Glasto’s variant “mosh pit”

Yes, Glastonbury won’t be doing more than live streaming this year, but many others will have social distant “mosh pits”, whatever the hell that is going to look like. What matters, is that without us truly understanding the unnatural immunity balance we find outselves in, thanks to the vaccine program, we might end up making the same naive fail as this piece by Richard Lyon.

To be clear, what’s unnatural about it is normally flu will spread up from the young and kill the old, with the young hitting herd immunity ahead of the old. However, vaccines mean we’ve bypassed the young, leaving them on about 25% immunity, and driven those 50+ to 70%.

Meaning vaccine hesitancy in the young will lead to a silent epidemic over the summer, that kills few, but runs high variant risk. This will push the HIT in the old slightly higher as it will provide far more seeding points into these older groups than epidemiology would normally expect.

I cite Richard as he decided to quote my latest effort to assess the correlation between infections implied from ONS by occurence deaths, and Google mobility data, without seek permission, and decided to totally misquote me.

I couldn’t just let it lie, as there’s a 0.6 to 0.82 rsq in the correlation, high.

I don’t support lockdown, I merely seek to justify to myself, and to others, that the sacrifice being made to honour our social contract, is not killing people in ways we don’t know, and that it is working to genuinely save more ives.

If it were not, I’d be shouting as loudly as Richard.

In doing what he does, you’re either:

• a sociopath, with no interest in that contract
• a sadist, who enjoys making people feel like they’re wasting their time when they aren’t
• or, you’re numerically incapable of understanding why what you’re spouting is horse manure.

His main observation is that deaths are invariant to mobility, is that true?

Now, the part of the plot which show this is entirely wrong is the pink dotted line along the top, which looks at the gradient, or rate of change of deaths. Here with the same data, this time red, with the smoothed version in pink, is a cut down version showing the data in the correlation.

Initially very high, but as mobility collapses into the first peak, this dives below 100% or 1, meaning each day has a lower number of deaths that the previous. This rate of decline, as shown by the green (orange in the main plot) tangent line remains constant through Apr, as does mobility.

It then begins to jitter up, at the same time as mobility is lifting. Bear in mind that mobility after the first peak is very different to that going into it. Before there were no social distancing measures in work places, so it likely took reducing mobility more to get cases to drop, after as we lifted measures, with everyone playing more of a part, the suppressed behaviour patterns meant we could get back far more movement with far less spread.

It seems to take getting to 40–45% commuting to by mid Jul to get this to zero, then permanently +ve in early Aug, getting very +ve by mid-late Aug.

This is part of why pre-variant I thought we’d cracked managing this with tiers and local measures. Like @MPIainDS, I believed they were working, and if anything I wanted us to move more towards German’s even more localised and targeted measures, the entire aim being to suppress spread more without needing to put rural Cornwall in lockdown. Seriously, has that really been of any benefit at any point?

Richard’s claim it is constant, and invariant to mobility, a plain lie.

As to his charge Christmas mobility does not make what happened then clear? It is the only thing that does. It was clearly a superspreader event, with mobility and infections peaking by Dec28, beyond mobility collapses showing people really hunkered down, complied with Tier4 measures (with schools closed). The 3rd LD then just kept those measures in place by not reopening schools, and stopping physical returns to work after Christmas breaks.

It’s entirely disingenuous to use the fact that the 3rd LD was a “keep measures in place” lockdown, not one that actively changed things.

Nothing Richard says makes any sense.

His understanding of seasonality is also very flawed.

Seasonality starts after a few seasons of a new virus, when we finally hit a lower summer herd immunity threshold for that virus, burning it out. Ironically the only comparable period in history he can find is the global period when HongKong flu first hit us. By a few years in winter spikes remained comparable to 2020 C19 with NPI suppression, but this is not normal influenza, and his curve for this year seems to be lie as it looks nothing like global cases or deaths from worldometers.

As to C19, from best R estimates it has a natural summer HIT of likely 60%, but last summer NPIs had suppressed spread to 10%. Sure, summer behaviour patterns on top of NPIs helped work against spread, and it getting colder into Sept/Oct did help drive up cases again, but the 2nd wave only took us to 25% spread, so his theory that seasonality is doing all the heavy lifting here is for the birds. NPIs were the main throttle.

So is there any evidence of “mosh pit” risk? Vaccines in the old not the young?

Well Sweden, that year long misused friend, paints our future. They are at 20% 1st dose, 7% full. Pretty much in those >50, or vulnerable, and having been afforded a more relaxed approach by having 1/3 our urban density, as I was pointing out last Aug in this piece, and it is clear their higher mixing since the start of the year has pushed up cases and ICU numbers. As they are now as high as they have ever been, yet (even factoring in their month delay in death reporting) they just aren’t seeing a spike up in deaths.

If we look at their case data by age, we can see cases are collapsing in the older age groups. If we use serology IFRs to imply an IFR from this blend of cases by age, we see that the avg to Jan was 0.68% likely to die, but this has now dropped to 0.26%, as there simply aren’t the cases in the old.

Howeve, this is still not the whole likely story, as these age banded serology IFRs will now be overestimating risk, as each of these age groups have had the higher risk vulnerable member vaccinated, meaning all thee IFRs by age are likely now lower, so the age wgted one will be lower too.

It could eaily be 1/10 what it was.

Sure, deaths are far lower. Vaccines are protecting those far more likely to die, but, this does not change the reality of spread, and makes the argument that those unvaccinated <50 are anywhere near their summer HIT without NPIs, just plain daft.

I don’t say this to argue for further lockdowns. We need to lift NPIs, I simply say it so that we all understand the risks with current policy. Without vaccinating those <50 spread is very likely to be larger than it has been so far, and that will carry with it home grown variant risk, one that might beat vaccines in our vulnerable cohort this coming winter.

We all need to get back to the “mosh pit”, but let’s know the risks, and keep getting tested, it won’t help you, but if they can catch a variant before it spreads, we’ll be able to keep the mosh pits open…