Thread by @AppgCoronavirus

🗣 Dr. Claire Steves:
“While I’m doing epidemiological research, I’m also a clinician that sees people with Long Covid as well. I’m sitting in my clinic today. I’m a geriatrician which means that I work in a memory clinic, that’s why I.. 1/
#APPGCoronavirus #LongCovid

“..see some patients that are referred through to me from Long Covid clinics but my role today in talking to you is I’m Cohorts lead for the CONVALESCENCE study which is one of the national core studies. Elements of it are looking at Long Covid but I’m also PI on the ZOE.. 2/

“..study on Long Covid so I’ll tell you about a bit of the research from there. You guys asked me just last Thursday to answer 3 questions so that’s what I’m going to focus on today. The first one is ONS data a reliable estimate of the no of people living with Long Covid.. 3/

“..in the UK? Then, do we know if vaccines cut the risk of developing Long Covid? And is there a difference in terms of recovery between those who were hospitalised and those who were not?” 4/

“I’m talking from 3 different population studies. One is the ZOE App which is a volunteer study so it’s not completely representative of the UK but it’s really good because it’s prospective logging of symptoms and we’ve got about 1mil volunteers who’ve been logging.. 5/

“..since right at the beginning of the pandemic so that’s a very rich data source on this. Also as part of the national core studies with the cohorts lead I’m looking at 9 longitudinal studies with pre-pandemic data to look for risk factors for Long Covid and to compare to.. 6/

“..pre-pandemic symptom reporting and then as part of that study been looking at electronic health records of people with a Covid-19 code. Which we can contrast against the other estimates.” 7/

“The study that we published in ‘Nature of Medicine’ in October 2020 which was one of the first studies to look at this in a population level in the UK and it’s brought out what ends up being the key factors that determine how prevalent Long Covid might be in your population..8/

“..that you might be studying. As Danny has pointed out, it’s prevalent in every population. But dependent on these 3 main factors it might be slightly different prevalences reported.” 9/

“Key things are: the severity of infections reported & we can see every study that’s looked at post-hospitalised patients, the prevalence of Long symptoms even going out more than a year is much greater than those where they’ve not been hospitalised.” 10/

“That’s not to say that people who have not been hospitalised won’t get Long Covid, it’s just the prevalence is lower. In our studies we’ve seen that even in people who are not hospitalised, the severity of that first week of symptoms is really important.” 11/

“This graph here is basically a nonogram which we published in this paper, if you just take 3 things: the no of symptoms experienced in the first week, the age and the sex of the person, you can get to a model which really quite accurately predicts their risk of going on to.. 12/

“..get Long Covid. You can use different thresholds based on what you might want to be doing. So if doing a trial that might prevent Long Covid, you might have a different threshold for treatment. The point is using those 3 things we can identify people who might be at risk.. 13/

“..from very early in the disease. The other factors that influence population studies estimates are about how they’ve actually determined the Long Covid diagnosis or label. In the ZOE symptoms study we’ve had lower prevalence because we’ve been looking at prospective.. 14/

“..symptom logging. We’re not asking people retrospectively to consider over their illness. We’re asking them on a daily basis. When we looked at the same groups of people, we’ve looked at prospective logging vs retrospective we find about a doubling of the numbers of people..15/

“..affected. In this paper we found that about 13.5% of people had symptoms for longer than 28 days. Prospectively longer that’s continuous symptoms and 2.3% of people had symptoms for longer than three weeks. When we asked retrospectively it was approx double that” 16/

“We were very careful to remove, because we could look at the prospective logging, remove people who had symptoms prior to their diagnosis and that was one of the strengths. I think this is another important factor in any study, you need to make sure that the study.. 17/

“..has taken account of baseline reports of these quite common symptoms. So ONS data, we’ve already talked about the 500,000 figure that is people maybe having symptoms for over a year at the moment. On the 6th Jan ONS’s estimate was 1.3mil people in the UK with Long Covid.” 18/

“They used for that their 3rd and most conclusive definition of cases. That was 11.4% of people with covid experiencing symptoms for more than 12 weeks. The advantages of this definition is the ONS study is a representative example, so it’s built based on the UK population” 19/

“They had a good coverage of symptoms that they were asking about and their self assessed method of assessing Long Covid, reflected lived experience. So people were saying ‘yes, I identify myself as having Long Covid’ and that was their primary definition.” 20/

“The troubles potentially with this method, there is a bit of a response bias in that people who maybe had fewer symptoms were maybe less likely to be involved in this particular wave of the ONS study. So there could be a response bias.” 21/

“Alternatively of course, people who are very, very unwell might not be able to respond to the survey. But also they were unable in this analysis to adjust the symptomatology in non-Covid populations. So we can have a look at how they use different definitions.” 22/

“Using a slightly different definition they found quite different numbers. In their first approach where they prospectively asked about symptoms then looked to see how many of them were experiencing symptoms after 12 weeks, it’s more like about 5% of people who’ve had.. 23/

“..Covid infection. You can see within the general population which is these control participants in green, the symptoms within these 12 symptoms they looked at were substantial. But there is a clear difference between Covid-19 and the general population.” 24/

“That’s something we also published in our original paper, but there’s a very clear difference between Covid-19 duration of symptoms and other respiratory viruses that present in exactly the same way, but people have a negative test.” 26/

twitter.com/i/broadcasts/1…