health authorities should aim to make the best decisions for the population, which are both science-based and risk-informed, given the need for speedy decisions in the absence of perfect data.
Following the rules
The imperative to avoid exhausting ICU capacity and so maximise the survival chances of those with a severe case of COVID-19 poses new challenges for pandemic crisis management, which have never arisen before. At the time of the 1918-1919 pandemic, treatment was largely symptomatic, aiming to reduce fever or pain. There was no urgency to suppress the number of infections so as to relieve pressure on ICU capacity. Even later for the 1957 Asian flu pandemic, governments did not consider suppressing R by introducing strict isolation measures, because ICU were not yet established, and ventilator technology was rudimentary. For the 1968 Hong Kong flu pandemic, there were few school closures, and businesses continued mostly as usual. For the 2009 Mexican flu pandemic, the case fatality rate was lower than for seasonal flu, and there was little need for social distancing other than some local school closures.
In writing his comprehensive report on the 1918-1919 pandemic, E.O. Johnson concluded that in urban areas there would be little to be gained from mass closures unless the community would accept great restrictions of ordinary activities. A century later, the threat of the NHS being overwhelmed and ICU capacity exhausted provided a compelling and persuasive argument, leading to widespread public acceptance of the severe restrictions associated with regional and national lockdowns. In contrast with previous UK pandemic experience, the availability of ICU treatment for severely ill patients poses unprecedented dilemmas for decision makers. Any form of triage is morally invidious and may lead subsequently to legal action against clinicians.
Because the Kent variant of COVID-19 is more transmissible than the mutated virus of the second wave of the 1918-1919 influenza pandemic, there must be high compliance with the rules in 2021, even as vaccines are rolled out, and even as pressure is eased on the NHS. But people with limited financial means deserve practical help from the government in following the rules.
In countries where there is high compliance with quarantine, special consideration has been given to financial assistance as well as medical and food provision. Financial support of £500 for quarantine was offered in UK only at the end of September 2020, after a King’s College London survey revealed that just 11% were prepared to abide by the quarantine rules.
Amongst those who still find it extremely difficult to comply are those financially struggling who do not qualify for the £500. In Yorkshire and Humber, 60% of applications have been rejected, and there are high rates of rejection in other council districts. A Department of Health and Social Care policy document has costed a payment of £500 to everyone at £2 billion a month. This is expensive for the Treasury, but for fear of losing their income, only 17% of people with COVID-19 symptoms have been coming forward to get a test. Furthermore, only one in four are fully self-isolating for ten days, and 15% carry on going to work.
Susan Michie, the SAGE behavioural scientist, has recommended extension of the financial assistance for those unable to work from home. She is also an advocate of the proposal that confirmed cases living in multi-occupancy households should be able to isolate in quarantine centres or hotels. Sir David King, the leader of independent SAGE, has also endorsed this measure, which has been adopted in New York to break chains of household transmission.
Removing infected people from contact with the uninfected is all the more important with the discovery at the end of 2020 of a new worrying Brazilian variant of COVID-19 which might potentially evade human immune response, and render those who already have had COVID-19 susceptible to reinfection. Because of the possible implications for vaccine efficacy, mandatory prior testing for arrivals into UK has been introduced in mid-January 2021. As in some other countries, visitors will be required to quarantine on arrival in hotels, rather than at home. This more stringent requirement is warranted by poor traveller compliance with home quarantine. This is estimated to be only 29%; but then Public Health England randomly samples only around 1,000 eligible arrivals per day to be checked by a phone call.
Awareness that many UK arrivals flout the quarantine rules discourages general public compliance with quarantine. Not following the rules is itself a contagious behaviour trait. In 2020, there were some highly publicised instances of rule-breaking by public figures, who may have acted as superspreaders of this rule-breaking ‘behavioural virus’.
Following the virus
An intrinsic characteristic of COVID-19 which has made it so much harder to contain than SARS in 2003 is the large amount of asymptomatic transmission. This has made diagnostic testing for low levels of infection such a crucial tool for controlling the spread of COVID-19. In the ONS Coronavirus Survey of private households in England, a substantial proportion of infections (45% to 68%) have been found to be in individuals not reporting symptoms around their positive test.
At the end of September, the estimated percentage of individuals testing positive was more than six times higher in young adults between 17 and 24 than in those aged 70 and over. The population percentage testing positive had shot up from 0.06% at the end of June 2020 to above 1% at the end of October. It was well known, and overtly publicised, that the free mixing of college students at the beginning of the academic year would significantly elevate the low infection levels of July and August. But without the availability of rapid inexpensive testing, containment of the silent spread of COVID-19 within college halls was a practical impossibility.
The stark consequence is illustrated in the figure below. The number of daily positive test cases at the end of September for Exeter University students staying in halls was 140. Exeter University has about 5% of the total number of UK students. At the end of September, the daily number of new UK cases was around 5,000, and still amenable to contact tracing. So half of the number of UK cases at the start of the second wave may have been associated with returning students. In 2021, students will need to be tested before they travel to campus.
The public health slogan in early January 2021 that every person should act as if they had COVID-19 would have been appropriate guidance for students in 2020, and would be prudent throughout 2021. Given the possibility of re-infection by a new evasive variant, this applies also to all those who know they have already had the coronavirus.
With the prevalence of COVID-19 expected to decline progressively with the vaccine rollout, the number of people presenting with COVID-19 symptoms should decline correspondingly. There then needs to be a sustained campaign of population testing to identify those who are infected, but are asymptomatic. Some of those who have been vaccinated may still harbour the virus, and be able to transmit it to others. The progressive relaxation of social distancing restrictions in 2021 will depend on public health authorities following the virus, not just the disease.
9 February 2021
 Edeling W. et al. (2020) Model uncertainty and decision making: predicting the impact of COVID-19 using the COVIDSIM epidemiological code. Royal Society RAMP initiative report.
 Rader B. et al. (2020) Mask-wearing and control of SARS-CoV-2 transmission in the USA: a cross-sectional study. Lancet Digit Health.
 Johnson E.O. (1927) Epidemic influenza: a survey. American Medical Association.
 Kupferschmidt K. (2020) New coronavirus variants could cause more reinfections, requiring updated vaccines. Science, January 15.
 Pouwels K.B. (2020) Community prevalence of SARS-CoV-2 in England from April to November 2020. Lancet Public Health 6, e30-38.