THURSDAY, 21 JANUARY 2021The COVID-19 pandemic has revealed diversity in the world's attitude towards healthcare and public welfare. One particularly striking aspect is how social values beyond scientific evidence shaped recommendations in Evidence Based Medicine. Here, I explore how COVID-19 response teams operated according to the principle of precaution. Precaution, as a guiding principle of theory and policy, encourages us to choose the intervention, evidence, and prognosis that is most pessimistic and cautious with respect to the future. Perhaps one key lesson to be learned from the post-COVID-19 ‘new normal’ is that the background values which shape our scientific models have fundamental consequences for our interpretation of scientific knowledge. Here, I explore the analysis and critique of these values by discussing the impact of being precautious on the epidemiological models produced by the UK’s leading COVID-19 response teams.
On 26th March 2020, Imperial College London’s COVID-19 response team published their report, The Global Impact of COVID-19 and Strategies for Mitigation and Suppression. Based on a series of models, they predicted that SARS-CoV-2 would cause 40 million deaths globally if left unchecked. The report came at a critical time in public health debates — the UK was set to follow a policy whereby the country would remain entirely open in order to generate herd immunity, where spread of the virus is prevented by enough of the population gaining immunity by contracting the disease and recovering. Imperial’s bleak prognosis catalysed the Prime Minister’s revision of the herd immunity stance whilst the shocking numbers in the report drove many to take the threat of the virus seriously.
Which values should we value?
At times when scientists and policymakers are expected to make decisions with potentially harmful consequences for the public, it is intuitive that they should consider the most possible detrimental outcomes of each option in their decision making process. This is the rationale that sits behind the precautionary principle.
Since the 1960s, philosophers of science have worked hard to devise an account for how scientists do and should use values such as the precautionary principle in their work. The clearest justification for the use of such values was first articulated by Richard Rudner in 1953: Argument from Inductive Risk (AIR). The argument states that: since no amount of evidence can guarantee that one hypothesis is correct over and above its closest rival, hypothesis choice is always underdetermined. Hence, scientists must use value judgements such as the social, ethical and political implications to navigate the choice by assessing the implications of choosing an incorrect hypothesis or disposing of a correct one.
According to Rudner’s argument, the choice between a herd immunity response or a lockdown response requires the inclusion of judgements beyond the confines of evidence and data. This is where the precautionary principle was exercised.
The poor quality of data in the early stages of the pandemic seemed to further justify a method that proceeded with maximum caution when designing effective public health responses. Furthermore, Rudner’s argument from inductive risk highlights that the less reliable the evidence is, the more essential extra-scientific values become. From looking at practitioners’ opinions on the quality of evidence during the pandemic, it was clear that values akin to the precautionary principle were necessary for good reasoning. Assistant Professor of history and philosophy of science at the University of Pittsburgh, Jonathan Fuller, explained that the focus on model-based strategies in COVID-19 epidemiology, as opposed to other evidence-based approaches such as population sampling, was the result of the dire quality of evidence collected at the beginning of the pandemic. The epidemiologist Professor John Ioannidis at Stanford University even labelled the scientific response to the virus an ‘evidence fiasco’. He also highlighted how the disparate methods of data collection across the globe have made the information on the rate of spread and the virus’s deadliness ‘utterly unreliable’.
A deeper look at Imperial’s report
Precautionary judgements were entrenched in the report from Imperial College, highlighting the centrality of this principle in guiding COVID-19 science and policy decisions. The shocking prediction that there would be seven billion infections and 40 million deaths globally were reached through worst-case-scenario modelling, whereby the R-number (the number of healthy people a single sufferer will infect on average) was set to 2.4–3.3, which was the upper band of postulated values estimated at the time. In mid-May, for example, Public Health England and the University of Cambridge calculated that the R-number in the UK was between 0.5–0.9. The fact that Imperial opted for such a high R-number highlights the depth to which being precautious and postulating on the worst-case scenario was rooted in this report.
The report concluded that, ‘given these results, the only approaches that can avert health system failure in the coming months are likely to be the intensive social distancing measures currently being implemented in many of the most affected countries’. It is no surprise that this would be the recommendation of a report that compared how a highly infectious virus would spread freely through a population undertaking no mitigation strategies, against simulations where extensive preventive measures were put in place. In other words, the recommendation of the model is a direct result of maximally dangerous premises and investigating only one avenue of mitigation, namely restrictions on movement. Thus, as one looks deeper into the Imperial report, the impact of the precautionary principle on the modelling style and interpretation of results was wide reaching, emphasising the need to scrutinise the justification for this value further.
A viable alternative?
The question will always remain on how different our COVID-19 journey could have been if a precautionary approach was exchanged for something else. One viable contender was offered by Professor Alex Broadbent, who discussed a ‘rational cost-benefit analysis’ approach. His suggestion was that our response to the current pandemic should be the result of careful compromise between a myriad of different problems, such as the supreme danger the virus poses to the sick and elderly, the damage to children’s educations from school closures, and the unprecedented impact on the economy lockdown has and will have.
He thinks that the most rational way to construct public health policy is to consider all possible consequences of COVID-19 strategy and trade-offs between them. It is important to note that Imperial’s report purposely excluded socio-economic considerations of the social distancing measures they adamantly endorse: ‘we do not quantify the wider societal and economic impact of such intensive suppression approaches’. As we enter a post-lockdown period of trying to restore our businesses, schools, and industries, only time will tell how successful the precautionary principle was at mediating the extremely complex landscape of irreducible and irreconcilable public health, social, and economic issues.
The celebrated and influential 20th century sociologist, William Edward Du Bois, defended a science free of values by emphasising that in a democratic society, scientists do not have the sufficient expertise to make judgements concerning what the outcomes of the knowledge they produced will be. That is the job for democratically elected politicians. Whilst precaution has been the watchword of risk assessment during the COVID-19 pandemic, perhaps allowing philosophers of science to scrutinise the values that we use to conduct epidemiological research in times of crisis would not only expose the values that shape our research, but also allow us to spend time conceptualising what the best values to use should be.
The lesson to be learned here is that the scheme of values that we prioritise in society has immediate and fundamental consequences for the science that we make and spread. Perhaps the outcome of this pandemic will inspire more careful evaluations of the judgements that sit behind our models and predictions so that we can justifiably say that we have learnt for the future to better evaluate our methods of judgement in situations of uncertainty.
Charlotte Zemmel is an MSc student in History and Philosophy of Science at Newnham College. Artwork by Natalie Saideman.