Attacks against Covid modellers undermine future pandemic response, say Sage scientists

Exclusive: Members of the Scientific Pandemic Influenza Group on Modelling hit back at critics and warn against dangers of eroding public trust in science

‘The newspapers tend to emphasise the worst-case scenario, so that’s the stuff that gets reported,’ said one scientist (AFP/Getty)

Repeated attacks against the government’s Covid modellers threatens to erode public trust and undermine future pandemic responses, senior scientific advisers have said.

Several members of the Scientific Pandemic Influenza Group on Modelling (SPI-M) – a sub-committee of Sage – told The Independent they were concerned by the “knock-on effect” of the criticism they have received in recent weeks for their modelling of the Omicron wave.

Fellow scientists and MPs have accused the advisers of bouncing the UK into repeated lockdowns and creating a “climate of manipulated fear”. SPI-M members fear such attacks will weaken trust in the modelling and hinder decision-making in the face of dangerous new Covid variants or future pandemics.

Professor John Edmunds, a SPI-M member, said there has been both “wilful misinterpretation” and a “genuine misunderstanding” of the group’s recent modelling, which did not predict what would happen over winter, but instead provided a variety of scenarios for ministers to review, ranging from the best to the worst.

This explains why one of the scenarios showed that 6,000 people could die a day, with tens of thousands of daily hospitalisations, the experts said. “The newspapers tend to emphasise the worst-case scenario, so that’s the stuff that gets reported,” said Prof Edmunds.

Others scenarios showed up to 400 deaths a day and just under 3,000 daily hospitalisations if restrictions weren’t imposed – a projection which has come close to materialising over winter.

In light of the modelling, and despite calls from Sage for the reintroduction of “more stringent measures” in mid-December, the government opted to stick with its plan B restrictions – a decision that supporters feel has since been vindicated, prompting many to accuse the SPI-M experts of getting it wrong.

“The main misunderstanding is that the models are not predictions,” said Professor Graham, chair of SPI-M. “The models are scenarios to help the decision-makers understand the implication of different policy choices.

“The second major problem is that because the focus is on the worst case, it looks as though we only model that, which we don’t. We are illustrating many possibilities for government.”

Prof Graham said he was concerned by the recent attacks, which could lead to an “erosion of trust” in the modelling. “You already have that in terms of vaccinations,” he added.

“I think it’s a severe problem,” said Prof Edmunds. “It’s the same with [modelling] climate change. Those guys face exactly the same issue. And you have the kind of bleating right-wing press trying to undermine it, for political reasons. It has a knock-on effect. So yes, it’s a big worry.”

Earlier in the week, Tory MP Bob Seely called the reliance on the modelling a “national scandal” and said “that never before has so much harm been done to so many by so few based on so little, questionable, potentially flawed data”.

Steve Baker, the deputy chair of the right-wing Covid Recovery Group, said projections produced by Professor Neil Ferguson, another member of SPI-M, had “disgracefully” bounced the prime minister into a lockdown at the start of the pandemic.

One SPI-M, who asked to remain anonymous out of fear, said critics expected the models to be “omniscient”.

“We are generally asked for modelling results at times of crisis, when things are going badly (eg, cases rising rapidly) and there is great uncertainty (eg, a new variant) – so it’s not surprising that our projections are imperfect,” they said.

“The analogy I’ve used before is that it’s like asking political commentators to predict the results of the next election based on current trends at a time of political crisis. We wouldn’t expect that to be accurate due to all the events that can happen in the future.”

The expert said they had been “hurt by some of the unwarranted comments about the modellers” and was reluctant to get publicly drawn into the discussion.

“Throughout this pandemic, I’ve managed to keep a relatively low profile and hence I haven’t suffered the mountains of abuse that others have received,” they added.

Others, like Prof Ferguson, haven’t been so fortunate, having been dubbed “Professor Lockdown” on account of his modelling.

At the start of the pandemic, his team from Imperial College London modelled a range of scenarios for the government of what could happen if restrictions weren’t imposed, including one that said up to 500,000 people could die from Covid-19.

“We didn’t see that because we locked down,” said Professor Mike Tildesley, a SPI-M member. “I think the media generally understands that, as do the public. It’s the people who don’t like controls or don’t like models that will sort of point this out to deliberately have a pop.”

These attacks nonetheless carry of a risk of polluting the public’s perception of the science and experts advising the government, he added.

All of those who spoke with The Independent also said that the wide range of uncertainty surrounding the modelled scenarios for Omicron had been impacted by a lack of data.

The variant took hold in early December, at which point little was known about severity and its ability to evade the body’s immune response. It took weeks for this data to crystallise.

“When Omicron first arrived, we knew very little about it, so the range of scenarios was going to be very wide,” said Prof Edmunds. “One of the important things we didn’t know was the severity. It was only shortly before Christmas that we had decent data that Omicron was actually less severe.

“The scientific process seeks works one step at a day, getting things wrong, learning and adapting and changing, and so on. It’s quite frankly ludicrous to assume that we knew everything about Omicron at the beginning. Our early scenarios reflected all of that.”