SAGE has acknowledged that its terrifying Omicron hospital admission projections were off the mark — and it is now warning of a ‘long tail’ of infections as the UK comes out of the fourth wave.
Modeling by the Government’s scientific advisor group showed that there might be as many as 10,000 Covid hospitalizations per day by mid-January, even though it was only December 23.
The reality is that daily admissions are now just above 2,000 in the UK. They have already fallen in England which gives the Prime Minister confidence to lift Plan B restrictions next Wednesday.
Last Thursday’s SAGE meeting was presided over by Sir Chris Whitty (chair) and Sir Patrick Vallance (chair). The group acknowledged that they had not seen the predicted surge in hospitalisations.
This was due to the combination of the country’s high-immunised population, waning vaccine protection faster than anticipated and vulnerable individuals taking precautionary measures.
Minutes were also made public by members today. They also admitted that the official projections of Omicron case numbers exceeding one million per day had not materialized.
SAGE stated that behavioral changes made by younger people may have led to lower models of infection over the past week.
International studies, mobility data, and other evidence over the past two years have all shown people changing how they behave and the relationships they make in response to rising cases.
MailOnline spoke with Professor Paul Hunter of University of East Anglia. He said that the main reason the models seem ‘overly pessimistic is because they fail to account for behavioural change.
In December, projections were not adjusted for Omicron’s decreased severity. This despite South African real-world evidence clearly indicating that the variant is milder.
The SPI-M modeling group, which feeds into SAGE, also published a separate document today. It claimed that Omicron’s peak was not due to natural immunity but because people were more cautious.
The warning stated that there were still many people susceptible to the virus despite having stayed away from it for eight weeks.
The Government’s scientific advisory committee predicted that by January mid-January there would be upwards of 10,000 Covid hospitalisations daily. London School of Hygiene and Tropical Medicine presented the doomsday scenario in a model
Warwick University released its December 30, infection, hospitalisation, and death projections. The group predicted that there would be up to 1.4million infections per day, nearly 3,000 hospitalizations, and a total of 1.3million deaths. This was based upon various levels of restriction and Omicron being 50% less severe than Delta.
LSHTM’s estimations are based on Omicron’s immunity escape potential and booster effectiveness
The SAGE minutes acknowledged the collapsed Omicron rates and stated that the increase in hospitalisations was not anticipated after the increase in elderly cases.
“This could be because of higher levels of vaccination against hospitalisation or slower waning vaccine protection. It may also reflect the effect of precautionary behaviors among the most vulnerable people and their families.
“Analysis by Bristol has shown that the intention to change behaviours over December 2021 led to lower models of infection over recent weeks, compared to risk mitigation).
SAGE also stated there was still uncertainty over hospital admissions over coming weeks. The case rate is high at around 90,000 Britons testing positive daily.
MailOnline spoke to Robert Dingwall, an ex-Government Covid adviser and sociologist. He said advisers should be careful not to use a simplistic excuse of “it’s all about behaviour and we can’t anticipate that”.
The fact that he accepted the challenge made modeling more difficult when you add in complex elements such as how frequently someone is out of their home, travels by public transport and interacts with other people.
He added, “But the…” [SAGE] modellers don’t seem to have had a serious discussion with people in other fields. There are sociologists and economists who model, so there is a lot of people that understand it.
“People are realizing that these models work.” [done by] the very narrow respiratory disease community… so their models haven’t really incorporated things others people know about.’
SPI-M published a consensus statement last Wednesday claiming that Omicron infection had not reached its peak naturally.
‘As of yet there are no signs that a purely immuno-driven peak has occurred in either the UK case data, or the CIS.
“Peaks of infection have frequently been asymmetrical in nature during COVID-19 with many of the more severe infections occurring at the population decline stage. Therefore, a long tail could still be needed to manage even after the peak.
SPI-M, however, expressed optimism regarding what would happen at hospitals during the weeks ahead. It noted that the rate of fatalities in recent waves has been much lower than the previous wave.
The statement adds: ‘This is highly likely a result of the combination of omicron’s decreased intrinsic severity and high vaccine effectiveness.’
MailOnline received a statement from Professor Hunter stating that SAGE “should” have been able to account for behavioural changes as there’s a lot of data on the effects of rising infections.
However, Professor Hunter stated that he believed natural immunity wasn’t being correctly accounted for in the models.
He mentioned Professor Lockdown Neil Ferguson’s projection of 250,000 Covid death without a lockdown by 2020, and said that was “not too unrealistic”.
Dr. Hunter says that the projections are getting worse because of our inability to properly account for natural immunity.
He stated that models worked well during the initial pandemic. That stick is a good idea. [aimed at Professor Ferguson’s 250,000 deaths model]Was unfounded, and a bit too extravagant.
“But, modeling endemic infectious diseases is more difficult than modelling epidemics. [SAGE is]It is not easy to model immunity.
“They must look at their assumptions regarding immunity, and consult people with a better track record of modeling immunity.