select search filters
briefings
roundups & rapid reactions
before the headlines
Fiona fox's blog
The latest batch of SAGE documents, giving the scientific evidence supporting the government’s response to COVID-19, have been published this afternoon.
Third-party comments:
Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:
“The level of uncertainty in the modelling results is really very large. The SAGE documents make it very clear that that’s largely because we don’t yet know enough about the Delta variant. You might wonder why we don’t know enough – the variant has been around for some time now – but quite extensive amounts of data are needed to make estimates of how much more transmissible the Delta variant is than the Alpha variant, or how effective vaccines are against the Delta variant in terms of reducing serious illness, hospitalisations, and deaths. So far it just hasn’t been possible to tie these down very closely. For instance, SAGE say that the Delta variant is estimated to be between 40% and 80% more transmissible than Alpha, but it makes quite a difference to the modelling projections whether it’s 40% or 80%, and anyway the SAGE documents make it clear that the estimates could change anyway. One of the SAGE documents (the SPI-M-O summary document from 9 June), in considering these uncertainties, says that, “the scale of that resurgence [in hospital admissions, without the four-week delay] is highly uncertain and ranges from considerably smaller than January 2021 to considerably higher. The difference between the optimistic and cautious effectiveness assumptions [on vaccine effectiveness against Delta] leads to a factor of three difference in the peak height; between 20% additional and 20% less transmission [infectivity] advantage leads to a factor of five difference.” This is very much not because of incompetent modelling – it’s because no model can give clear projections without good data, and the modellers (and everyone else) just haven’t got good enough data yet.
“The latest released summary document from SPI-M-O, the SAGE group that deals with modelling, says that these uncertainties mean that “SPI-M-O cannot determine with confidence whether taking Step 4 of the Roadmap on 21st June would result in a peak that might put unsustainable pressure on the NHS.” That’s relevant, because not putting unsustainable pressure on the NHS is one of the Government’s four tests in the roadmap for lifting restrictions. That sentence perhaps means that SPI-M-O can’t tell whether that test is passed, but, for me, it mainly emphasises how uncertain things are and how nobody clearly knows how big the risks are. All the indications are that there would have been increased risks, without the four week delay, but it’s not possible to say how big an increase, and certainly not possible to say exactly how increases in serious illness and death might have traded off against the economic and social costs of continuing restrictions. These are hard choices.
“The four-week delay will, of course, help in terms of getting many more people vaccinated, though the uncertainties mean we can’t be really clear on exactly how much that might help in reducing serious illness. But it will also help a lot in buying time to learn more about Delta and about how vaccines work against it. I’d argue that that could be just as important as the extra vaccinations. Not only will it make it clear the Government is doing the right thing on July 19 after the delay, it will also allow evidence-based discussion on important aspects of our lives after that. The virus isn’t going to disappear, and we might have to get used to the idea that some aspects of what we do, just a few I hope, may never be the same as before the pandemic. But I really wouldn’t want to accept anything like that without good evidence.
“A final note – I do wonder how the Government can make good decisions on the balance between restrictions on what we can do, if they have detailed modelling of infections, vaccines, hospitalisations and deaths (including information on the likely uncertainties), but no detailed modelling (that I’ve seen) on the economic and social costs of the restrictions.”
Prof Mark Woolhouse, Professor of Infectious Disease Epidemiology, University of Edinburgh, said:
“Projecting the course of the Covid-19 pandemic remains a challenge. The UK government is being guided by epidemiological models calibrated to the latest data for the delta variant, which is significantly harder to control than the alpha variant. These models generate projections – and the (very large) uncertainties around these projections – of the next phase of the UK epidemic, for scenarios where Stage 4 relaxations are implemented on June 21st or are delayed, for periods of up to several months. These projections have been published by SAGE today.
“Despite considerable uncertainty, there is a consistent pattern that delaying Stage 4 for 4-5 weeks has a significant public health benefit. It reduces (by 20-30% in the central scenarios) the total number of Covid-19 hospitalisations over the coming year. It has a bigger impact on the peak (40-50% reduction) and pushes the peak further into the autumn.
“However, the same models also generate a significant wave even without any Stage 4 relaxations at all. If we do progress to Stage 4, in the worst case hospital admissions could be on a comparable scale to the past winter.
“The model outputs also show – but do not highlight – the expected situation on July 21st. Daily numbers of cases are projected to be several times higher than they are now and to still be increasing at that date. Though the number of hospitalisations may still be low, the models generate a rapid increase after July 21st. On the basis of the data we should see if these projections are correct, this could make it extremely difficult for the government to decide that it is safe to lift restrictions on July 21st. Conversely, if daily numbers of cases are falling by July 21st then that will indicate that the more pessimistic model projections were not accurate and the delay was only precautionary.”
Prof Rowland Kao, the Sir Timothy O’Shea Professor of Veterinary Epidemiology and Data Science, University of Edinburgh, said:
“The modelling of the impact of the delta variant was conducted by three independent research groups working with different modelling approaches. That the three models agreed on the potential for a considerable increase in the impact of the delta variant on serious illness and deaths, is a good indicator of the seriousness of the situation. The considerable uncertainty associated with the outcomes is a real indicator of the uncertainty in the data – there is much we simply don’t know. Any decisions announced today reflect that uncertainty, and thus it is entirely possible that the outcome over the next few weeks will be substantially fewer cases and deaths than the extremes of the modelling scenarios – while all models can of course be wrong, such an outcome should not in itself be an indicator of fundamental flaws in the models themselves.”
Comments from scientists involved in the modelling that has been published today:
Prof Neil Ferguson, Director of the MRC Centre for Global Infectious Disease Analysis, Imperial College London, said:
“We’re at a critical point in the ongoing race between the virus and our vaccination programme. The country is starting to see an uptick in hospitalisations and our understanding of the precise efficacy of vaccines against the Delta variant is still developing. Step 4 will increase virus transmission when it goes ahead, so the delay announced gives more time to boost vaccination coverage and to refine assessments of future risk before we move forwards once more.”
Dr Anne Cori, Lecturer in Infectious Disease Modelling, Imperial College London, said:
“Unfortunately the Delta variant is growing quickly and has become the dominant variant in the UK. Our latest models suggest that this variant has the potential to cause a third wave of hospitalisations and deaths, of a magnitude that is still highly uncertain. In all the scenarios that we looked at, delaying the easing of restrictions by a few weeks would reduce the size of the wave and result in fewer deaths.”
Dr Marc Baguelin, Lecturer in Infectious Disease Modelling, Imperial College London, said:
“The number of infections, hospitalisations, and even deaths could grow rapidly in the next month. There is uncertainty around the size of a third wave and more time is needed to assess the transmissibility of the virus and the efficacy of vaccines against the delta variant in particular for the most severe outcomes. Delaying the easing of restrictions would reduce transmission and enable more people to get protection through vaccination.”
Dr Nick Davies, Assistant Professor of Mathematical Modelling, London School of Hygiene and Tropical Medicine, said:
“I think it can be hard to understand how the modelling can suggest we may be facing a third wave of a comparable magnitude to previous waves, when so many people have been vaccinated. Surely, we would expect a smaller wave this time around because of vaccines?
“The crucial difference is that wave 1 in April 2020, and wave 2 in January 2021, were both stopped short by lockdowns, which prevented either peak from reaching their full heights. For this third wave, we are returning to almost-normal behaviour, and the projections are showing what would happen if no further action was taken by policymakers to attenuate any resulting third wave. So these waves are reaching their full height rather than being cut short by any lockdowns.
“Another important difference is the increase in transmissibility, lower efficacy of vaccination, and (in some scenarios we prepared) increase in severity for the Delta variant relative to previous variants.”
All our previous output on this subject can be seen at this weblink:
www.sciencemediacentre.org/tag/covid-19
Declared interests
Prof Kevin McConway: “I am a Trustee of the SMC and a member of its Advisory Committee. I am also a member of the Public Data Advisory Group, which provides expert advice to the Cabinet Office on aspects of public understanding of data during the pandemic. My quote above is in my capacity as an independent professional statistician.”
None others received.