Omicron news

[elbows]

The usual story then, same sort of stuff as Whitty mentioned recently, ie the last sentence from this bit of the abstract of that study:

Early analyses suggest a reduced risk of hospitalisation among SGTF-infected individuals when compared to non-SGTF infected individuals in the same time period, and a reduced risk of severe disease when compared to earlier Delta-infected individuals. Some of this reduction is likely a result of high population immunity.

I think its been pretty clear from the start of Omicron that it does not lead us 'all the way back to square one' in this pandemic. But beyond that, I mostly have to wait for the full UK picture of how much its eroded our protection to emerge. The number of infections we allow in this country will certainly provide huge amounts of data in time.

 

[Supine]

A bit of UKHSA leakage

 

[elbows]

This is why I hate the ignore function, I already mentioned that but my pandemic rants are not to Supines taste and so they probably didnt see my post on the subject.

 

[littlebabyjesus]

A bit of UKHSA leakage

One other piece of potentially good news not included there is that London appears to be following a Gauteng-like curve. Cases shot up rapidly but look like they may have peaked rapidly as well, as happened in Gauteng. Tentatively, London may have hit a peak around the middle of last week. That's led to increased hospitalisations this week, but those may be nearing or at their peak now as well.

So when they say that hospitalisations may shoot up, there is also some reason to suspect that they won't, that we will continue to follow a path that is comparable to that in South Africa.

This is Gauteng's omicron curve so far:



And this is London's (grey bars are incomplete, but none of the grey bars is likely to exceed the peak bar that represents last Wednesday):



One obvious difference is that London didn't start from near-zero like Gauteng, but Gauteng peaked within a few days of omicron really exploding, and London may be doing the same thing.

 

[magneze]

 

[Sasaferrato]

But do a test anyway.

Just FYI - there has been a bit of norovirus about.

Fuck. The year that keeps on giving.

 

[ska invita]

One other piece of potentially good news not included there is that London appears to be following a Gauteng-like curve. Cases shot up rapidly but look like they may have peaked rapidly as well, as happened in Gauteng. Tentatively, London may have hit a peak around the middle of last week.

All social engagements got binned for me two weeks ago as trying to get over the line so as to see people at xmas without being positive....needs factoring in, rather than Omi has run out of people to infect. Anecdotally true for lots of people.

 

[klang]

Yes, I know quite a few people who have been laying low the last couple of weeks.

 

[elbows]

One other piece of potentially good news not included there is that London appears to be following a Gauteng-like curve. Cases shot up rapidly but look like they may have peaked rapidly as well, as happened in Gauteng. Tentatively, London may have hit a peak around the middle of last week. That's led to increased hospitalisations this week, but those may be nearing or at their peak now as well.

So when they say that hospitalisations may shoot up, there is also some reason to suspect that they won't, that we will continue to follow a path that is comparable to that in South Africa.

As of yesterdays published data, hospital admissions figures only went up to Sunday, so that needs to be taken into account when estimating timescales of hospital peaks in the data.

Also a path similar to that seen in South Africa will involve hospitalisations shooting up. I havent done much with South Africas data because the format befuddles me at times, have you attempted any exercises to see what a comparable number might be here if the hospitalisations trajectory and peak proportions turn out to be similar?

You know that I consider your instincts and analysis of the first few UK waves were far wide of the mark. However with every passing wave and with every expansion of the vaccine programme, in theory the reality should become closer to your expectations than it was at the start. Mostly because of the evolving immunity picture, which gradually brings reality closer to what you incorrectly hoped was the case after the first wave. Uncertainties about the extent to which Omicron can evade prior immunity of various forms, most crucially protection against severe disease, continues to make it very hard for me to judge the chances of you being right this time. We are at least at the phase of the pandemic where I wont just automatically shit all over your views with a high degree of confidence at this particular moment.

 

[littlebabyjesus]

All social engagements got binned for me two weeks ago as trying to get over the line so as to see people at xmas without being positive....needs factoring in, rather than Omi has run out of people to infect. Anecdotally true for lots of people.

Yep. True of a lot of people. Wandering through London the last couple of days, it has been deserted. Pubs still open but empty.

Whatever the reason for it, the rapid rise does appear to have halted, although we do still need this week's numbers to be sure of that. By far the biggest rises in infection in London came in the 20-40 age group, though, which is the age group least likely to drastically alter its behaviour.

 

[elbows]

When todays data comes out I will provide some graphs for the London region showing cases by age. Especially as the timing can vary in different groups, the hospitalisation prospects vary by age group, and one of the large differences between us and South Africa is the number of people in older age groups.

In the delta wave the different rates and timing between age groups meant that we could not rely on the previous simple approach of expecting hospitalisation peaks to follow x days after peak in overall case numbers. Because the July peak was sponsored by younger people, and things actually got worse in older age groups later on, which was also reflected in subsequent hospitalisation peaks much later. So I have to apply lessons from data in that wave to our current view of the Omicron wave.

Our attempts to tell simple stories via hospitalisation data are also going to be complicated by hospital infections which will affect the headline numbers.

 

[l'Otters]

This is why I hate the ignore function, I already mentioned that but my pandemic rants are not to Supines taste and so they probably didnt see my post on the subject.

Fwiw I find getting both/all versions offered here of how info is framed useful - often takes a few different methods of explaining the same thing for something to go in. (for me anyway.)

 

[Monkeygrinder's Organ]

Yep. True of a lot of people. Wandering through London the last couple of days, it has been deserted. Pubs still open but empty.

Whatever the reason for it, the rapid rise does appear to have halted, although we do still need this week's numbers to be sure of that. By far the biggest rises in infection in London came in the 20-40 age group, though, which is the age group least likely to drastically alter its behaviour.

I think the particular timing of it - ie that getting it now means Christmas in isolation - probably has a big additional effect in terms of behaviour to what would normally be the case

 

[gentlegreen]

.

 

[elbows]

That's led to increased hospitalisations this week, but those may be nearing or at their peak now as well.

I forgot to explicitly check whether that means you really think the London daily hospital admissions figures will not exceed levels similar to the amount shown in yesterdays published data?

Even if we get a relatively happy scenario that doesnt sound plausible to me.

(Recent figures which now include an additional days worth since your comments, which I'm sticking on the UK thread because there was a story that government are going to make some decisions based on whether the figure exceeds 400 by the end of the week #44,796 )

 

[magneze]

I think the particular timing of it - ie that getting it now means Christmas in isolation - probably has a big additional effect in terms of behaviour to what would normally be the case

I think we're almost too close now. Getting it now - you might not test positive until after Christmas.

 

[teuchter]

Do we have a clue from south Africa what the length of the cases--admissions lag might be for Omicron might be, and is it likely to be shorter than we have been used to so far?

 

[littlebabyjesus]

I forgot to explicitly check whether that means you really think the London daily hospital admissions figures will not exceed levels similar to the amount shown in yesterdays published data?

Even if we get a relatively happy scenario that doesnt sound plausible to me.

(Recent figures which now include an additional days worth since your comments, which I'm sticking on the UK thread because there was a story that government are going to make some decisions based on whether the figure exceeds 400 by the end of the week #44,796 )

If cases really did peak (or at least plateau) last week, we are likely to see a peak (or plateau) in hospital admissions this week if previous patterns are anything to go by. Nothing is certain of course, but that would also follow roughly what happened in South Africa. Monday was 300, exactly a week after the first day that new cases went over 20,000. Will that storm past 400 this week? Perhaps. Or perhaps it will sneak past 400, or it will level out around that level. Will we see much shorter stays in hospital like South Africa? Dunno, but as with all this stuff, I don't see a compelling reason why not.

400 is a pretty arbitrary number to choose, though, especially if the numbers show it probably won't go much higher than that. Also, I don't know how many of those 300 on Monday tested positive on admission for something else. In South Africa, a high proportion of patients on covid wards a couple of weeks ago were people being treated for something separate, with either mild symptoms or no symptoms.

But I was actually quite clear I think in what I said - that if cases peaked last week then admissions were likely to peak this week. I didn't give a day, and I didn't give a specific level. I'm watching with interest to see what happens this week following the record high level of cases last week.

 

[zahir]

A new record today.

UK Covid live: research into Omicron severity in England finds ‘moderate reduction’ in hospitalisation risk compared with Delta – as it happened

Latest updates: researchers at Imperial College London stress that reduction in severity must be balanced with larger risk of infection

www.theguardian.com

 

[elbows]

If cases really did peak (or at least plateau) last week, we are likely to see a peak (or plateau) in hospital admissions this week if previous patterns are anything to go by. Nothing is certain of course, but that would also follow roughly what happened in South Africa. Monday was 300, exactly a week after the first day that new cases went over 20,000. Will that storm past 400 this week? Perhaps. Or perhaps it will sneak past 400, or it will level out around that level. Will we see much shorter stays in hospital like South Africa? Dunno, but as with all this stuff, I don't see a compelling reason why not.

400 is a pretty arbitrary number to choose, though, especially if the numbers show it probably won't go much higher than that. Also, I don't know how many of those 300 on Monday tested positive on admission for something else. In South Africa, a high proportion of patients on covid wards a couple of weeks ago were people being treated for something separate, with either mild symptoms or no symptoms.

But I was actually quite clear I think in what I said - that if cases peaked last week then admissions were likely to peak this week. I didn't give a day, and I didn't give a specific level. I'm watching with interest to see what happens this week following the record high level of cases last week.

Thanks for the answer.

I too heard about the proportion of cases in South Africa thing some time back, but I fear it wasnt accurate. I'm pretty sure I saw someone visualising such data from South Africa more recently which showed the proportion of those cases that were in hospital for 'non-covid' reasons had actually remained quite consistent with the past, and still quite a low proportion (perhaps around 10%). But I dont know if I'll be able to find what I saw. I'll try and will post it here if I find it.

I have reasons to doubt that London admissions will remain at recent sort of levels, my brain still tells me to prepare for large jumps. But another part of my brain is pumping much more uncertainty into the mix than I had with the first few waves. This uncertainty was also the case for me with the Delta wave, but this Omicron wave features larger case numbers so my brain is a bit of a mess right now, a big tug of war between fearing a very bad wave and retaining some hope that worst fears wont come close to being met.

 

[teuchter]

A new record today.

UK Covid live: research into Omicron severity in England finds ‘moderate reduction’ in hospitalisation risk compared with Delta – as it happened

Latest updates: researchers at Imperial College London stress that reduction in severity must be balanced with larger risk of infection

www.theguardian.com

It will be very surprising if we don't see "new records" most days for the next couple of weeks - it's what is predicted and it's going to get a bit boring if it keeps being presented as news.

 

[elbows]

And no I dont know why the government picked 400 as a measures threshold, if that story in the news is even true.

 

[elbows]

It will be very surprising if we don't see "new records" most days for the next couple of weeks - it's what is predicted and it's going to get a bit boring if it keeps being presented as news.

The modelling cover a large number of scenarios and they always point out that it is not a prediction.

 

[elbows]

This points in the same direction as other early studies:

Omicron wave appears milder, but concern remains

The risk of needing hospital treatment could be 30 - 70% lower with Omicron than other variants.

www.bbc.co.uk

I havent done the maths on case numbers to guesstimate what the implications still are. And I'll really have to wait till more people in older age groups catch it and have time to show up in hospital numbers before I get too carried away.

 

[weltweit]

If it is milder then how lucky were we? Imagine if it had been as infectious as it is and more deadly than Delta, we would be up shit creek.

 

[elbows]

Can still be up shit creek via transmissibility alone.

Because if it only generates a third of the hospitalisations of other variants, but you have three times more cases, you get the same number of hospitalisations.

I had a quick go at using the ridilously oversimplified maths to apply that to Londons peak in cases (so far) & hospitalisations already seen this time compared to peak delta wave figures for that region. I think it already fails to work that simply, but maybe I did the maths wrong. But I wouldnt expect it to work out that simply because of age-dependant hospitalisation rates. Would have to drill down into cases by age to even do the simple maths version more correctly.

 

[bluescreen]

Do the data distinguish between people hospitalised because of Covid and those admitted for other reasons who on routine testing are shown to happen to have it as well? Eg, admitted for broken leg?

 

[kabbes]

Can still be up shit creek via transmissibility alone.

Because if it only generates a third of the hospitalisations of other variants, but you have three times more cases, you get the same number of hospitalisations.

I had a quick go at using the ridilously oversimplified maths to apply that to Londons peak in cases (so far) & hospitalisations already seen this time compared to peak delta wave figures for that region. I think it already fails to work that simply, but maybe I did the maths wrong. But I wouldnt expect it to work out that simply because of age-dependant hospitalisation rates. Would have to drill down into cases by age to even do the simple maths version more correctly.

I suspect there is a complication that arises due to the fact that cases are not so simply multiplied. If the transmissibility is x100 then you don't have cases x100, because that would be more than 100% of the population. On the other hand, if transmissibility is x1.01, it's probably reasonable to assume cases x1.01 as well. Somewhere between those extremes, you have omicron. Maybe 5 times as transmissible as delta but what that translates to as a multiple of cases depends on how much delta there would have been in the population anyway, and what peak prevalence can theoretically be based on the epidemiological dynamics.

 

[elbows]

Do the data distinguish between people hospitalised because of Covid and those admitted for other reasons who on routine testing are shown to happen to have it as well? Eg, admitted for broken leg?

No and I often go on about that, official daily data is for 'hospitalisations/diagnoses'. and includes people who went in for other reasons and thsoe who caught it in hospital.

There is seperate data about reason for hospitalisation that can be used to get a rough idea about that though. It doesnt come out daily but I will talk about it and its implications at the next opportunity.

Also we cant just remove all those 'other reasons' cases from the picture because unfortunately a fair chunk of the death burden in previous waves was caused by those sorts of cases - people that were already rather sick or frail for whom covid on top was fatal. There have been a really large number of hospital-acquired infections during this pandemic with quite major consequences . Vaccines should have reduced the consequences but not down to negligible levels.

 

[elbows]

I suspect there is a complication that arises due to the fact that cases are not so simply multiplied. If the transmissibility is x100 then you don't have cases x100, because that would be more than 100% of the population. On the other hand, if transmissibility is x1.01, it's probably reasonable to assume cases x1.01 as well. Somewhere between those extremes, you have omicron. Maybe 5 times as transmissible as delta but what that translates to as a multiple of cases depends on how much delta there would have been in the population anyway, and what peak prevalence can theoretically be based on the epidemiological dynamics.

I didnt do any such maths to the case numbers. I took the actual recorded case numbers, and checked how many more times that amount has already been reached in London at its highest Omicron point seen in the data so far, comapred to the Delta wave peak case numbers. Then I took peak hospital rates from Delta wave for London and multiplied them by the same factor. Then I divided them to account for the decrease mentioned in the aforementioned study. Number came out significantly lower than a number already seen in the London hospital data in this Omicron wave.

Reasons why I would not have expected this crude attempt to match the actual figures at all includes the need to take into account age of cases, hospital infections, and those who went to hospital for other reasons but were already infected before admission as discussed in previous post. I only did this exercise briefly, and without discussing the actual figures, mostly so I could then warn others not to base their expectations on this simplified method or how many hospital admissions there were at peak in London in the delta wave.

Other problems would include the difference between the detected number of positive cases and the actual case numbers. Reasons for that including attitudes to testing, availability of testing, and the fact the published case numbers dont count reinfections (although they do in Wales).