National COVID-19 Infections Survey reveals changes to pandemic over time | University of Oxford
Woman in a supermarket looking at fruit wearing face mask
Woman in a supermarket looking at fruit wearing face mask

Credit: Shutterstock

National COVID-19 Infections Survey reveals changes to pandemic over time

Data from the National COVID-19 Infection Survey, done in partnership between the University of Oxford, the Office of National Statistics, Public Health England, University of Manchester and the Wellcome Trust, has revealed detailed characteristics of England’s coronavirus pandemic, including which factors have contributed most to case numbers over different phases and the prevalence of asymptomatic infections.  

The study, published today in The Lancet Public Health, illustrates the substantial drop off in cases over the summer, before cases began rising again from the end of August, a trend that has continued through the autumn. The data analysed was from 26 April to 1 November 2020, using a representative sample of private households in England (totalling 1,191,170 coronavirus test results from 280,327 individuals). 

Working outside the home and having a patient-facing role in health or social care was most associated with a positive COVID-19 test in the spring 2020 peak of the pandemic, while age (young people under 25) was the most important factor for positive tests in the autumn 2020 peak. The reasons for this difference could be attributed to changes in behaviour and movement. A possible explanation for the lower rates in vulnerable groups during the second peak is the development of antibodies in those who previously had the virus, as well as better measures to reduce the chance of infection for key workers.   

There were really large differences in which kind of people were at highest risk of infection in the first and second peaks, illustrating how quickly things can change with a highly transmissible virus like SARS-CoV-2.

Sarah Walker, Professor of Medical Statistics and Epidemiology at the University of Oxford and Chief Investigator and Academic Lead for the National COVID-19 Infection Survey, says: ‘There were really large differences in which kind of people were at highest risk of infection in the first and second peaks, illustrating how quickly things can change with a highly transmissible virus like SARS-CoV-2. Just knowing who was most at risk in April – May didn’t tell us much about who was going to be most at risk in September.  

‘We should expect this could change again in any third wave. Studies like this, which allow us to monitor the pandemic in real-time, are important for identifying groups where infections are rising the fastest in the community, giving us the chance to intervene specifically in those groups to try and prevent spread of infection and subsequent increases in hospitalisation and deaths.’  

Dr Koen Pouwels, a senior researcher in Oxford University’s Nuffield Department of Population Health and lead author on the study, says: ‘There is inevitably a lag between increases or decreases in the number of infections observed in community testing for those with symptoms, and then subsequent hospitalisations and deaths, especially when case rises start in young people who have lower risks of these poor outcomes.   

‘Without large community surveys such as ours, changes in case numbers may be missed, which could lead to delayed implementation of effective interventions needed to control spread of the virus.’  

Without large community surveys such as ours, changes in case numbers may be missed, which could lead to delayed implementation of effective interventions needed to control spread of the virus.

Another important finding of the survey was the prevalence of asymptomatic coronavirus cases. Between 45% - 65% (depending on the point of the pandemic) of positive COVID-19 tests were found in individuals without any symptoms.  

Professor Walker says: ‘This is a huge challenge, because they may unknowingly transmit the virus on to other people, and probably explains why the virus has been so hard to control. This also probably explains why large-scale lockdowns, which restrict the movement of everyone regardless of symptoms, have been some of the most successful interventions.’  

Professor Sir Ian Diamond, UK National Statistician, said: ‘The UK infection survey has proved a vital source of intelligence on the spread of the pandemic and is probably the largest regular study of its kind worldwide.  It is likely to remain important for some time, especially as it starts to monitor the impact of the vaccines now being deployed.’ 

Dr Yvonne Doyle, Medical Director at Public Health England, said: ‘We cannot assume there is any single pattern to who and how this virus strikes. Over the past ten months the people who were most likely to test positive has changed considerably, and a very high proportion had no symptoms at all. This underlines the need for all of us to remain vigilant. 

‘We have made a huge step forward this week with the first vaccines being rolled out. However, in order to control the virus and save lives, you should remember to make sure to reduce your contacts. Keep your distance, wear a face covering in enclosed spaces, and wash your hands regularly.’ 

We cannot assume there is any single pattern to who and how this virus strikes. Over the past ten months the people who were most likely to test positive has changed considerably, and a very high proportion had no symptoms at all. 

The study will continue monitoring the pandemic in the UK on a weekly basis to look for early warning signs for rising infection rates in different regions, sub-regions, and demographic groups. This will be particularly important in January after the Christmas holidays. 

The National COVID-19 Infection Survey is the first longitudinal community survey of SARS-CoV-2 infection at the national and regional levels in the UK, using data from a representative sample (nearly 300,000 individuals aged 2 and up) in England.  Participants completed a questionnaire and nose and throat swabs were taken on a number of occasions, with a median of four times per individual over the period between 26 April – 1 November 2020. This has allowed researchers to use statistical modelling to develop a real-time picture of the pandemic in the community, which has informed government and SAGE decision-making.  

Full paper: Community prevalence of SARS-CoV-2 in England during April to November 2020: Results from the ONS Coronavirus Infection Survey