November 4, 2021

expert reaction to press release from the ZOE app saying that ‘COVID cases have probably peaked for 2021’

The press release from the ZOE COVID Study suggests that COVID-19 cases may have peaked.

An anonymous disease modeller, said:

“This is optimistic given we have just had half-term. ZOE may be right but I’d want to see the trend in hard metrics (hospital admissions and deaths) changing too before I called this one. It might also be instructive to look at the hospitalisation trends in older and more vulnerable individuals for Scotland.”

Dr Simon Clarke, Associate Professor in Cellular Microbiology, University of Reading, said:

“It’s impossible to predict what the Covid-19 case numbers will be next week or next month, based merely on recent infection numbers.  As a prediction, it appears to be based more on a ‘finger in the air’ than any robust analysis which would normally be expected to include things like seasonal changes in people’s behaviour and movements and vaccine booster uptake.

“Back in August the ZOE Covid study predicted that the population’s immunity would continue to decline, reaching as low as 50% by Christmas.  Similarly, that was not based on robust data analysis either and I do not see how predictions of continually waning immunity can sit alongside predictions of infections having peaked.

“The ZOE app is a useful tool in monitoring people’s symptoms and that is what it relies on for its data, but on its own it cannot accurately monitor the number of infections in the population, as some have claimed. 

“The app is dependent on a self-selecting cohort, so even though the sample size is large, it does not necessarily use a study group which is reflective of the UK population. 

“If we’ve learnt anything in the last 18 months, it’s that infection trends can turn ‘on a dime’ and that we should not be tempted into making attention-grabbing predictions.  That the future remains uncertain is our only current certainty.”

Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:

“Well, it would be good news if Covid cases have peaked for 2021, but I’m not sure about the evidence that they have ‘probably’ peaked. Obviously it’s true that the ZOE estimates had a peak a little over a week ago, and there have been similar declines in the numbers of new confirmed cases on the dashboard at coronavirus.data.gov.uk. Also, the latest preprint from the REACT-1 survey, published yesterday and presenting results from data collected between 19 and 29 October, showed some evidence of a downturn in people testing positive in England over the later part of that period. There’s no sign yet of a reduction in infection rates from the ONS survey, because so far its estimates go up only as far as 22 October, and the patterns of reductions in infections in the other data sources didn’t start until about then – but there will be more ONS survey results tomorrow.

“What we’ve got to bear in mind, though, is that these reducing trends are, so far, very short-term (not much more than a week). They generally coincide with the school half-term holidays, when people’s movements and the amount of mixing with others is different from during term time. Very high infection rates in school children have been a major driver of patterns of infection generally over the past several weeks. So, while I very much hope personally that the decline continues, I really don’t think we can be anywhere near certain that it will, and it could well be too much to say that cases ‘have probably peaked for 2021’. The ZOE press release does give some plausible reasons why the decline might continue, but a lot of them seem to be expressions of hope more than definite predictions. Tim Spector is quoted as saying “hopefully we’re over the last great peak of COVID in 2021” and “we’re hopeful that the trend will continue.” Well, I hope so too, but hope isn’t evidence.

“In yesterday’s REACT-1 preprint, the researchers urged caution about assuming that the decrease they had seen in infections would continue. I think that’s sensible. The level of infection is still very high, as the ZOE press release points out. We can hope that the peak for 2021 has been reached, but we still need to plan accordingly for what should be done if it hasn’t been reached.”

Prof Mark Woolhouse, Professor of Infectious Disease Epidemiology, University of Edinburgh, said:

“The ZOE project has been a valuable source of independent data on COVID cases in the UK for most of the pandemic. As the authors of today’s report point out, because ZOE monitors symptoms it can detect changes in the course of the epidemic earlier than government data or surveys based on laboratory-confirmed cases. To date, ZOE has proved impressively accurate, as demonstrated by the good agreement with those other data sources. The ZOE estimates confirm that approximately half of COVID-19 cases in the UK continue to go unreported, which is a grave concern.

“The ZOE data nicely capture the complex trajectory of the UK’s COVID epidemic over the past few months, with different trajectories in different regions and, most strikingly, in different age groups. This complexity isn’t surprising – there have been large changes in people’s behaviour and in levels of herd immunity over that period, both of which can have big impacts on cases numbers – but it does make the course of the epidemic over the next few weeks extremely difficult to predict.

“The report suggests that the number of cases in the UK has now peaked, but I don’t think that is the most important feature of the ZOE data. Recent trends in cases have been driven by big waves of infection firstly in younger adults and then in school-aged children. The great majority of those age groups are at minimal risk from novel coronavirus and are unlikely to end up in hospital.

“Almost hidden by the high and volatile cases numbers in young people, since June there has been a steady, gradual upward trend in cases in older and much more vulnerable age groups. This trend continues. Even though the vaccines give very good protection against severe disease, they do not give complete protection, and these age groups continue to dominate hospital cases.

“I would not want to conclude that the UK COVID-19 pandemic is in decline until it declines in older and more vulnerable age groups. That has not happened yet.”

Prof Paul Hunter, Professor in Medicine, UEA, said:

“The suggestion that covid cases have probably peaked for 2021 is likely to be correct. As was reported in the past few weeks most, but not all, of the models for COVID presented to SAGE have suggested that the epidemic would peak late October/early November before declining towards the end of the year. In addition daily reported cases numbers for the UK and for England have been falling since about the 18^th October, based on specimen date. One of the uncertainties with the daily reported cases has been whether the half term holiday could have led to an apparent fall because of reduced testing of children when not attending school. Although this may have contributed to some of the fall it could not account for all as case numbers started to decline the weekend before the start of the holiday and so far we have not seen a rebound increase because schools are now back this week, though we should wait another couple of days to be certain. So although the Zoe group state that they are always a week ahead of other incidence surveys, they do appear to lag behind the daily case reporting.

“The most recent ONS infection prevalence survey data published last Friday showed a continuing increase over the previous week but the dates on which that data is based straddle the peak date based on daily reporting. It will be tomorrow before we see any impact of a decline in rates in this data and even then because of the fact that this is prevalence and not incidence data rates are likely to fall more slowly than the incidence rate.

“So in summary today’s report from Zoe is consistent with a number of other sources that cases may have peaked for this year. Although we can expect to see waves of the infection next year, it is likely that we won’t see infection rates as high as they are now.

“Hospitalisation rates have pretty much plateaued over the past week, though still drifting upwards by a couple of % a week on average. Nevertheless, I would expect to see hospitalisation rates to start falling soon in part because falling infection rates overall and partly because of the rollout of the booster vaccine to those age groups most likely to be hospitalised if infected. Emerging evidence shows that the booster dose substantially reduces the risk of hospitalisation compared to people who have just had the first two doses https://www.medrxiv.org/content/10.1101/2021.10.07.21264626v1. This analysis suggests that a third dose of vaccine would reduce severe disease and deaths by over 90% compared to those who have only completed the initial course.”

Prof Jonathan Ball, Professor of Molecular Virology, University of Nottingham, said:

“Most schools have just had their half-term break and this likely reduced transmission amongst children; the group where significant numbers of cases have been reported.

“It may well be that we will start to see a slight rise as schools return this week.

“But one consequence of all this virus transmission in younger people is that they will become immune and as immunity builds transmission will slow. Whilst I don’t expect massive increases in case numbers I think case numbers over the next few months will still be lumpy as the virus continues to circulate in unvaccinated people and in those whose immunity has started to wane. That’s why it’s so important for people to get vaccinated and, when invited, get their booster jab.”

Prof Rowland Kao, the Sir Timothy O’Shea Professor of Veterinary Epidemiology and Data Science, University of Edinburgh, said:

“Though there are some biases in the ZOE app data (the data are collected voluntarily, and so the population being sampled may not be representative), the indicators from the ZOE app are consistent with other sources, and thus the broad indications that cases have peaked should be taken as good news. Hospitalisations should be watched closely as they are a key metric and always lag behind cases – and as has always been the case, hospitalisations are more likely in those with existing health issues, and in the older population. As the peak in cases has been driven in part by teens and younger adults, we need to be careful in overinterpretation of the raw data. Further caution should also be taken in considering mixing patterns as we head towards the winter – should these result in increased risky contacts, then a further upturn is possible. We also have continued uncertainty about other respiratory infections which can also cause severe disease and increase burden on hospitals, further complicating winter planning. Thus existing measures and recommendations, including ensuring good access for uptake of boosters and doses for teens, and continued emphasis of the importance of non-pharmaceutical measures (masks, physical distancing) remain important.”

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.”

Prof Mark Woolhouse: No COIs to declare

None others received.