martinhutch
It’s possible, if they’ve got a large enough number of patients, and they may well revisit the dataset to look for that, and publish further papers*. But not in the first wave of a pandemic where you need to get the headline result out and published as quickly as possible.
EDIT: Actually, they haven’t even published a paper yet, just rushed out the most important findings, which are clinically significant enough to be of value immediately to medical teams worldwide.
I guess funding has to come from somewhere but it looks to me more like some soundbites to be used by politicians.
I did ask many pages ago if anyone knew if we were collecting data for example on genetics or were we cremating the data...
Sat where I am it seems that most of the world is in "ICU" mode. That is they are simply treating symptoms and not trying to understand why some people are more vulnerable to different aspects of the virus and acting pre-emptively.
Globally all sorts of political nonsense statements are being made masking any understanding. A friend of mine posted a "stop the lockdown" tweet about how in Massachusetts 98.5 of fatalities had "underlying conditions". When I pointed out that over 50% of over 18's in MA have something to be counted as an underlying condition (including him) he deleted the thread...
This was all the more disturbing because of who his wife is because if anyone has access to this information she does.
So, that Spanish study was published in February?
Have to wonder how many people in the UK would still be alive if we had been using this stuff since then.
That is they are simply treating symptoms and not trying to understand why some people are more vulnerable to different aspects of the virus and acting pre-emptively.
A bit frigging arrogant to assume that scientists all over the world are just sitting around scratching their arses and only you have the brilliant insight to try and understand the root causes of the illness.
You're wrong by the way I've read about plenty of efforts in this area, as you'd expect because scientists aren't idiots despite the public's best efforts to portray them as such.
Molgrips
A bit frigging arrogant to assume that scientists all over the world are just sitting around scratching their arses and only you have the brilliant insight to try and understand the root causes of the illness.
You’re wrong by the way I’ve read about plenty of efforts in this area, as you’d expect because scientists aren’t idiots despite the public’s best efforts to portray them as such.
Scientists? I'm one of those... I work on what I'm funded to work on with the funded resources.
This is what I mean by "ICU mode" as was explained to me by the head consultant at St Thomas's PICU
To paraphrase ... "our job isn't to work out what is wrong, just keep him alive long enough and we are very very good at that"
They did an incredible job... and when he recovered enough he was transferred to Gt. Ormond.
To me it looks like the overwhelming funding at the moment is still "ICU mode" and whilst that is wholly understandable and appropriate initially, the political will seems more focused on either a vaccine or simply keeping people from dying in a ICU way.
only you have the brilliant insight to try and understand the root causes of the illness
Not at all... I'm sure most researches would jump at the chance of being funded for understanding over treating symptoms after they have prepared a business case.
I'm disappointed that taking the BBC article and political statements made by jokers like Matt Hancock and worse a PM who can't read a side of A4... that funding isn't being made available.
You’re wrong by the way I’ve read about plenty of efforts in this area, as you’d expect because scientists aren’t idiots despite the public’s best efforts to portray them as such.
I'm pretty certain the engineers at McClaren aren't idiots either. Indeed when they got asked to design a ventilator from scratch they asked what it needed to do and said "surely it would be better to start with an existing design"
However it appears Boris thinks they are idiots and despite all the (idiot) experts telling him otherwise only he is intelligent enough to chuck money at what the experts say is a waste of money. You know... the same guy who said we had more than enough PPE, turnbed down the invitation to be part of the EU procurement ... sent infected patients to care homes with noi PPE because "care homes are safe".. and then prevented care homes actually buying PPE...
The guy who was ignoring scientific advice and bragging about shaking hands with Covid patients and then going home to his pregnant mistress? Prevented the chief nurse form attending briefings ???
Yeah... I'm sure he should be telling the scientists and clinicians what funding they get!
I was expecting a breakdown of what types of patients but perhaps thats not possible in clinical trials?
Early days - those analyses may follow if they have collected enough covariates (I'm sure they have), but just getting an interim analysis out at all for such a multi-centre trial is a big challenge. You will have no idea of the shear amount of data cleaning and processing needed as this trial has been conducted, just to arrive at any analysis. It's a huge undertaking.
Much smaller numbers, but yes.
Mortality trials need big numbers of trial participants. Those are the rules of the game.
Let's not forget, there were no Covid patients in Kettering General Hospital to shake the blonde blob's hand. Misleading the public with a self-aggrandising lie which was instantly uncovered, not so very bright at all then.
TiRed
Early days – those analyses may follow if they have collected enough covariates (I’m sure they have)
I hope so... sadly in my field I spend a lot of time asking "but why didn't you collect/keep that data"... then an equal amount of time being asked "why don't you just use the resampled (below nyquist) data"
but just getting an interim analysis out at all for such a multi-centre trial is a big challenge. You will have no idea of the shear amount of data cleaning and processing needed as this trial has been conducted, just to arrive at any analysis. It’s a huge undertaking.
and an even greater amount if time where "it would be too expensive to reformat and collate the data"
The thing is they are saying this at the same time as they are throwing away the new data or failing to keep original etc. so in 2 or 5 years time they still don't have the data or it's too expensive to reformat etc.
Usually the scientists and engineers fully realise this but they have been told by management/IT that they need to justify keeping every bit of data and write a retention policy.
"but I want to do something with it later" is usually challenged by "have you got a signed off business case"
Have to wonder how many people in the UK would still be alive if we had been using this stuff since then.
Thousands of people have been receiving it, that forms the treatment group in this study. Hopefully they called quits on the control (non-treatment) group pretty early on. Studies like this one will save lives by changing practice elsewhere, and allowing authorities to increase manufacture.
You have to have studies like this because, although it might seem intuitive to give steroids to counter a runaway inflammatory response, you need to be certain that you're not making things worse, and that's not always obvious in a situation where nearly half your patients are dying anyway, and even a good outcome may involve some significant organ damage.
And of course, now everyone has the green light, even those already doing it can stop querying if they are on the right track, and focus their efforts on identifying patient sub-groups who stand to benefit more (or less), refining the dose for maximum benefit, and coupling it with other drugs and therapies to work out the best combination.
I - You have to have studies like this because, although it might seem intuitive to give steroids to counter a runaway inflammatory response, you need to be certain that you’re not making things worse, and that’s not always obvious in a situation where nearly half your patients are dying anyway, and even a good outcome may involve some significant organ damage.
And of course, now everyone has the green light, even those already doing it can stop querying if they are on the right track, ....
...
...
...
II - and focus their efforts on identifying patient sub-groups who stand to benefit more (or less), refining the dose for maximum benefit, and coupling it with other drugs and therapies to work out the best combination.
This is what I'm missing... I hope as TiRed say's they did actually collect and collate the data as they went through the trial but the BBC article says nothing about part II.
Given the size and time spent on this study it can't be more than a few hours to segregate patient subgroups ... but there is no mention of this..
It's not that I don't appreciate the huge task in aggregating and cleaning the data but put simplistically it doesn't take appreciably longer to clean 6000 rows x5 columns as 6000 rows x 100 columns.
Hopefully they called quits on the control (non-treatment) group pretty early on.
My perspective/reading is they also continued treating some patient sub-groups for which the outcome was negative? It's just that they don't know yet and possibly need to run another study or set of studies to find out???
So, that Spanish study was published in February?
Have to wonder how many people in the UK would still be alive if we had been using this stuff since then.
Question for medically knowledgeable folk:
the Spanish study addressed ARDS, which I gather is a frequently-fatal complication of a number of diseases including COVID19. So how much of a leap is it to study use of the drug in COVID19 cases? Does the British study actually merit all the fanfare? It seems more like confirmation of what might have been already deduced from the Spanish study?, rather than an original discovery? Am I being unfair?
None of this is new or news...
The use of steroids in any kind of sepsis, or in sepsis in respiratory failure is not a new topic. It has been researched many times over the years with no actual gold standard ever produced.
Dexamethasone is not a wonder drug despite what you've been told.
https://www.cochrane.org/CD002243/EMERG_corticosteroids-treating-sepsis
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5949415/
https://www.thebottomline.org.uk/blog/steroids-in-sepsis/
https://www.jwatch.org/na48210/2019/01/09/more-about-steroids-sepsis
The recovery study is the first study to show an increase in survival. We like to say death is the hardest endpoint. You can’t argue with it (unlike time to get better). So yes it’s an important result. But no it’s now unexpected. It’s A not particularly large effect either, sadly, but other drugs will be tested on top of this new standard of care.
As for the conduct of the trial itself. This analysis will have been preplanned and protocolled. When we conduct trials, we write the analysis plan often years in advance. This must be absolutely strictly adhered to. You can’t just test what you want. That’s called “post hoc” and violates the principles of statistical inference. If you accept 1/20 as statistically significant, then guess what - you’ll find if you repeatedly test/torture your data, then left-handedness is a predictor of mortality in about 1/20 trials. So we don’t do it. Except we do sometimes in our early trials to see if subgroups might benefit. Never in “confirmatory” Phase 3 trials.
Now the Oxford Study is an academic study, but they will have employed a CRO (Clinical Research Organisation) to do the heavy lifting of data management. I have not seen the reporting analysis plan (RAP) but they will have one before doing the analysis.
People have no idea how rigorous a double blind randomised placebo controlled trial is. One Unplanned look at the data and your £200M Phase 3 trial is deemed invalid. You just don’t. Instead you have an Independent Data Monitoring Committee (IDMC) who look at the data from time to time and tell you if you should stop or not (either for futility - common sadly, or overwhelming benefit - rarer but very good news). At the end of the trial you run your planned analyses. With all the covariates you’ve collected. We just stopped a trial for PrEP in HIV for overwhelming efficacy for a monthly injectable against tablets. This is the same against placebo. It won’t be ethical to conduct placebo trials for survival now.
[TL:DR] Gold standard evidence that steroids prolong life in the very sickest COVID19 patients. Not unexpected, but good news.
So it's still good news then.. Let's hope there's more next week, the week after and the week after that..
Meanwhile in app land.
<makes shocked face>
TiReD
The recovery study is the first study to show an increase in survival. We like to say death is the hardest endpoint. You can’t argue with it (unlike time to get better). So yes it’s an important result. But no it’s now unexpected. It’s A not particularly large effect either, sadly, but other drugs will be tested on top of this new standard of care.
I guess this is my frustration ....
It seems to me that a lot of covariates would fall under "not unexpected" and that a biased sample will easily mask not unexpected results. If the overwhelming majority of patients are over 80 and have weak immune systems then this would easily mask the much lower numbers of people with a overactive immune system that it wouldn't be unexpected if they showed better results.
This analysis will have been preplanned and protocolled.
It seems to me that presenting the results without the plan and protocols is at best mis-leading. The BBC article doesn't mention the constraints .... however I find it more disturbing when some politician who is incapable of reading a full side of A4 14pt bold then makes statements and influences or even directly controls funding.
Meanwhile in app land.
What’s really had me fuming, is Matt Hancock telling us that “we found people responded better to the human side of the tracking and tracing” … and Dido going on about the app just being “the cherry on the cake” … we already knew from other countries that the human side of the track/trace/isolate programme is the most important part … this isn’t news … what is news is that, no only is that part late … but the extra and very useful tool that is a mobile phone contact tracing app is not only very very late, but is still being developed in a way that means uptake and usefulness will be far lower than would be possible if they listened to people who know what they are talking about, rather than kept up their Cummings controlled circle jerking.
It seems to me that presenting the results without the plan and protocols is at best mis-leading.
Both available online. But nobody will be interested in the RAP! I’ve only actually read the protocol (and did a little bit to help with it)
The stratification by ventilation, oxygen and no support is planned. There will be the categorical scale for severity too, but death is the simplest endpoint (hence fastest to report and the primary endpoint). I’ll read the RAP later. It will be carefully protocolled. You can’t make up analyses - that’s against the rules of inference. You pick your primary endpoint first (death here) and analyse that. The. You can have all sorts of rules to control for multiple testing of secondary endpoints (test 20 and guess what?).
Edit. Covariate marching below, so lots of subgroup potential!
Age at randomisation
Sex
Ethnicity
Time since COVID-19 symptoms onset
Time since hospitalisation
Current respiratory support requirement
Currently requiring renal dialysis or haemofiltration
Comorbidities (diabetes, heart disease, chronic lung disease, tuberculosis, human
immunodeficiency virus, severe liver disease, severe kidney impairment)
If female, known to be pregnant
Second randomisation
In addition to the above:
Type of ventilation support currently required (none, CPAP alone, non-invasive ventilation, high-flow nasal oxygen, mechanical ventilation, ECMO)
Latest oxygen saturation measurement (%)
Latest CRP measurement (mg/L)
Latest ferritin measurement (ng/mL)
Latest creatinine measurement (μmol/L)
Allocation in first randomisation
Interval between first and second randomisation
To nobody's surprise, government funded IT project is complete disaster zone for everyone except the company coining it in for supplying the thing.
Much as predicted:
https://www.nytimes.com/2020/06/17/world/europe/uk-contact-tracing-coronavirus.html
Trying to link a NYT on test and trace but it ain't having it
Seems like a good place to ask if I am being too sensitive or rightfully cautious...
A big group of leadership team and chief engineers from the business units have been "requested" to attend a capability strategy meeting on site next week. Between 15-20 people taking up a large meeting room/lecture room for 6 hours. It's in the middle of the building so no natural ventilation, just aircon.
I have no doubt we can all be way over 2m apart but it seems very excessive, if not just due to the shear cost of the meeting in person hours.
BTW, case rates and all that.
Lining up the peaks of the death figures seems to show that we had very similar deaths/per million pop at the peak, vs spain/italy, yet at this point, we are now double the deaths/per million pop at the comparable time later.
I notice the exact same problem when lining up the infection rate / per million population for the three countries, in fact on that one the UK seems to be about triple that of spain and italy at this point in time.
Hmm.
Rest of world is, once more, laughing at our predictable shitshow.
Internal news in this country is going the direction of 'tractor production is up this month'.
Having no luck with this link but I'd strongly recommend a squizz at that NYT article on Britain's response to the pandemic. https://nyti.ms/2Bj89f8
Lining up the peaks of the death figures seems to show that we had very similar deaths/per million pop at the peak, vs spain/italy, yet at this point, we are now double the deaths/per mi
Yes, Strength of lock down affects rate of decline as well. Hugely overlooked by one and all, but there is a reason why Sweden have effectively endemic transmission and deaths. It’s not just “lives saved”. Plot it on a log scale to make it even more obvious.
Yes, I can't see the UK having anything other than C19 slowly burning through the population now.
Sweden's infection rate seems to be climbing again.
TiReD
Both available online. But nobody will be interested in the RAP!
I'm struggling to match that SAP with the BBC link....
As I remember it didn't mention the randomisation or anything else... just drew some conclusions that are from Part B at this stage... which are not the quoted parts from the actual researchers.
In other words it reads to me like the Health Editor and Health correspondent didn't actually read it but spent some time instead making a infographic.
I didn't watch Hancock or Boris present this either ... but I think it's unlikely they read it either.
Yes, I can’t see the UK having anything other than C19 slowly burning through the population now.
How will a stronger but shorter lockdown prevent this from happening in other countries? Surely this will just mean that it will take the virus a little longer to re-build the infection rate.
How will a stronger but shorter lockdown prevent this from happening in other countries? Surely this will just mean that it will take the virus a little longer to re-build the infection rate.
I guess if you get the infections down to a certain point then you can track and isolate new outbreaks. Assuming, that is, that you actually have a track and isolate systen in place and not just politicians' blether.
How will a stronger but shorter lockdown prevent this from happening in other countries? Surely this will just mean that it will take the virus a little longer to re-build the infection rate.
You need to leave lockdown with low infection levels, paired with effective track/trace/isolate measures being in place. Anything else is just accepting ongoing infection and deaths.
I guess if you get the infections down to a certain point then you can track and isolate new outbreaks
That makes sense - but for that to work, you'd need a permanent qurantine in place for anyone who enters your country by any means, until global eradication or a vaccine. You'd also be very vulnerable to an infected person evading border security.
As for a government-developed track and trace system, I am almost 100% certain that the money will just go down the drain. I'm the stupid one for not working out how to get on that gravy train!
How will a stronger but shorter lockdown prevent this from happening in other countries? Surely this will just mean that it will take the virus a little longer to re-build the infection rate.
The practicalities of lockdown mean that even a half-hearted version has to end at some point.
If you can properly suppress case numbers before you release (eg New Zealand, hopefully still after this week's events), then it gives an effective track and trace system the opportunity to keep a lid on potential hotspots. This means there would be a substantial delay before any further peaks, perhaps you could suppress this altogether, maybe until a vaccine was available.
If, when you are forced to release lockdown, there is still a small but signficant baseline number of cases grumbling away in the community (which is where we are), not only does this make the chance of re-emergence much more likely, but it makes the job of your 'world beating' track and trace system much, much harder, and shortens the likely timescale to any second wave.
The end goal is a vaccine, we hope, and the aim is to reduce mortality to that point in time. If you can somehow delay or suppress your second (or third) wave past the point at which a vaccine is available, that saves thousands of lives.
Even if there were no vaccine, the results from the RECOVERY trial are a reminder that, month by month, doctors around the world will get better at saving people from this virus. Every month you delay your next wave of infections, there is a better chance that treatment protocols for the most seriously ill patients will have improved, and their chance of survival/good quality of life post covid increases too.
TL:DR Pushing back cases in time saves lives.
I’m the stupid one for not working out how to get on that gravy train!
No, you’re not stupid, there are very strong links between those earning the money and the no10 team.
No, you’re not stupid, there are very strong links between those
earninggetting the money and the no10 team.
FTFY
So in the best case scenario, we need to live under permanent semi-lockdown and with all movements being tracked and recorded, and under the risk of being told at any time we need to go back to our houses and isolate for days/weeks. We would need the track and tracing to be incredibly thorough, and for our country borders to be completely locked down, for this to work. The system would be very vulnerable to any slip in track/trace, or infected people entering the country. I just don't see how this is feasible in the real world, and how life could continue and support an economy that can pay for that.
How ‘severe’ all those measures need to be is dependent on ‘us’ getting infection levels low enough before opening up lock down… and the ‘completeness’ of that lock down determines how long that takes. ‘We’ failed in the UK on both counts… no real lock down and opening up while infection still ongoing.
Oh, if you can’t imagine how keeping infection under control once low enough is ‘feasible’, try looking at other countries. And, again, it won’t be good for the economy not to do so.
So in the best case scenario, we need to live under permanent semi-lockdown and with all movements being tracked and recorded, and under the risk of being told at any time we need to go back to our houses and isolate for days/weeks.
Not really, if case numbers are low enough, then lockdown can be lifted completely, with the proviso that new cases are aggressively tracked and their contacts isolated. This is what has happened in NZ.
Obviously we're not in that happy place, so some form of lockdown restrictions will carry on, with each easing accompanied by lots of praying and crossed fingers that numbers stay where they are. Losing your summer fun/livelihood/family life is the price you pay for a poor government which wasn't prepared to make the difficult decisions because it didn't want to be unpopular with a vocal minority.
What other countries have shown is that to be successful track and trace needs to be done on a local level and done through trusted, open, accountable local agencies like healthcare trusts, allowing action to be rapidly tailored to the needs of specific places
So the government have done their usual and ignored all evidence from experts (who needs those eh?) as it doesn't fit their chosen ideology.
Instead, they've done what they always do - ignored all evidence and advice and done the polarv opposite. They've centralised the decision-making in Whitehall, kept it opaque and secretive, and farmed the implication out to the same private sector companies (Capita, Serco) that have repeatedly demonstrated that they are incapable of delivering anything like this and certainly can't be trusted with personal information.
And does anyone actually believe that a fully functioning app will ever be operational?
They're a bunch of ****ing clowns!! Utter incompetents!! But, as per bloody usual, I bet their mates are making an absolute killing out of this shambles. Both metaphorically and literally
Oh, if you can’t imagine how keeping infection under control once low enough is ‘feasible’, try looking at other countries.
Are there any countries in particular that are claiming success due to a hard, short lockdown followed by track and trace? Personally, I think that the story there is far from over.
Not really, if case numbers are low enough, then lockdown can be lifted completely, with the proviso that new cases are aggressively tracked and their contacts isolated. This is what has happened in NZ.
We will see how that develops, but it's already not looking great again for NZ. I think it will only demonstrate that trying to maintain anything like a normal economy, complete with large numbers of international arrivals, is not compatible with a "complete elimination" policy.
In other words it reads to me like the Health Editor and Health correspondent didn’t actually read it but spent some time instead making a infographic.
Exactly. The analysis will have been done by the IDMC. It’s the headline look to see if it ethical to stop or continue. That analysis has shown a robust effect, hence it is time to stop. The other planned analyses will follow (including subgroups). Interestingly (for me) they have not used the standard 7-point categorical scale that everyone else is using. One measure is fraction of patients making a one and two point improvement. 7 is death btw. Remdesivir gave a modest one point improvement in mild patients.
but it’s already not looking great again for NZ
You have a strange idea about what looks great.
How many new deaths in NZ in the last week? How many here?
How about schools, shops, cafes etc… how many are back in proper use there? How about here?
You have a strange idea about what looks great.
Not really. Sure, living with very little restrictions does look great, but they've already had two known infected people fly in, leave quarantine early for "compassionate reasons", and make known contact with 320 other people. Other people are known to have left quarantine early, and just today someone was caught flying while infected with Covid-19. At the rate that they are having to make contact with people who have potentially been infected, pretty much everyone will be back under lockdown in NZ very soon. My point is only that this approach is not practical, unless you completely rigidly enforce your country's borders.
To explain a bit re Enzed - there is some cultural stuff around funerals that makes it harder to limit them.
Queen Cindy's face at the briefing about the latest developments almost made me cry!