Coronavirus

[LHC]

I've been puzzled why the only AZ vaccine has been singled out for these suspensions, seeing that both clotting and low platelet counts have also been observed, at least in the UK, with the Pfizer vaccine (though no more than what could be expected in the general population).

It seems unreasonable to suppose that it is because of a general anti-AZ attitude in the EU that has been fostered by ill considered pronouncements by various politicians. Nevertheless I do have a sneaking suspicion that this might have contributed. (Mind you, the somewhat haphazard AZ Phase 3 trial didn't help matters.)

I think it has contributed, and now any negative stories about the AZ vaccine are being seized upon as confirmation of the negative attitudes already expressed by several European leaders. How else do you explain the Italian police's decision to seize nearly 400,000 doses of the AZ vaccine after an Italian man died after having been vaccinated.

https://www.reuters.com/article/us-health-coronavirus-italy-astrazeneca/italy-prosecutors-seize-batch-of-astrazeneca-vaccine-after-death-of-man-idUSKBN2B71QQ

[Dave2002]

“There's no proof the Oxford vaccine causes blood clots. So why are people worried?” - David Spiegelhalter

Guardian Article puts it in perepective:
https://www.theguardian.com/commenti...a-causal-links

I'm sure he's right, but he also writes this about some of the vaccine trials

Some kind of adverse events were reported by 38% of those receiving the real vaccine but, rather remarkably, 28% of those who received the dummy also reported a side-effect. This shows that the vaccination process itself causes about two-thirds of all the reported harm.

So on the one hand he is denying the use of post hoc ergo propter hoc causality arguments, then a few paragraphs down he slips up and does the same himself!

[johnb]

Are we seeing the effect of the school's reopening - the current Manx lockdown was triggered by transmission in both pre-teen and teenage students (proRata the Manx infection rate is much higher than the UK)

I am sure that will be the case but I wouldn't have expected to show up in the cases up to the w/e 10th March.

However, I've just plotted the following and I think the recent increase in CV-19 test might (?) have had an influence in the recent slowing down of the rate of decrease:



Once again, apologies for the charts.

[Frances_iom]

I wonder if we are seeing the false positives from the lateral flow tests being done in the week or before the schools reopened - there had been comments that without a check on the positivity that these tests could skew the figures

[Simon B]

A conservative estimate on the false positive rate of even the better lateral flow tests (i.e. hopefully the ones the govt has authorised to be used - most of the candidates were rubbish IIRC) is 0.1%. Can't find the source right now, but Lancet, BMJ, something like that...

There hasn't just been a significant increase in testing recently - it's huge. +100% since Mar/1 - 6e5 a day to 1.2e6. Most of those 600k extra tests are presumably lateral flow in schools etc - so that's likely at least 600 additional false positives a day introduced for a start - 10-15% of the current daily total.

At this point it is probably necessary to look at the % +ve to make sense of the d/dt of +ve tests rather than the absolute numebrs.

Without having the time to do it myself, gut feeling is that +100% on # of tests could turn the previous -25%/week trend in # of +ve tests into an apparent -2% (the current week-on-week change in 7 day mean) without the underlying phenomenon having changed.

Getting into "Best ask a statistician" territory with too many variables changing simultaneously.

[french frank]

I'm sure he's right, but he also writes this about some of the vaccine trials

Some kind of adverse events were reported by 38% of those receiving the real vaccine but, rather remarkably, 28% of those who received the dummy also reported a side-effect. This shows that the vaccination process itself causes about two-thirds of all the reported harm.

So on the one hand he is denying the use of post hoc ergo propter hoc causality arguments, then a few paragraphs down he slips up and does the same himself!

I think that a psychosomatic side-effect could be caused (as with placebos) - I had a very slight 'tenderness' on my arm, but when I had to report on the app I also thought I had an ache in my other arm, both appearing and disappearing at the same time . I doubt blood clots could be psychosomatic.

[teamsaint]

I think that a psychosomatic side-effect could be caused (as with placebos) - I had a very slight 'tenderness' on my arm, but when I had to report on the app I also thought I had an ache in my other arm, both appearing and disappearing at the same time . I doubt blood clots could be psychosomatic.

One might think that the blood clots issue could be pretty easily dealt with. Background rates must be known, and the yellow card ( or similar abroad) reporting on post vaccine incidence must surely be reliable too.

Break it down by batch to check for problems at batch level, check got any increased incidence against obvious factors like age, and job done, I would have thought ?

[Cockney Sparrow]

One might think that the blood clots issue could be pretty easily dealt with. Background rates must be known, and the yellow card ( or similar abroad) reporting on post vaccine incidence must surely be reliable too.......

Well, yes - Spiegelhalter - #5195 above.

“There's no proof the Oxford vaccine causes blood clots. So why are people worried?” - David Spiegelhalter

Guardian Article puts it in perepective:
https://www.theguardian.com/commenti...a-causal-links

As it happens, I had an asthma attack for the first time since my teens (when I had many, many of them) on the night of my Astra Zen 1st dose. And the following night, both on retiring (so going horizontal). Not evident since, and after the 2nd dose I'll try dropping the preventative medicine if the GP agrees.

I'd had very, very slight chestiness the week before, but it had abated and disappeared and I didn't think of that for a moment when offered the vaccination. A nurse I saw said she had a lot of pain in her arm about 10 hours after the injection and a sense and pain of someone pulling on her hair - pony tail - (on her 2nd dose). But we both agreed - hey - its a lot better to have the protection than be at risk.

As an addition on edit - I have to say having the vaccination hadn't preyed on my mind either before the day nor on the day - really I'm not a worrier, but my approach is to take action to limit risk but not worry about whatever it is that might go wrong. It was a complete surprise to experience the onset of Asthma as I retired.

[oddoneout]

One might think that the blood clots issue could be pretty easily dealt with. Background rates must be known, and the yellow card ( or similar abroad) reporting on post vaccine incidence must surely be reliable too.

Break it down by batch to check for problems at batch level, check got any increased incidence against obvious factors like age, and job done, I would have thought ?

First reports mentioned a suspicion of connection with a particular batch but I don't know what the current state of play is with that. I have also read mentions that the incidence of post vaccine blood clots is actually lower than incidence in the general population. What I think may be more relevant is whether the nature of the blood clots is different, ie more likely to be related to the vaccine itself.
The matter of side effects and sorting out actual risk as opposed to perceived, let alone placebo effects, can be difficult. Some years ago an osteoporosis drug I took briefly had as its first common side effect Heart attack, which is a bit disconcerting until I looked a bit further at what that was based on, which was a small trial in sedentary over 80 year olds (residential care) with pre-existing cardiac and vascular conditions. That has limited value when compared to the majority of patients likely to be taking the drug. Also what those bald two words didn't say was 'non-fatal' - even in the trial group. I gather that further usage and reporting has produced a more realistic picture.

[johnb]

A conservative estimate on the false positive rate of even the better lateral flow tests (i.e. hopefully the ones the govt has authorised to be used - most of the candidates were rubbish IIRC) is 0.1%. Can't find the source right now, but Lancet, BMJ, something like that...

There hasn't just been a significant increase in testing recently - it's huge. +100% since Mar/1 - 6e5 a day to 1.2e6. Most of those 600k extra tests are presumably lateral flow in schools etc - so that's likely at least 600 additional false positives a day introduced for a start - 10-15% of the current daily total.

Yes the increase in LF tests is huge, going from 330k daily in late Feb to an average of 994k in the w/e 14th March.

I've just been looking at the Test & Trace stats. One thing that stands out is the the overall positivity rate for LF tests in educational establishments (nurseries. primary, secondary and further education) is 0.10% for w/e 24th March and 0.07% for w/e 3rd March. With these very low levels of positivity it is very likely that the reported positives are dominated by false positives.

It does make you wonder how much of a benefit they are in this situation.

(There have been various reports of the level of false positives with the LF tests. A joint study, last November, by Porton Down and Oxford University arrived at a figure of 0.32%, which was lowered to 0.06% in a lab setting.)

[teamsaint]

Yes the increase in LF tests is huge, going from 330k daily in late Feb to an average of 994k in the w/e 14th March.

I've just been looking at the Test & Trace stats. One thing that stands out is the the overall positivity rate for LF tests in educational establishments (nurseries. primary, secondary and further education) is 0.10% for w/e 24th March and 0.07% for w/e 3rd March. With these very low levels of positivity it is very likely that the reported positives are dominated by false positives.

It does make you wonder how much of a benefit they are in this situation.

(There have been various reports of the level of false positives with the LF tests. A joint study, last November, by Porton Down and Oxford University arrived at a figure of 0.32%, which was lowered to 0.06% in a lab setting.)

Quite a benefit, if you are being paid to run the testing system.........

[Anastasius]

I'm sure he's right, but he also writes this about some of the vaccine trials



So on the one hand he is denying the use of post hoc ergo propter hoc causality arguments, then a few paragraphs down he slips up and does the same himself!

Well, if you've got all those column inches to fill....logic goes out the window, I guess.

Digressing slightly but still highlighting just how few people actually think when consuming stuff from the media, we happened to watch Gogglebox where they were showing the reaction to 'that' interview. Not a single one of them actually stopped to ask "Is she telling the truth or just being highly manipulative".

[Anastasius]

First reports mentioned a suspicion of connection with a particular batch but I don't know what the current state of play is with that. I have also read mentions that the incidence of post vaccine blood clots is actually lower than incidence in the general population. What I think may be more relevant is whether the nature of the blood clots is different, ie more likely to be related to the vaccine itself.
The matter of side effects and sorting out actual risk as opposed to perceived, let alone placebo effects, can be difficult. Some years ago an osteoporosis drug I took briefly had as its first common side effect Heart attack, which is a bit disconcerting until I looked a bit further at what that was based on, which was a small trial in sedentary over 80 year olds (residential care) with pre-existing cardiac and vascular conditions. That has limited value when compared to the majority of patients likely to be taking the drug. Also what those bald two words didn't say was 'non-fatal' - even in the trial group. I gather that further usage and reporting has produced a more realistic picture.

Very true. Same thing has happened with 'Is wheat glucose OK for coeliacs'. CoeliacUK say 'fine'. When pressed, they refer to a Finnish (IIRC) study claiming wheat glucose was OK. Digging deeper, the sample size was in the 10's. There was just the single sample of wheat glucose tested. Said sample provided 'by the industry'. The fact ? Yes, wheat glucose is fine for coeliacs if it's manufactured properly. That's a very big 'IF'.

[oddoneout]

Well, if you've got all those column inches to fill....logic goes out the window, I guess.

Digressing slightly but still highlighting just how few people actually think when consuming stuff from the media, we happened to watch Gogglebox where they were showing the reaction to 'that' interview. Not a single one of them actually stopped to ask "Is she telling the truth or just being highly manipulative".

Although I agree about the tendency to just accept what is being presented unquestioningly, without seeing all the footage I'm not sure it's possible to know that there was no discussion of the sort you mention. Those responsible for the finished product, whether papers, TV, online, and whatever the subject, will put out the version that meets their aims.

[Frances_iom]

...
So on the one hand he is denying the use of post hoc ergo propter hoc causality arguments, then a few paragraphs down he slips up and does the same himself!

No you misunderstand the statistics - you take two supposedly identical groups jab one with vaccine and jab the other with saline then see how many report any side effects - the difference in this experiment was 12 IIRC with 38 in vaccine grp + 26 in non vaccine displaying I presume a similar range of, I assume slight, side effects hence the comment that the vaccine did cause some additional side efffects over + above those that would be caused by a jab on its own.

Now the Norwegian cases (of which we have seen few if any details) state that in X number of jabs (X not being given) then we saw 7? bad effects within Y days of the jab (Y also not being given) then the test to apply is given X non jabbed drawn from a similar population then over Y days how many similar bad effects would we expect to see - statistics from the several million UK jabs show that those jabs have a lower change of bad effects than a non-jabbed population.

This then raises a question was there some other common factor that those jabbed shared that was different from a random population - one suggestion is that they may already have been incubating covid against which the jab would give no protection