Should We Have Relied More on Historical Comparisons and Less on Epidemiological Models?

During the pandemic, the British government has relied heavily on epidemiological models when deciding what course of action to take (e.g., whether to tighten or loosen restrictions). The advice it has received in this regard comes from the Scientific Pandemic Influenza Group on Modelling (SPI-M), a body comprising 82 scientists from institutions across the U.K. 

Most influential (and infamous) have been the epidemiological models developed by Neil Ferguson’s team at Imperial College London. Indeed, the Government initially appeared to be following a focussed protection strategy; it was only after the publication of an alarming report by the Imperial College team that lockdown became the official policy. (Ferguson and colleagues’ report has since been described as the “catalyst for policy reversal”.)

As late as March 5th, Chris Whitty told MPs on the Health and Social Care Committee: “We will get 50% of all the cases over a three-week period and 95% of the cases over a nine-week period.” He explained: “What we’re very keen to do is not intervene until the point we absolutely have to, so as to minimise economic and social disruption.” And he added that “one of the best things we can do” is to “isolate older people from the virus”.

Dominic Cummings has since confirmed that the Government did abandon its original plan at the last minute. He claims, “No10 was made aware by various people that the official plan wd lead to catastrophe.” 

However, the epidemiological models that served as the basis for lockdown – both here and elsewhere – have come under substantial criticism. They made highly untenable assumptions, such as that seasonality and voluntary behaviour change do not affect transmission. This, in turn, led to disastrous forecasting errors. For example, Neil Ferguson’s team predicted there would be 85,000 deaths in Sweden; to date, there have been fewer than 15,000 (and that figure’s probably an overestimate).

In a new paper published in History and Philosophy of the Life Sciences, George Heriot and Euzebiusz Jamrozik argue that we should have relied more on historical comparisons, and less on epidemiological models. 

They point out that “twenty-first century human communities may bear greater resemblance to communities in the eighteenth and nineteenth centuries than to an abstracted representation within an epidemiological model”.

The authors note that the transmissibility and lethality of COVID-19 are “are well within the range described by respiratory viral pandemics of the last few centuries”, whereas the Spanish Flu of 1918 is “the clear outlier”. They suggest that the 1889 flu pandemic (sometimes termed the “Russian Flu”) offers a particularly close historical analogue to COVID-19.

According to the authors, “The historical record may provide a richer and more useful understanding of the range of medium- and long-term consequences… than even the most complex mathematical model.” And they go on to say: “Every established respiratory pandemic of the last 130 years has caused seasonal waves of infection and has culminated in viral endemicity.”

Heriot and Jamrozik’s article provides a much-needed antidote to the Government’s modelling malady and is worth reading in full.