select search filters
briefings
roundups & rapid reactions
Fiona fox's blog
The Office for National Statistics (ONS) have released the latest estimates of the prevalence people with self-reported “long COVID” using data from the UK COVID-19 Infection Survey.
Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:
“This new ONS release on self-reported ‘long Covid’ symptoms looks at the four-week period up to 6 June. In some ways it doesn’t add a great deal to our understanding of the pattern of long-lasting symptoms, compared to the previous ONS release* on the subject a month ago. The same analyses are provided, except that the analysis of how the odds of having long-lasting symptoms depend on various characteristics, after statistical adjustment for other characteristics, is not included. Generally the patterns of prevalence mostly have not changed much since the previous report, and the changes that have occurred do generally match what I’d have expected from the patterns of infection since the previous report. That said, the new bulletin does certainly confirm that the results in the previous bulletin weren’t some sort of abnormal statistical fluke. These long-lasting symptoms are really fairly common. An important strength is that the data come from a survey of a representative sample of people across the UK community population, so the results should represent what’s actually going on across that population. Because the condition is a new one and because treatments and medical interventions for it aren’t universally available or agreed, data based on health service sources are likely to be very incomplete, and using a survey that represents the whole population can get round that limitation.
“The overall estimate of the number of people in the UK community population who report ‘long Covid’ symptoms was lower in the four weeks to 6 June than it was in the four weeks to 2 May – 962,000 compared to 1,021,000. And the estimate for the four weeks to 2 May, was slightly lower than that for the four weeks to 6 March, in the ONS bulletin before last, which was 1,094,000. There’s some statistical uncertainty about these estimates, as is inevitable with survey estimates, so the trend isn’t known precisely. But a decrease is what I’d expect. These are estimates of the numbers of people who would report symptoms persisting for more than four weeks after the first suspected Covid-19 infection that were not explained by something else. The numbers of new Covid infections over this period have been decreasing fairly rapidly. So, although the number of people that would be ‘eligible’ to have symptoms that lasted four or more weeks by 6 June is higher than the number by 2 May, it isn’t all that much higher, and meanwhile some of the people who would have reported long-term symptoms in the four weeks to 2 May would have recovered and no longer have those symptoms. At least the decreases do show that long Covid isn’t something people have forever. The ONS results provide estimates for many subgroups of the population, according to the extent to which their symptoms reduce their ability to carry out their day-to-day activities, the length of time the symptoms have lasted, their age, gender, ethnicity, where they live in the UK, and more. Looking at the estimated numbers with symptoms lasting at least four weeks, most of those estimates for subgroups have also reduced between the previous bulletin and this one, though some of them have a high level of statistical uncertainty because the number of survey participants in those groups is not very high. The rate of reporting long Covid symptoms was greatest in people aged 35 to 69 years, females, people living in the most deprived areas, those working in health or social care, and those with another activity-limiting health condition or disability. Those broad patterns have not changed since the previous bulletin. Of the 962,000 reporting symptoms that lasted at least four weeks, about two-thirds (66%, or 634,000 people) said that the symptoms reduced their ability to carry out day-to-day activities, and of those, 178,000 said that their ability was reduced ‘a lot’.
“The ONS data also provide similar estimates for people who report that their symptoms have lasted at least 12 weeks, and at least 12 months. The number who have had symptoms lasting at least 12 months increased, in the four weeks to 6 June compared to the four weeks to 2 May. It went up from 376,000 to 385,000. In fact that increase is small enough to be due, potentially, to statistical variability and nothing more. However, it certainly is plausible that it might really have gone up – and even if it hasn’t, the numbers involved are worryingly high. And 69% of those people with symptoms lasting at least a year said that the symptoms reduced their ability to carry out day-to-day activities, with 75,000 saying this ability was reduced a lot. That’s a high level of new disability.
“Looking again at the wider group of people who report symptoms lasting at least 4 weeks, the proportion whose activities are limited a lot is higher than average in those whose symptoms have lasted between 4 and 12 weeks (possibly because symptom severity tends to reduce over time, so people with longer-lasting symptoms may not be affected on a day-to-day basis as much as they previously were), in people who were hospitalised, and to some extent in older people.
“This isn’t the only information about long-lasting symptoms that is based on self-report from a population survey. Results in preprint form from the REACT-2 survey** from Imperial College came out a week ago, and produced much higher estimates of the prevalence of long-term symptoms than is estimated in this ONS bulletin. I commented*** then for SMC on some reasons why the results from the two surveys are so different, on the surface, and the points that I made then still apply. Really the underlying issue behind the differences is that there’s no clear definition of what is meant by ‘long Covid’ – and arguably it helps to have more than one survey looking at data from different questionnaires, to help us sort out what an appropriate definition and approach might be.”
** https://spiral.imperial.ac.uk/bitstream/10044/1/89844/9/REACT_long_covid_paper_final.pdf
Dr Elaine Maxwell, Scientific Advisor National Institute for Health Research (NIHR), said:
“The data is exactly is expected, tracking the infection rates. As the infection rate dipped during the spring of 2021 the number of people reporting Long Covid symptoms in the short term dropped. The increase in people reporting symptoms from at least 12 months increased but at a slower pace than it did from March to May, again matching last year’s infection rates. It is therefore likely that the recent rise in infections will see an increase in Long Covid lasting 4 weeks in the next ONS release. By the autumn/winter of 2021 we should expect to see a sharp rise in those reporting symptoms for a least 12 months, mirroring wave 2 infections.
“Self-reporting of Long Covid has been similar to recent big data studies from the USA that studied insurance claims, including UnitedHealth who found 14% of people with confirmed or suspected COVID19 infection had new diagnoses and FAIRHEALTH which reported 23%.
“Different prevalence rates relate to different case definitions (including the number of symptoms included).”
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.”
Dr Elaine Maxwell: “I am the author of the two NIHR reviews on Long Covid, although no conflict of interests”