2021 covid trends in highly vaccinated new england

despite 80%+ of over 12's fully vaccinated, new england looks worse than last year

covid is a highly seasonal virus and these seasons vary by region. the classic “flu season” period in the north gets rolling in late october and peaks in december and january. covid seems to follow this.

similarly, we see the peak in souths states tend to occur in august. this makes comparing southern states to norther states at any given point in time a fraught process as comparing one in peak season to another that is in low season can lead to conclusions that are generally useless.

this issue finds a confluence with a variation in vaccination rates. many southern states are less vaccinated that their northern counterparts and the southern peak season has been used to claim “you need more vaccination” but as we saw HERE there is basically no correlation between state vaccination rates and the current trajectories of cases or hospitalizations. there is good correlation to latitude.

so perhaps another approach to assessing vaccine efficacy is in order: we can use high vaccine states as their own controls.

• there have been many more cases of covid and considerable additional amounts of acquired immunity have been generated since this time last year.

• there have also been meaningful depletions of high risk groups that will lessen overall

• delta variant has become predominant. this variant is somewhat more contagious and has ~2/3 lower CFR than prior variants (such as alpha) and a roughly 45% lower hospitalization rate.

• (though in the most recent UK VoC report, secondary attack rate from delta looks near identical to alpha p56 so whether delta is still meaningfully more contagious is a more open issue that many seem to suppose and may no longer be the case. (i need to work on this)

• this effect does not appear to be a result of vaccination. delta CFR has been rising as more people have been vaccinated.

• and vaccination, which was 0 at this time last year, now exceeds 80% of all people over the age of 12 in new england and in all states but one.

so, we should expect, even in a world without a covid vaccine, an ambiguous result on cases as possibly higher contagion works in opposition to more acquired immunity, but one should expect dramatically lower hospitalizations and deaths.

unfortunately, this is not what we are seeing.

i took a snapshot (sept 20th) and compared it to last year using 7 day moving averages to limit data artifacts from day of week reporting and capture a wider swathe of data. i chose this date because it was the most recent one for which all the data was available. data is from NYT.

i adjusted the cases level for the testing level in the last column. i believe this to be a much better metric than raw cases and that it is not fair to compare 2021 to 2020 on raw cases because the testing levels are generally around twice as high and you mistake sample rate for disease prevalence and overstate 2021 disease relative to 2020.

i have used only “fully vaccinated” for this discussion to avoid issues from the “worry window” of negative vaccine efficacy early on.

this figure is the actual % increase, so a 100% number implies that last year was X and this year was 2X. obviously, these hikes are dramatic.

i stacked the states from north to south because i think something interesting emerges when we do: the further north you go, the worse the acceleration seems to get. (vermont had no deaths in september last year (and does this year) so no ratio could be computed)

i added NY because connecticut seemed to be bucking the latitude trend, but given the cross pollination of fairfield county and NYC, i wanted to see if that might provide an explanation. it seems like it might.

but what we see here is unsettling. new england is still at low absolute rates of death and hospitalization, but they are rising rapidy and are tracking at multiples (sometimes extreme multiples) of last year and cases are exploding even when adjusted for sample rate, which augurs poorly for future hospital and deaths data.

this also seems to imply that as the cold settles further south, they may need to brace for something more like what’s happening in maine and vermont. (this could also be that maine and VT had milder exposures previously and are now playing catchup. MA, RI, CT have already been pretty well pummeled. this may be protective)

this result is deeply counter to expectation even without vaccines. with 80% of the over 12’s vaccinated (and under 12’s amounting to an insignificant % of hospitalizations and essentially zero deaths) if vaccines were working to stop spread or severe outcomes, this becomes a difficult claim to explain.

pretty much everyone (even the CDC) has given up on claiming these vaccines are sterilizing. they do not stop spread and this shows up clearly in the UK data where the vaccinated are showing meaningfully negative VE in terms of “cases”. (HERE)

there is very plausible case to be made that these vaccines are actually increasing spread and may well be selecting for hotter, more dangerous virus. leaky vaccines are a known problem and can pose large dangers. and these are definitely leaky vaccines.

while is shows amplification of case risk, this UK data does seem to show VE in hospitalizations and deaths, albeit at levels more like ~50% than the 97-99% figures once claimed. but even 50% VE at an 80%+ penetration rate should be showing a dramatic overall effect.

(though the idea that it’s all “the unvaccinated” dying and clogging up hospitals is clearly refuted by this UK data. in the last 2 weeks, 54% of emergency care, 59% of hospitalizations overnight, and 76% of deaths were vaccinated.)

the US is playing some very aggressive games with definitions and with failure to report status on the vaccinated. these claims of “our whole ICU is unvaxxed” as about as credible as those of “we called everywhere and the closest ICU bed to idaho was connecticut when idaho was not even at 90% ICU capacity” or the crazy claims about floods of ivermectin deaths. these are simply false and the US data on vaccination status has become a disaster. it’s simply not trustworthy. let’s see if we can keep the disinformation and manipulation out of the data as much as possible.

but all in all, what were seeing here looks worrying. all this vaccination is not slowing covid down at this point. it’s very early in the season to be making big calls on the northeast, but this year looks seriously aberrant and blaming this on delta given the much lower hospitalization and death rates (and possibly no higher spread rate) therefrom seems fraught.

last year hospitalizations in maine did not start to even tick up until the last week of october. this year, with 84% of over 12’s fully vaxxed, they are already nearing the highs from the peak last winter.

vermont shows the same pattern and will be a very interesting bell-weather to follow. they have had very low covid to date and 88% of over 12’s vaxxed. so it’s an even purer test of vaccine efficacy.

cases, even on a test adjusted basis are ALREADY higher than the peak season winter peak vs not even really ticking up yet this time last year.

new hampshire is not as bad, but shows similar pattern.

MA and RI, while worse than last year, do not look anything like as bad. they have also had A LOT more covid previously than their northern neighbors.

MA, RI, CT were among the worst states in the US for deaths per million (cases are hard to comp due to testing level variance)

NH, ME, VT were among the best. (data)

this may be inverting. everybody gets a turn eventually and it looks like holding out until high vaxx rates are achieved is not helping so far…

they are also still warmer. fall comes early in maine and vermont.

it will be informative to see how MA, RI, CT fare through october vs the higher latitude lower initial immunity north.

but all in all, every single new enlgand state has far more cases and hospitalizations than this time last year. the margins are not slim.

all but 2 have much higher deaths and the two that do not are already numbers 7 and 8 in the US for per capita deaths, so cohort depletion and developed immunity are likely suspects.

this is not what one would look for or expect from 80%+ overall vaccination rates among people over 12.

this does not disprove vaccine efficacy against severe cases and outcomes. cases are spiking enormously (far more than the variance in delta contagion would predict) and that pushes all metrics higher in absolute values even if CFR and HFR drop.

(i want to do some work on this with a better dataset)

but it IS what we’d predict from the “vaccinated superspread hypothesis” which is looking more and more worrying the more data i look at.

even if vaccines drop hospitalization rates among the vaccinated by 50%, if they also increase disease spread by more than 100% (and this data is worrying in that regard) the actual number of vaccinated people hitting the hospitals will rise.

i increasingly fear that vaccine driven spread is being wrongly allocated to delta R0.

if, this hypothesis holds and the vaccinated constitute high VL asymptomatic superspread bombs, then this campaign may have kicked the whole pandemic up into a higher valence. it will also be brutal on the unvaccinated who, ironically, would need protection FROM the vaccinated which, in a double irony, will make vaccines look more effective because they intensified spread.

the confluence of this with the tendency of leaky vaccines to select for deadlier strains could be especially pernicious. CFR on delta has been rising in the UK. (best data racking i know of, though if others have good sets on this, would love to see them)

that’s strongly contrary to evolutionary expectation.

we’re in some weird, uncharted waters here and working with data of poor quality and limited reach, but such piece of this mosaic as can be pinned down are cause for concern on what’s coming this winter.

will keep gathering data and trying to coax some signal from it.