1. Any idea why 1940 to 1960 is so much better than other years?
2. Well, actually Cochrane's RCT reviews do tell us something. Namely that "we have no clue". But people don't absorb that info very well. Aside from that, IMO, the only flu vaccine studies that tell us anything are "regression discontinuity designs". That approach says fluvax may kill. Pasting my notes on that below.
RDD studies by their nature combat a variety of confounding found in other observational studies.
In RDD’s of vaccines, the comparison is between younger versus older populations. Since aging is not a choice, group assignment is in a sense randomized. In some contexts, some may even prefer the results of RDDs over RCTs when these two designs yield conflicting results on the grounds that RDD reflects the “real world”, whereas RCTs may not be generalizable.
Regression discontinuity study of flu vaccine finds “Turning 65 was associated with a statistically and clinically significant increase in rate of seasonal influenza vaccination. However, no evidence indicated that vaccination reduced hospitalizations or mortality among elderly persons. The estimates were precise enough to rule out results from many previous studies.”
RDD could be expected in this context to reduce the impact of confounding by contraindication, reducing it to a “self-protection bias” among those above the 65 age threshold who choose to avoid vaccines. This act does not cause a group reassignment, rendering any risk-concealing bias less significant.
Unforunately, RDD is presumably not suitable for studying vaccines and autism, as it can take years for ASD to develop, age thresholds for vaccination are quickly passed, those thresholds may not be adhered to tightly enough, and other childhood vaccines or treatments associated with the age threshold can confound.
Replies to the RDD paper raise further signals of possible harm. As pointed out, all reported point estimates point in the harmful direction, some approaching statistical significance. For example, in men the VE against all-cause mortality was −8.9% with C.I. of (−19.6% to 1.8%).
This is consistent with other RCTs which suggest that in patients without CAD vaccination may increase risk of cardiovascular events, cardiovascular mortality, and all-cause mortality.
This is fascinating. Influenza/ flu-like illnesses relabelled as Covid presumably to create panic and ensure compliance with the NPIs. Striking that the only people who were/ are able to see through the bulls..t are those of us who have jumped off the merry-go-round and can't be cancelled/ defunded/ harassed by professional bodies!
I see how the flu data got all goobered up. We know that a significant number of COVID deaths, weren't actually COVID deaths because of coding shenanigans. But, I'm skeptical that the vast majority of COVID deaths were just miscoded flu deaths.
At a 50 thousand foot view, there was an uptick in all cause mortality. Are you saying that the uptick was so small that seasonal flu variations can account for most of it? Are you saying that not very many people actually died of COVID, i.e,, less than the flu? Your COVID cobbler chart seems to show this.
Not the vast majority, just 17,942 (see part 1). The jiggery-pokery comes in artificially separating underlying cause of death from direct cause of death and forcing both of these to be unicausal, with COVID taking priority. COVID only deaths (direct cause of death) are few and far between. Prior to WHO rule 3 COVID deaths would simply be a subset of pneumonia but this relationship has been flipped on its head.
After adjustment to put pneumonia back in the driving seat there is indeed a small uptick in 2020 (the nutty bit) that aligns with COVID but this doesn't stand out as an outlier using the ARIMA technique. In sum it was respiratory business as usual (apart from the rather peculiar spring 2020 catastrophic health collapse spike that took place over 6 weeks) and I've a series of articles showing this. Erase that spike and you've nothing out of the ordinary to see, even with the cock-eyed coding that ONS are running. My wife nailed it when she said "it's the same bucket of balls but they've decided to paint them all red."
Great question. I'd like to see the ASMR first and will rummage to see if I can produce this. That being said my gut tells me there is still likely to be a ramp... and a rather awkward one at that!
Sweet sweet icing.
1. Any idea why 1940 to 1960 is so much better than other years?
2. Well, actually Cochrane's RCT reviews do tell us something. Namely that "we have no clue". But people don't absorb that info very well. Aside from that, IMO, the only flu vaccine studies that tell us anything are "regression discontinuity designs". That approach says fluvax may kill. Pasting my notes on that below.
Notes on RDDs:
https://www.bmj.com/content/349/bmj.g5293/rr/764146
https://en.wikipedia.org/wiki/Regression_discontinuity_design
RDD studies by their nature combat a variety of confounding found in other observational studies.
In RDD’s of vaccines, the comparison is between younger versus older populations. Since aging is not a choice, group assignment is in a sense randomized. In some contexts, some may even prefer the results of RDDs over RCTs when these two designs yield conflicting results on the grounds that RDD reflects the “real world”, whereas RCTs may not be generalizable.
https://pubmed.ncbi.nlm.nih.gov/32120383/
Regression discontinuity study of flu vaccine finds “Turning 65 was associated with a statistically and clinically significant increase in rate of seasonal influenza vaccination. However, no evidence indicated that vaccination reduced hospitalizations or mortality among elderly persons. The estimates were precise enough to rule out results from many previous studies.”
RDD could be expected in this context to reduce the impact of confounding by contraindication, reducing it to a “self-protection bias” among those above the 65 age threshold who choose to avoid vaccines. This act does not cause a group reassignment, rendering any risk-concealing bias less significant.
Unforunately, RDD is presumably not suitable for studying vaccines and autism, as it can take years for ASD to develop, age thresholds for vaccination are quickly passed, those thresholds may not be adhered to tightly enough, and other childhood vaccines or treatments associated with the age threshold can confound.
https://sci-hub.se/10.7326/L20-0828
Replies to the RDD paper raise further signals of possible harm. As pointed out, all reported point estimates point in the harmful direction, some approaching statistical significance. For example, in men the VE against all-cause mortality was −8.9% with C.I. of (−19.6% to 1.8%).
https://sci-hub.se/10.7326/L20-0828
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6041244/
This is consistent with other RCTs which suggest that in patients without CAD vaccination may increase risk of cardiovascular events, cardiovascular mortality, and all-cause mortality.
My word, what a magnificent and informative comment - I hope readers scoff this with relish and wolf it down!
This is fascinating. Influenza/ flu-like illnesses relabelled as Covid presumably to create panic and ensure compliance with the NPIs. Striking that the only people who were/ are able to see through the bulls..t are those of us who have jumped off the merry-go-round and can't be cancelled/ defunded/ harassed by professional bodies!
Quite. Boiling a multicausal clinical endpoint down into unicausal coding was always going to lead to nonsense.
I see how the flu data got all goobered up. We know that a significant number of COVID deaths, weren't actually COVID deaths because of coding shenanigans. But, I'm skeptical that the vast majority of COVID deaths were just miscoded flu deaths.
At a 50 thousand foot view, there was an uptick in all cause mortality. Are you saying that the uptick was so small that seasonal flu variations can account for most of it? Are you saying that not very many people actually died of COVID, i.e,, less than the flu? Your COVID cobbler chart seems to show this.
Am I missing something?
Not the vast majority, just 17,942 (see part 1). The jiggery-pokery comes in artificially separating underlying cause of death from direct cause of death and forcing both of these to be unicausal, with COVID taking priority. COVID only deaths (direct cause of death) are few and far between. Prior to WHO rule 3 COVID deaths would simply be a subset of pneumonia but this relationship has been flipped on its head.
After adjustment to put pneumonia back in the driving seat there is indeed a small uptick in 2020 (the nutty bit) that aligns with COVID but this doesn't stand out as an outlier using the ARIMA technique. In sum it was respiratory business as usual (apart from the rather peculiar spring 2020 catastrophic health collapse spike that took place over 6 weeks) and I've a series of articles showing this. Erase that spike and you've nothing out of the ordinary to see, even with the cock-eyed coding that ONS are running. My wife nailed it when she said "it's the same bucket of balls but they've decided to paint them all red."
Ee’s a boot
Once again more gold! Love the story and as a non-Covidian love the reality you present.
What do you think might be the reason why flu and pneumonia deaths rose steadily since after WWII?
Great question. I'd like to see the ASMR first and will rummage to see if I can produce this. That being said my gut tells me there is still likely to be a ramp... and a rather awkward one at that!