Living on Plague Island

Previous posts on this site have detailed the declining amount of data available to track what is happening with Covid in the UK and to inform our assessment of how safe we are at various points in time. However, there may be some light at the end of the tunnel. It has been decided to roll out a new winter slimmed down version of the previous COVID-19 infection survey, but how good is this news?

This decline and the consequences are well summarised in a recent ‘subtract’ from Christina Pagel posted on X (formally twitter) that includes an overview of the wealth of information that used to be available from the ONS infection study which was terminated in March 2023. This is well illustrated in the graphs below.

Households like mine have been left completely at sea regarding the level of infection which makes it impossible for us to assess risk. This has made us more cautious – not less cautious. There is plenty of evidence that many clinically vulnerable people and their households feel the same. The consequences of this for the economy are potentially very significant.

Good News at Last?

To the delight of many, and possibly in the light of an obvious rise in Covid cases and hospitalisations which is impossible to miss, despite the lack of data, the UK Health and Security Agency recently announced that they were introducing something called the Winter Covid-19 Infection Survey.

The new survey is described as ‘a different study’ to the original ONS Infection Survey and will ‘involve up to 32,000 lateral flow tests being carried out each week, providing key insight into the levels of COVID-19 circulating across the wider community. ‘

However, there are a number of potential issues with what is proposed

Firstly, the reliance on self administered LFTs raises a number of issues about reliability. The fear is that overall prevalence will be underestimated. It remains to be seen how much guidance the sample receive on carrying out the LFTs as there is plenty of anecdotal evidence of people being unable to carry out testing in a way that yields an accurate result.

Reliance on LFTs also means that it will not be possible to carry out any sequencing to identify which Covid variants are rising and falling at different points in time. In the absence of wastewater screening it will be very difficult to track variants. This is a potentially serious omission as it is important to understand whether, for example, a new variant is more transmissible or leads to more serious illness than previous variants.

Neither will the new survey involve any finger prick blood tests to test for antibodies. This is a potentially serious flaw because this data would be useful to understand the level of antibodies across different population sub groups and, potentially, inform targeting of vaccine booster programmes.

It is not clear from the press release or guide to participants whether the new survey will ask whether respondents are suffering from Long Covid. A question on LC was included in the previous survey, and although it was far from perfect, it was the only measure of Long Covid available.

The main potential problem, however, is whether the sample will be representative of the population as a whole. This is critical if we are to gain an accurate picture of changes in levels of infection over the period of the survey.

Will it provide a representative picture?

The key problem is that unlike the Infection Survey the new winter survey does not offer survey participants any financial incentive to participate. This is a drastic change as those participating in the previous survey were paid £20 per test/survey completed in the form of a high street voucher of their choice. This is not an insignificant level of incentive given the cost of living crisis and is likely to have persuaded a representative range of people to participate, including people who consider that they have moved on from Covid and do not regard it as a threat.

The obvious danger with offering no incentive, beyond a few free LFT’s is that the sample will be biased towards people who are interested in and concerned about Covid. It is very possible that these people, including my household, are less likely to be infected because they are taking precautions to avoid getting infected.

I am aware that UKHSA and ONS say ‘This sample will be broadly representative of the population according to key characteristics’, but controlling for key measurable characteristics such as age, gender etc will not pick up whether the sample is biased in terms of peoples attitudes to Covid and whether it is an on-going risk.

Concluding comments

Both myself and my husband were part of the previous ONS infection survey sample and we have both been invited to, and have accepted, participation in the new survey. Our participation in the survey will be confirmed or otherwise by ONS over the next week or so. I will update this or a new blog in the light of our experiences of participating, the contents of the questionnaire and of course the reports to emerge from the survey.

It is not currently clear when the first report of findings from the new survey will emerge but my best guess is that it will not be before mid November at the earliest.

Update 12 November

I and my husband have both been invited to take part in the new survey and we received our introductory letter and packs of Lateral Flow Tests over 2 weeks ago, with a note saying ONS would be in touch shortly.

We heard nothing until 9 November when we received an extraordinary confusing email from Ian Diamond, Chief Statistician as follows:

Coronavirus (COVID-19) Infection Survey

We’re writing to update you about the COVID-19 Infection Survey, following the pause in the survey in March 2023. 

The UK Health Security Agency (UKHSA) has confirmed that COVID-19 Infection Survey data collection will not restart. 

Having reviewed its approach to COVID-19 surveillance, the UKHSA have decided that while the COVID-19 Infection Survey will stop, some studies will continue and new ones may start to monitor the nation’s health effectively.

As explained in the participant information sheet shared previously, https://www.ndm.ox.ac.uk/covid-19/covid-19-infection-survey/protocol-and-information-sheets, we will continue to use information from NHS Digital, linked to existing COVID-19 Infection Survey data collected up to March 2023. This will help us find out whether there are long-term health impacts from the number of COVID-19 infections people had during the first three years of the pandemic.

If you agreed to this, you may receive, or already have received, information about other studies in which you may be interested in taking part, as we explained in the participant information sheet. For example, you may have received information about a Winter COVID-19 Study.

These studies will have been approved by a Research Ethics Committee, or by the UK Health Security Agency or Department of Health and Social Care. It is up to you whether you choose to participate in these studies. You do not have to take part in any other studies we send you information about. However, as valued survey participants you would be uniquely placed to help support future work.

We would like to thank you for your sustained commitment and dedication throughout the COVID-19 Infection Survey and for making it such a success.

Yours sincerely
Professor Sir Ian Diamond – National Statistician 

Ian Diamond letter 9 November

I originally took this to mean that the new survey was cancelled but then had second thoughts following a second read. However, most people on social media seemed to think that the survey had been cancelled. As a consequence some people will given away their LFTs to people who happen to be ill, who knows…

I contacted ONS and, in response, they issued a correction (see below) on Friday the 10th.

Dear Gillian Smith

You will have received, or will soon be receiving, an email from office.for.national.statistics.surveys @notifications.service.gov.uk
letting you know that the COVID-19 Infection Survey (CIS) has formally closed. 

Please note, this is not the same as the Winter COVID-19 Infection Study (Winter CIS).

The Winter CIS is a new study that you have agreed to participate in. 

You may have already received your lateral flow test kits from the UK Health Security Agency (UKHSA). 

Please do not start testing using the lateral flow test kits until you have received an email letting you know the testing window is open.

We will notify you of your testing window in the coming weeks. 

Your participation will provide vital information on the effects of COVID-19 on the lives of individuals and the community. It will also help us understand the potential winter pressures on our health services. Thank you for taking part.

Yours sincerely 

Winter COVID-19 Infection Study (Winter CIS) team
Office for National Statistics and UK Health Security Agency

ONS 10 November 2023

I await further communications from ONS about the opening of our testing window but I am worried. Conducting surveys, particularly when you want participants to do something for nothing requires thought and careful handling. One of the first rules is be clear when communicating with survey participants and never waste their time.

We await the next development with some trepidation. This is a very important survey that is already running late. We are fully committed to it, but will everyone be equally so after the chaos of last week. We shall see.

Update 18 November and 28 November

The leaves have largely fallen from the trees, the first frosts of winter are in evidence, but we have heard nothing from ONS about when out testing windows for the Winter COVID-19 Infection Study will open. Neither do commentators know when the first results will be published.

On 28 November we finally received an email from ONS saying that the testing window was open. We duly did our tests and completed the survey questionnaire.

The first full results from the survey were due to be published on 7 December. Alas, this did not happen. For details on this continuing saga see my blog of 10 December for an update.