SAGE expert says wildly wrong Omicron death predictions failed to account for behaviour change

16

View
comments

Covid models predicting that deaths from Omicron could peak at 6,000 a day were wildly wrong because they failed to account for voluntary behavioural change, experts who advised the Government have admitted. 

Professor Graham Medley, chair of the Scientific Pandemic Insights Group on Modelling (Spi-M), which has advised the Government on its pandemic response, answered tough questions from MPs yesterday.

He said models of potential infections, hospitalisations and deaths from Covid were scenarios and not predictions.

But Professor Medley also admitted that not including in the models people’s changes in behaviour, such as attending fewer social occasions which could stem the spread of Covid, was ‘maybe an error’.

Professor Medley told the Commons Science and Technology Committee that in mid-December – when ministers were under pressure to impose another national lockdown in response to fears about the new Omicron variant – experts on SAGE presented the Government with a range of deaths scenarios. 

In their best-case scenario, they predicted deaths peaking at 600 a day but said they could reach 6,000 between mid-January and mid-March. Instead, they peaked at 306 on January 1. 

‘The worst thing for me would be for the Government to say, “Why didn’t you tell us it would be that bad?” So, inevitably, we’re always going to have a worst case above reality,’ Professor Medley said in The Times. 

The models were wrong in part because they did not factor in behavioural changes which people made themselves in response to warnings about Omicron, without being legally compelled to do so. 

It meant that socialising and contact between households plummeted before Christmas, even though the Government opted not to impose another lockdown.  

Professor Medley also admitted that it was nearly impossible to predict behaviour. 

‘The epidemic is dynamic. People’s responses to the situation in March 2020 were very different to those in November 2020 and very different again in January 2021,’ he said.  

Another expert who gave evidence to the Commons committee said he feared the incorrect models could prompt Britons to ignore warnings in future.  

Cambridge University epidemiologist Dr Raghib Ali said: ‘My concern is next time there is a new variant doctors will go on TV and say “this is bad” but people will say, “you got a lot wrong last time, we’re not going to listen”, and that’s very dangerous.’

The claim that the Government practised ‘number theatre’ during the pandemic was backed at the Committee hearing by the head of the statistics watchdog.

Health Secretary Sajid Javid was last year slapped down by the Office for Statistics Regulation after saying in mid-December that there were 200,000 new Omicron infections a day.

Ed Humpherson, head of the watchdog called for better ‘transparency’ while giving evidence.

Mr Humpherson said: ‘A couple of times during the pandemic, we’ve seen examples of where we think the models have been put out in a way which doesn’t give enough information to enable people to make sense of them.’

Referencing critical comments by statistician Professor Sir David Spiegelhalter on numbers given by the Government, the head of the statistics regulator added: ‘Most recently on an estimate that Omicron would produce 200,000 infections a day, David Spiegelhalter memorably described it as number theatre – it’s just putting a number out.

‘As a result of speaking to the UK Health Security Agency, they then released the underlying logic of their model.’

It took officials three days to justify the 200,000 infections figure in December, after the Office for Statistics Regulation said it ’caused confusion’.

The Science and Technology Committee is scrutinising statistics on Covid following controversy over how they are communicated.  

Behaviour changes were included in models created in Denmark, the committee heard.

Dr Raghib Ali, an epidemiologist and consultant in acute medicine at Oxford University Hospitals, who correctly predicted that the omicron wave would be mild, told MPs: ‘People’s behaviour does change.’

On people’s willingness to be more careful in response to rising numbers of cases, he said: ‘What has an impact is knowing that your friend, relative, colleague has got Covid – of course, during Omicron, that was very common – every one of us knows someone that got Covid.’

Doctors had seen early signs the Omicron variant was not as severe, Dr Ali said, suggesting there should have been more doctors on the Sage panel of scientists advising the Government.

Mr Humpherson also accepted criticism of NHS statistics showing how many people have died within 28 days of a positive Covid test.

He said the snapshot figure had ‘limitations’ in failing to separate out people dying from Covid from those who happened to die while they had Covid. 

It comes after Britain’s daily Covid cases rose for the first time in a month yesterday in a sign the outbreak may be growing again, while hospitalisations also ticked upwards. 

Government dashboard data shows another 44,017 infections were detected over the last 24 hours, up 11 per cent on the tally last Wednesday.

It brings an end to more than four weeks of tumbling daily cases, with about 33,700 cases now being recorded every day on average.  

Latest hospital data showed 1,040 people were admitted to hospital with the virus on February 26, up seven per cent on the 970 from the previous week.  

But the seven-day average number of daily admissions is still falling, meaning today’s rise could be a blip. Daily Covid deaths, however, have continued to fall, with the 74 victims announced today down 54 per cent in a week. 

The rise in infections coincides with a more infectious version of Omicron, scientifically named BA.2, becoming dominant in England last week. 

Experts warned it may cause some fluctuations in case rates, but called for calm saying there is no evidence that it is more severe than the original strain.

Plummeting testing numbers, which have dropped by a tenth in a week, also make it more difficult to track trends in infections.

Last week, Boris Johnson ditched all of England’s remaining restrictions, with requirements to wear face masks on public transport and isolate when infected coming to an end.

Mass testing is also set to be dumped from April 1, with the country instead set to rely on the Office for National Statistics’ Covid infection survey. 

But experts today warned this could take two weeks longer to pick up new potentially dangerous variants.