The graph above is self-made, but based on data linked to in the text of this article.
Over the last two weeks, data has emerged from a number of sources which indicate that infections peaked nationally in the week before lockdown – and in most regions – including areas in Tiers One, Two and Three. (All findings below are for England and are shown in figure above.)
First, the ONS infection survey, the COVID Symptom Study and RCGP surveillance data all show cases rising throughout September and October, but then starting to fall at the beginning of November (possible with a slight uptick in the days between lockdown being announced and starting, as people made final visits to friends, relatives, pubs, restaurants, etc.)
Second, estimates of R from SAGE, the MRC Cambridge Biostatistics Unit and the COMIX study also match this infection data, with R rising through September but then starting to falling and reaching one at the end of October (again, with a slight uptick between lockdown announcement and starting.)
Third, hospital admissions up to November 18th have stabilised, with the seven-day average beginning to fall for the first time, and a peak possibly reached on November 11th. This cannot be due to lockdown, since this would take at least two weeks to have an effect on admissions, and so reflects changes in infections pre-lockdown.
And finally, deaths have also now stabilised, with the seven-day average being steady since November 11th and, again, this can only be due to decreases in infection pre-lockdown, since it would take at least three weeks for the lockdown to impact deaths.
(As I have pointed out before, this also matches the experience from the first wave, where infections must have started to fall during the week before lockdown, given when hospital admissions and deaths peaked, but at that time we didn’t have the infection data to demonstrate that.)
Also, as hospital admissions have started to fall, the total number of Covid-19 positive patients in hospital has also started to stabilise for the first time. This has meant that NHS acute and general bed and ICU bed capacity has not gone above normal levels for the time of year in England or in any NHS region – although some Trusts have, sadly, had to cancel elective operations.
This data all indicates that that the regional tier system was working effectively enough in controlling community spread, particularly in the over 60s, and in keeping hospitalisations below a level that the NHS could cope with.
But of course, this data was not available at the time the decision was made that a second lockdown was needed. Indeed, the data on hospital admissions at that time was genuinely alarming, and it was essential to ensure that the NHS could continue delivering all services given the huge backlog of cancer screenings, tests, procedures and operations which had arisen and is still causing so much suffering.
We had to learn the lessons from the first wave and keep all NHS services running this time so we can save and improve lives from all causes – not just Covid-19. And given the NHS has limited capacity both in beds – and particularly staff – the only way that can be done is by keeping Covid-19 admissions below a certain level – which I think we can now say is roughly at the level where we are at now nationally, but of course there are regional variations.
So what explains these findings?
Although immunity from Covid-19 infections in the first wave could explain part of this, especially in London, where antibody levels indicate their rates of infection were more than twice as high as any other region, it doesn’t explain why infection rates have also started to fall in regions with very low levels of infection from the first wave (e.g. East Midlands, South West)
The findings are however consistent with changes in behaviour reducing Covid-19 transmission, which has now been demonstrated in three separate studies – the Covid-19 Social Study, the ONS Opinions and Lifestyle Survey and the COMIX study.
All showed increasing compliance with measures (e.g. social distancing, wearing masks, washing hands, self-isolation and reducing social contacts) over the weeks preceding lockdown – and in all three tiers. This has been shown most clearly in the large Covid-19 Social Study, which has been longitudinally tracking compliance in over 70,000 people nationally asking respondents anonymously to self-report the extent to which they are following recommendations such as social distancing and staying at home.
Compliance improved over the last two months as Covid-19 infections were increasing– both nationally (as shown in the figure below) but also regionally – including in areas in Tier One (and even before the Tier restrictions came in.)
Further evidence for this also comes from the Google mobility data (also shown in the figure below) where reductions in journeys are seen from mid-September as Covid-19 infections rise and again before mandatory restrictions came in. (A trend only disrupted by the late-October spike likely caused by the leak of the second lockdown.)
This improvement in compliance shows that voluntary behaviour change is an important factor – as would be expected in that people naturally change their behaviour as they perceive that their risk of infection is increasing.
We also need to focus on improving voluntary compliance with existing restrictions which have been shown to most effective – and least harmful – before adding in new ones. Simple, consistent public health messaging is critical, and seeking persuade people without resorting to fear – which also causes significant health harm, as we saw in the first lockdown, when thousands died and suffered because of fear – either of catching Covid-19 or of putting a burden on the NHS. We need to try different messages and different messengers, including greater use of hospital doctors and GPs, who are most likely to be trusted.
As I have continually stressed in previous articles, sustainable compliance is critical to any successful strategy to manage Covid-19– especially with social distancing and self-isolation. Restrictions are only effective if people are willing (and able) to comply with them – which is also why greater support for those required to self-isolate is essential.
We now have the first evidence from England that people can change their behavior enough to control community spread, particularly in the over 60s, and so to keep hospitalisations below a level that the NHS can cope with. However, this work very regionally and so it makes sense for any measures which are required (if voluntary compliance is not sufficient) to be targeted – but not to have blanket restrictions across the whole country.
It is also worth noting that it was possible to get R below one and infections falling while schools were open, and both the Government and Opposition should be congratulated for holding their ground on this.
In conclusion, the Government is right to be ending the lockdown on December 2nd and returning to a targeted regional strategy. While we await the details, I hope the evidence that has now emerged will persuade them to adopt a more voluntary approach which will significantly reduce the harms of restrictions – to health, education, society and the economy.
Although there is now genuinely light at the end of the tunnel with the impending arrival of safe and effective vaccines, we do have to continue to live with the virus for now and maximising voluntary compliance will help to ensure that we get through the winter while minimising overall harm.