are covid vaccines the superspread vector for omicron?
it's really starting to look that way
i’ve spoken quite a lot about why leaky vaccines drive viral evolution to select for vaccine advantage. rolling out non-sterilizing vaccines is is terrible idea. it’s like only taking half your course of antibiotics and then wondering why MRSA is suddenly everywhere. it’s just a selector for resistance and advantage. the basic mechanism is really very simple: antigens that were recessive become dominant in order to evade the fixated immune response generated by an inoculant that did not work to stop colonization, carriage, and contagion.
this is the simple, predictable, and inevitable outcome of herd level antigenic fixation whereby most people are all locked into the same increasingly ineffective immune response and fail to generate new responses when faced with novel pathogens. this gets called OAS/hoskins effect.
it’s also how you get a throwback variant like omicron which did not descend from delta but looks to be a second serotype whose last common ancestor with D was pre-alpha. it was a going nowhere failed mutation that lacked evolutionary fitness. but then the world changed and vaccines selected for omicron. it’s just simple evo pressure that works like this.
and it looks to be intensifying rapidly. since the emergence of omicron and especially of some of its later sub-variants, particularly the BA series, there has been dramatic change in viral behavior.
we saw the canary in the covid mine in the UK where risk rates were blowing out and even as they switched to “3 dose boosted” to measure relative effect, the fact that risk rates were not only higher in the vaxxed than unvaxxed, but rising fast over time became unavoidable.
covid was already vaccine advantaged, but was rapidly becoming more and more so. full data HERE.
notoriously, this led the UK to discontinue reporting of this series. it appeared they suddenly did not want to talk about this anymore. but it’s cropping up everywhere.
it can now be seen in UK all cause deaths data.
we also saw it in some US hospitalization data.
and in the data from israel.
this set me to wondering where else we might test this theory that “more injection = more infection,” especially with new variants.
the timeframe for investigation was easy to pin down so we have the makings of a testable hypothesis:
if these variants are more vaxx enabled, we’ll see more rise in prevalence in places of high vaccination.
if this prevalence is swamping the lower virulence of the variant, we will also see a rise in severe outcomes (because 4X cases at 1/2 severity is still 2X the number of severe outcomes)
so here we go:
i reached out to longtime gatopal™ and noted datahawk ben to get a look at some US data that he is far better at automating and slicing than i.
the results were extremely provocative.
first, we plotted the relationship between vaccination rates and covid case rates in 2021 from spring to mid summer:
we can see a mild association between vaxx and cases, but it’s nothing that screams “crisis.” r2 = 0.079
but look what happens in 2022 as vaxx rates are higher and the new omicron variants are dominant:
now the relationship between more injection more infection becomes pronounced. R2 is now quite high for such a noisy series at 0.484.
it’s the same timespan in the same states, so this should cause seasonal signals to mostly cancel out.
i think we’re seeing a strong vaccine signal here and the highly vaxxed states are tallying on the order of twice the total covid counts based on about a 25 percentage point difference in vaccination rate.
that’s a massive signal and starts to align ominously with some of the UK data. if 25% of your people get 4-5X the infection rate, that roughly doubles the overall rate of infection, which is just what happened. this is more than a bit provocative.
cases are a tough metric to work with for a great many reasons and so i then sought to find further validation of what was going on. this led me to the data from healthdata.gov. the data is, to put it kindly, disordered and dumping it into csv and thereby to excel and doing this by hand is time consuming (and i’d love to find a way to automate it if someone has the time/chops) so i just pulled one state of particular interest because it’s big, has enough data, has been widely watched, and i know it to be representative of what’s going on in many others: new york.
i chose hospitalization rates among the old as a metric because it’s a severity measure and the vaxx rate there is now near total in NY.
it’s obvious just from the NYT chart that this year does not look like last year. things were going well, but then in march/april, the trend reversed and we got a dramatic departure from past seasonal signals.
the chart above uses the healthdata.gov data, so i pulled that data and ran the series (in absolute numbers, not per 100k)
we can see “this was going better and then it suddenly wasn’t” very clearly if we overlay the years.
and it’s even more pronounced in 80+
to make this easier to see i calculated the year on year change in hospital admissions (and applied a 7 day moving avg for smoothing).
something reversed in late march/early april and by late april it was really taking off.
that’s right when the BA sub variants hit. note that these are also the same variants that caused a near step function rise in the UK as can be seen in the chart above.
it was worse in 80+
this is, of course, exactly what one would predict if the driver is antigenic fixation: new variants make the effect pronounced and it expresses most highly in those most dependent on antibodies because their generalized immune systems are weaker.
(a caveat: some seasonal variability (like early summers) may account for some of this yoy rise just based on comps to very low prior season so as we start to compare vs summer peak from last year, the % will likely (and hopefully) drop. but even at 200-400% elevation instead of 700-1400%, that’s still very, very bad.)
as one final issue: look at the overlap of vaxx and hotspots within NY
this association just keeps cropping up everywhere. it’s in the german data too.
and europe as a whole:
graphic from dr claire.
i wish i had better news, but this signal is prolific.
cases in NY are ~10X what they were this time last year. no way that’s “early summer” or “sample rate.” it’s a massive change in infection profile.
covid hospital admissions are 4X a year ago and ICU doubled. deaths, while low, are roughly 3X last year same time.
overall US case count for 7/17/22 is 130k vs 32k last year, hospitalized is 41k vs 24k and deaths are 420 vs 270. (all 7 day mvg avgs)
all of this is strongly consistent with the idea that we have a much milder variant that is still producing a greater number of severe outcomes because the case count is so much higher and this case count is being driven by the vaccinated, not the unvaxxed. their risk ratio for catching covid looks to be 3-5X higher.
this is the outcome no one wanted/everyone hoped against. alas, reality is not optional and the reality is that this was entirely predictable because this is how evolution works. it’s why we do not use leaky vaccines. the herd level antigenic imprinting such drugs produce becomes an evolutionary selector and the outcome will always be this (unless, like flu, the vaxx is so ineffective and the pathogen so variable that the imprinting is too wide of the mark to fixate).
your immune system becomes a one trick pony and the virus learns to trick the pony and spread like wildfire.
that’s nature for you.
this pattern is all over the US states and far more pronounced in the high vaxxed than the low.
how this ball bounces next is not easy to call with precision, but in general, i’d expect the omicron sub variants to become ever more vaccine advantaged, the risk ratios to stay elevated and possibly rise further, and for durable covid immunity among the vaxxed to remain elusive, especially in those with weaker general immune systems (mostly, the aged).
boosters seem to be making it worse and evidence that variant specific boosters will provide efficacy is absent. tests of such have shown them to be ineffective and this stands to reason: if the actual virus cannot overcome vaxx induced OAS it’s unlikely a vaccine can either.
i take no joy in this hypothesis gathering such compelling evidence and keep trying to falsify it because, frankly, i’d sleep better.
but the data is what the data is. so we must follow it.
please keep the ideas for and insight into these analyses coming.
will keep at this.