select search filters
briefings
roundups & rapid reactions
before the headlines
Fiona fox's blog
NHS Test and Trace have released the latest data for COVID-19 in the UK.
Prof Sheila Bird, Formerly Programme Leader, MRC Biostatistics Unit, University of Cambridge, said:
“Background: On 23 July 2020, the Royal Statistical Society made recommendations – thus far ignored – to remedy the failures of Test & Trace (T&T) to glean intelligence.
“First, we need information about the symptomatic test-positive rate for T&T’s two high-risk quarantined groups:
“a) Members of the household of symptomatic index cases
“b) External close contacts of symptomatic index cases.
“Second, robust intelligence is needed – requiring home-visits on randomly sampled days – about asymptomatic swab-positive rates in the first five days of quarantine and in the next nine days for T&T’s two high-risk groups.
“Third, those random home-visits allow adherence to “stay at home” to be properly audited.
“Today’s evidence session before the eminent House of Commons Science & Technology Select Committee included penetrating questions about end-to-end performance of Test & Trace (contacts reached within 72 hours of swab-date when index case was positive); why adherence to “stay at home” was not being audited in a well-designed manner, as recommended by RSS; and why T&T was failing to learn about in a rigorous manner (such as via random swab-visits) about asymptomatic positivity and how soon those quarantined – either a) in the household of symptomatic index cases or b) as external close contact – were swab-test positive.
“Reasons given for T&T’s failure to answer such scientifically-astute questions were:
“We have bit-by-bit performance measures [because end-to-end is inevitably less favourable . . . ?]
“Discoveries by record-linkage or random-visits won’t change the chance that a close contact becomes symptomatic [true but we’d learn what percent do develop symptoms and test positive; about explanatory factors for positive-rates; and how soon positives are recorded. If positive-tests were always in first 7-days, this could and should lead to refinement of quarantine period from 14 days!]
“We’d learn too much too rigorously about non-adherence when we prefer a consent-based approach anyway [essentially, if adherence were badly incomplete, we might then have to deal with the problem . . .]
“No to random-visit swab-testing because the public already thinks that testing negative equals “get out of quarantine early”! [but random-visits represent an opportunity to explain to quarantined individuals that a negative swab-test does not currently allow them to shorten their quarantine period but their random-test-result, combined with others, could provide robust evidence to check-up on whether quarantine-period could safely be shortened in future. Random-visits are also an opportunity to thank us for our compliance and demonstrate scientific method in practice.]
“Mid-September 2020 update: 82,000 symptomatic index cases had been transferred to T&T by 2 September 2020: 65,000 were reached but only 48,000 identified close contacts. Of 178,000 identified non-complex close contacts (61,000 external), only 104,00 were reached (39,600 external) and asked to self-isolate by T&T. Table add results for 3 to 9 September and end-to-end summary.
“Without heed to timeliness, which makes matters worse: 85% positives referred to T&T * 83% of referred index cases being reached by T&T * 83% reached index cases identify contacts * 64% non-complex contacts being reached by T&T and asked to self-isolate implies only 37% as end-to-end percentage for self-isolation of contacts of non-complex symptomatic test-positive index cases.”
Dr Simon Clarke, Associate Professor of Cellular Microbiology at the University of Reading, said:
“These latest statistics show that the number of newly diagnosed infections continues to climb sharply in the UK, while the number of deaths is currently low. This has sparked much debate about whether this constitutes a true ‘second wave’ of infections, or whether the increase is down to more testing, despite the delays in the testing system.
“It remains the case that younger people are over-represented in the new diagnoses in the UK and that they are at lower risk of developing serious disease. But it’s now becoming painfully clear from France that infections in low-risk groups can spread to people at higher risk.
“The developing culture clash about whether this is a second wave, or whether testing data is significant, is losing sight of the fact that Covid-19 has the ability to put large numbers of very unwell people into hospital, very quickly, preventing healthcare systems from treating their normal caseload. When the hospitals start filling up, who cares about the definition of a second wave?”
Prof Kevin McConway, Emeritus Professor of Applied Statistics at The Open University, said:
“The large rise in the number of positive cases in the latest Test and Trace figures for England for the week ending 9 September, compared to the previous week and the week before that, leaves no doubt that infections are really increasing. As always throughout this programme, the great majority of people, who are tested, don’t have a positive test result. But the percentage that do have a positive test result has gone up as well. It was 3.2% in the most recent week reported, compared to 2.3% the previous week and 1.5% the week before that. Other things being equal, that would also indicate an increase in the number of positive cases in the country.
“That’s all certainly worrying. But I should mention that these figures can’t give us a precise measure of how fast the true number of positive cases in the country is increasing. That’s basically because you can’t be counted as a positive case in these statistics unless you have been tested. And whether you’re tested depends on whether tests are available, where they are available, and the sort of people that can be tested. Those things can all change over time, and may have been changing quite rapidly. A lot has been said about the availability of tests recently. This report shows that a lot more people were tested in the week ending 9 September compare the previous week – 27% more, so that more than 5 people were newly tested that week for every 4 newly tested the week before. Other things being equal, which they never completely are, that should lead to an increase in the number of people found to be positive – but the increase in positive cases was larger, which is why a greater percentage of the tests gave a positive result. But other factors may be acting too. We’re told that testing has been more concentrated in areas where the infection rate is known to be high – that makes sense in terms of getting on top of the disease, but if you look in the places where people are more likely to be infected, you’ll find more infected people, even if the total number of infections in the whole country isn’t rising so fast. We’ve also all seen many reports that tests can be difficult to get – other things being equal, that might lead to the increase in new positive cases being lower than the real increase in the country. But people are also being told, more emphatically than before, not to request tests unless they have symptoms. Previous data from, for example, the Office for National Statistics (ONS) infection survey and a large survey (REACT-1) run by Imperial College have indicated, not surprisingly, that people with typical Covid-19 symptoms are more likely to get a positive test result than are people who don’t have those symptoms. So if more of the people being tested have symptoms, then, other things being equal, you’d expect more positive cases, whether or not there are really more positive cases out there in the country. And it’s plausible (but here I’m speculating, really) that people who have no symptoms, or have only very mild symptoms, could be more easily put off trying to find a test, if it’s hard to get a test appointment, than would be people with more marked symptoms. All these different possible tendencies can work in different directions, and this report doesn’t include data allowing us to assess most of them.
“So yes, we can be pretty well sure that the number of infections in the country has increased substantially, but we can’t be sure, from this report, how fast it is increasing. That might well matter in terms of deciding on the best policies to deal with the increase. The latest data from the ONS Infection Survey should be published tomorrow. The people tested in that survey are tested simply to measure the overall level of infection in the country, so its results are not affected by changes in Test and Trace in the number of tests carried out, the availability of tests, or where the tests are being done and exactly who has access to them. That should provide clearer information for policy makers (and the rest of us).
“One more point, for clarity. The report says that the number of new positive tests in the week ending 9 September showed “a substantial increase of 167% compared to the end of August.” That’s true, but perhaps it’s not entirely clear which comparison is being made. The comparison is between the week ending 9 September and the figure for two weeks before that, the week of 20-26 August. Previously these Test and Trace reports have more usually made a comparison with the previous week, not two weeks before. The number of new positive cases for the week ending 9 September is 75% up on the week before (ending 2 September). That’s still a substantial rise for one week, though a good bit short of the doubling of new cases in a week that was being suggested in several recent media reports. (The new confirmed cases for the week ending 2 September were 52% up on the week before that (ending 26 August), and putting that together with the 75% increase for the most recent week produces the figure of a 167% increase for the latest week compared to two weeks earlier.)”
Prof James Naismith FRS FMedSci, Director of the Rosalind Franklin Institute, and Professor of Structural Biology, University of Oxford, said:
“The case numbers are now clearly growing, I expect the ONS data tomorrow to show the same thing. At the moment, it seems the virus is mainly in the young, who are largely unaffected by it. If the virus continues to spread, we will see hospitalisation and serious illness when it reaches into the older population. Entering into winter with a growing number of cases is of concern. Whilst the virus remains deadly for some, we can take some comfort from scientific and medical breakthroughs that will greatly reduce the number of deaths.
“The trace system is beginning to ramp the numbers it handles as the case numbers increase and although performance fluctuates it seems reasonably steady. The number of people entering tracing increased by 74 % yet the number of positive cases grew by 167 %. A similar disparity was seen last week, this could reflect the differences in timing (the test comes first, transfer comes later). It would be helpful to report on the number of people who do not enter tracing within 2 days of testing, the trend in this number will be a good guide for the robustness of the system. Last week I questioned the value of satellite / home testing given the delays. Tests which take many days to report and action, are of no value in suppressing the pandemic.
“The wider question which is often ignored in a media focus on number scoreboards, is what was the system set up to achieve? Cases are growing and we are introducing increasing lockdown measures. Objectively, I think many of us hoped a robust test, trace and isolate system would have prevented both of these.
“The system is not currently achieving its goals. I do not see enough discussion about this bigger picture.
“Obviously in the first wave, the UK was one of the worst hit countries and we have a correspondingly harder task since the level of infection was higher than almost anywhere else.
“More testing will help but I do not believe the issue is the number of tests, whilst there are obvious issues, the positivity rate suggests this is not a fatal problem. The tracing effort seems to have stabilised, it would be great if we could improve the turn around time and reach a higher percentage of contacts.
“However, SAGE estimated that 20 % of those with symptoms isolate and the figure for asymptomatic but infectious individuals is lower still, I fear we are ignoring this elephant in the room. It’s a three legged stool, test, trace and isolate; more focus is needed on isolation urgently.”
Dr Daniel Lawson, Senior Lecturer in Statistical Science, School of Mathematics, University of Bristol, said:
“Due to the lack of a system to record any excess of tests being requested over those available, we have unfortunately reached another point where the critical data needed to track COVID-19 are not present in the official statistics.
“The scale of the problem is hard to tell from anecdotal error, but from a social and statistical perspective this is a problem. With tests being unavailable, some people will elect to not get tested, either to free tests for others or due to limited enthusiasm for fighting to get a test. Lack of trust is likely to reduce testing rates into the future. Statistically we lose power to distinguish what the true infection rate is in the population and management is severely compromised. Test and trace is going to be strongly impacted because it relies on timely tests being performed; even a small delay decreases its efficacy.
“These may therefore be the last reliable figures for some time, ending on the 9th September. The system was still coping at that time, with test and trace functioning as ever and only a small increase in test time. The infection rate now will be substantially higher, and there are no data on how many people want a test but cannot get one, or what delays this will induce.”
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 member of the SMC Advisory Committee, but my quote above is in my capacity as a professional statistician.”
None others received.