The govt has recently started trials in schools to replace 10 day quarantine for contacts with daily testing with rapid (LFD) tests for 7 days. The govt is using these as 'green light' tests contradicting both MHRA & CDC recommendations. 🧵…
There is very little transparency around this - to my knowledge the govt has not released ethics documents. It's clear from the information sheets provided to parents that consent is being sought only from children/parents who are willing to substitute isolation with testing.
This doesn't make any sense from an ethical perspective because the entire bubble is exposed to risk when a child who's been exposed is not isolating and rather being tested with LFDs that are known to miss infections, which is why they are only recommended as 'red light tests'
What I find all the more surprising as a researcher is that the participant information sheets shared with me by concerned parents do not mention insurance or liability/compensation, which the current research guidelines in the UK require mention of:…
It's very unclear what liability or compensation will be provided to children or staff or parents or household members who may develop long COVID, or even get serioously ill, possibly due to spread of infection that may have been missed by LFDs.
The govt appears to have based much of the evidence for the basis of the ethics of the trial on modelling studies- that suggest that LFDs would miss a significant proportion of infections, despite optimistic assumptions that may not reflect reality - discussed below.
When more realistic assumptions are considered, the comparison suggests this approach is even more sub-optimal than the model seems to suggest at first. Essentially, the study shows that in all scenarios LFDs miss infections to different extents relative to quarantine.
The study states '7 days of sequential testing at approximately 20% higher adherence matches the effectiveness of 10-days quarantine'.

So LFD testing would only match effectiveness of quarantine if adherence was 20% higher for testing than quarantine. How likely is this?
Earlier model scenarios suggesting good LFD performance in a previous study assumed 50% adherence to quarantine & 67% to self-isolation. Given schools often identify contacts, who are then asked not to come into school for 10 days, adherence with quarantine is likely to be high.
So it is very unlikely that adherence would ever be 20% higher for testing than quarantine - which means that LFDs as per the model would almost always miss cases relative to quarantine.
This paper states "the likelihood of detecting a case is determined by the Ct value at time of testing". While there is a correlation between Ct values & detection with LFDs, we know that this is not consistent. The model acknowledges results are v. dependent on this assumption.
The following study has been used by the DfE to provide evidence for LFDs in detecting 'infectiousness':
'modelling has found that they would be effective at detecting up to 90% of the infections that the individuals then passed on to their contacts."…
Is this generalisable to the current school setting?
No. This study essentially considers virus levels among people who were PCR positive in the community who had contacts who were also PCR positive, and suggests what LFD accuracy may be in this situation looking at Ct levels.
Given PCR testing is generally carried out for symptomatics, who may have different virus levels compared with asymptomatics (also this is largely based on adult virus loads & transmission), this study is likely to overestimate the accuracy of LFDs in identifying 'infectiousness'
There are many studies where even with low Ct values sensitivity for LFDs has been low. It is thought that much of transmission happens at Ct values below 33. Sensitivity has varied across different settings - depending on what Ct threshold is used.…
This BMJ report suggests that even with a Ct threshold of 25, sensitivity was only 47% by contrast to the assumption in the Oxford study presented by the DfE where assumptions are far more optimistic - suggesting LFDs would pick up most cases with Ct<30.…
This distribution of Ct may not generalise to asymptomatic contact testing in schools, where the sensitivity of these tests may be lower if virus levels are lower(or lower test accuracy in children)- but spread may occur as tests are not perfect indicators of 'infectiousness'.
Accuracy increases as Ct drops (higher virus loads), but there is no indication that sensitivity of LFDs in identifying 'infectiousness' would be high enough in the school setting among exposed contacts to justify the current trial, given the risk to children, staff & households
Indeed, the results show that in all cases LFDs are always associated with more infections than quarantine. Important to remember that we know even less about LFDs capturing infection accurately among children, given even accuracy of PCR tests can be lower among children.
This is not to say at all that LFDs are not useful. They are- they have very high specificity - and are good 'red light tests' for those who are at low risk. But as the MHRA has stated, using them as 'green light tests' is dangerous- because using them this way can increase risk.
As mentioned before, a previous model model with rather unrealistic assumptions has previously shown that 'daily contact tracing for five days with LFTs is only 12% less effective than 14-day quarantine in averting onward transmission' This seems a small difference, doesn't it?
It isn't. The study only models the risk of only a single generation of many superspreading outbreaks. Given that a lot of spread is schools is asymptomatic, and children are less likely to show symptoms, 12% misses in infection in one generation can have big downstream impact.
If these are asymptomatic, they would not be picked up by the process, and would continue to transmit asymptomatically & exponentially, so the overall impact would be more than 12% given several generations of spread could go undetected under before a symptomatic case is found.
The authors state that delay in contact tracing also determines no. of cases missed by LFD vs quarantine- with more being missed the earlier the tracing is done. This often happens in schools where contact tracing can be immediate, if the index is a child/staff member in school
Notably, the model the govt appears to have based it's own evidence on concludes by suggesting that rapid testing is not a substitute for other measures, but rather an adjunct. Why then is it being used as a substitute rather than an adjunct? Against the advice of the MHRA?

• • •

Missing some Tweet in this thread? You can try to force a refresh

Keep Current with Deepti Gurdasani

Deepti Gurdasani Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!


Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @dgurdasani1

28 Apr
I keep hearing that the UK should open up now... because Israel! There's seems to be a misunderstanding about what actually happened in Israel. Israel has had fairly strict mitigations in place & schools were opened only after 57% of the population was fully vaccinated. Thread
I keep hearing that because infections in Israel have declined during vaccination, this is a sign other countries can open up. What people forget is that Israel didn't follow a vaccine-only policy, but a multi-pronged strategy aimed at containing transmission during roll-out.
Here are a few examples:
Schools in Israel only opened on the 18th April after 57% of their population was *fully* vaccinated, and children wear masks in schools.

They had a mask mandate outdoors & indoors until very recently (much stricter than the UK)…
Read 11 tweets
26 Apr
I was invited today to on @NickFerrariLBC to comment on the recent letter by 22 scientists on ending mask use in schools, and opening up the UK without masks & social distancing by June 21. I was *not* told that this would be a 'debate' with Anthony Brooks, one of the signatories
This is a letter that's been signed by Carl Heneghan, Sunetra Gupta, Karol Sikora - the architects of the pseudoscience & several debunked papers that led to the Great Barrington declaration - that caused untold damage to pandemic response across the world.
Both Heneghan and Gupta predicted there would be no 2nd wave. Gupta said early last year that 50% of the population was already immune, and we would reach 'herd immunity' soon. Both opposed pandemic control - & suggested we should let the virus spread through young people.
Read 19 tweets
24 Apr
Concerning that in India, not only is the cost of Covishield (Astra/Oxford) vaccine above the $3 max price set by Astra, but this cost is not being borne by govt, but rather by the public (esp for those <45 yrs). Current pricing is $5.3 and $8+ through state & private hospitals🧵
The Serum Institute is the manufacturer of Covishield in India, and originally sold 100 million doses to central govt for a price of $2 per dose. The govt has decided to split vaccine roll-out between central govt (50%) and state govt & private sector (totally 50%) from May.
This means that anyone over 18 can now access vaccination through the state & private sector - these are not free, though. People 18-45 yrs have to pay between $5.3-$8 for a dose, while vaccines are free for over 45s through the govt vaccine centres.
Read 8 tweets
24 Apr
The cases in India show no sign of peaking - and they will not, because no action is being taken to prevent this in most parts of India. Cases in all states rising exponentially now & nothing imposed except night/weeked curfews in most places. Why? 🧵
Here's the situation in India - every state has its own decision making capacity & what we're seeing now is exponential rises in *all* states. There is no part of India that is spared, but exponential rises have started at different time points - some later and some earlier.
What is absolutely clear is that even the places that are not where Delhi is will be there unless they do something to prevent this- because the rate of growth looks similar across all places - it's just at different points.
Unfortunately most state govts are not doing much.
Read 13 tweets
23 Apr
I've been avoiding writing about this because it feels too close to the bone at the moment. But important to discuss- so here goes. What's unfolding in India is absolutely tragic & horrific (and was preventable). Unfortunately, it's likely to get worse before it gets better. 🧵
We've been seeing rapid growth in the pandemic, which hit some parts of India (e.g. Maharashtra, W. Bengal, Delhi) earlier than others - but cases are now growing exponentially in every part of India. India has had >300K *reported* cases and >2000 *reported* deaths in 24 hrs.
Deaths & cases are being substantially underreported. The positivity rate in Delhi is 36% and testing isn't available for many. It's estimated that deaths are being underreported by ~10x, with level of underreporting varying from place to place.
Read 25 tweets
21 Apr
One of our government's 4 tests was not exceeding NHS capacity. An interesting discussion with @Kit_Yates_Maths recently with @OwenJones84, where he astutely pointed out that this hasn't been defined got me thinking about how we might go about defining this.

A few thoughts. 👇
So when is NHS capacity overwhelmed?
Is it when we run out of oxygen in hospitals?
Is it when we run out beds despite surge capacity?
Is it when we run out of ventilators?
Is it when we start creating surge capacity by using resources from other routine care settings?
Worth remembering surge capacity is often created at the cost of routine care - public & private. Also when we create surge capacity, the number of staff treating the patients are often still the same- it's just the no. of staff per patient has changed.
Read 12 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!