select search filters
briefings
roundups & rapid reactions
Fiona fox's blog
The MRC Biostatistics Unit at the University of Cambridge have released their latest nowcast and forecast of COVID-19.
Prof James Naismith FRS FMedSci, Director of the Rosalind Franklin Institute, and University of Oxford, said:
“These latest nowcasting estimates show the number of infections increasing at around 77,000 a day (an estimated range of 53,000 to 113,000). Their model implies a doubling of the virus in 15 days.
“They estimate that by mid-November we will see between 380 and 740 deaths a day. This is consistent with a rule of thumb estimation and is now unavoidable. What matters is what happens to the number of infections over the next few weeks.
“Nowcasting has proven reliable and uses well established methods. However, there are other equally reliable methods of estimating the spread – ONS and REACT – that will help to complete the picture.
“It was depressing to see widespread cherry picking of models and data to advance an argument against the lockdown measures.
“It may be possible that the virus has plateaued, but this is not what nowcasting shows. It was not what the ONS data showed – we will have to wait for the new release on Friday to see if this continues to hold. It was not what the REACT data showed. We will hopefully soon have future REACT data to study.
“The only survey data that reports stabilisation is the ZOE model from KCL. This model is a new approach – this does not mean it’s wrong, just that it does not have the track record of the ONS-type approach.
“Neither ONS nor nowcasting use the word spiralling. Rather, what these surveys show is that the virus is continuing to spread.
“I would like nothing more than for the current measures to have worked to stop the spread of the virus. However, I see no evidence to believe that the tier lockdown model works at the country-wide level. I accept that there is some evidence that in some localities it may be doing so. However, at a whole of England scale, I don’t see it yet.
“We have to make a decision without perfect knowledge – there is, I am afraid, no option but to use judgement. This is my judgement of the numbers.
“We can choose to decide that nowcasting, ONS and REACT are reasonable guides to the next few weeks, and thus if we continue as now, we will have over 1000 deaths a day by Christmas.
“If current measures have worked (i.e. the virus has plateaued), then the lockdown will inflict further misery for a limited gain (i.e. a faster decrease in the circulating virus). For those already under Tier 3 restrictions, the loss added by these additional restrictions will be less. Clearly for those areas under Tier 1 restrictions, the new burden is heavier.
“We can choose to continue to wait to be sure/follow the KCL ZOE model. Should we follow this policy, we must be explicit about the risks and who will face them if it is wrong. The costs are hundreds of unnecessary deaths every day by Christmas.
“I strongly support scientists taking contrary views, and identifying flaws in current models or estimates has never been more important.
“It is for elected politicians to decide what to do. In making decisions, they have to consider advice not just about the number of deaths or infections, but must also consider the NHS limits, the costs of lockdowns and other factors. I have confidence that the SAGE system gives them the information.
“However, it also helps the public when scientists point out what data does not fit their theory, what their estimations of uncertainty are and what the consequences of them being wrong are. False certainty is unhelpful in a pandemic where we are learning something new every day.”
Dr Michael Head, Senior Research Fellow in Global Health, University of Southampton, said:
“The modelling estimates 77,000 new cases each day, with a doubling rate of every 15 days. This will translate into hundreds of COVID-19 deaths each day and huge numbers of patients in hospital. There are currently around 1,500 COVID-19 patients being admitted to NHS hospitals every day, and 11,000 patients are already in hospital.
“Dealing with the infection control consequences of all these patients takes up a huge amount of hospital resource, and there are around 30,000 NHS staff off work right now due to infection or need for self-isolation. This will inevitably have a huge additional impact upon other hospital services, such as for cancer care. Therefore, interventions such as those scheduled to come into force this week are urgently required to get cases low and to relieve the pressures on the NHS ahead of the intensely busy winter period.”
https://www.mrc-bsu.cam.ac.uk/nowcasting-and-forecasting-3rd-november-2020/
All our previous output on this subject can be seen at this weblink:
www.sciencemediacentre.org/tag/COVID-19
Declared interests
None received.