Mar 31, 2022·edited Apr 1, 2022Liked by T Coddington
Thanks for the update.
While the math that leads to "somewhere in the 33% lower range" might not have changed much for the overall timespan, it probably doesn't apply to the discrete rate for December - it seems unlikely that as many pregnant women would have put off injection until October or later rather than having already made their decision by then.
I see your point. On the other hand, my previous calculations arriving at the 33% number did not include an adjustment for socio-economic status. I've not written here about this, but I estimate based on economic status alone (vaccinated women are much less deprived), the vaccinated women should have stillborns ~10% less often than unvaccinated women, so this pushes us more in the area of concern again. Let me see if we have the data to estimate what fraction of Dec vaccinated births were vaccinations in 3rd trimester.
The horrible study from Scotland suggests a drop-off in uptake after September (https://www.nature.com/articles/s41591-021-01666-2 Fig 3). Technically September is still in the window, but only a fraction of it.
Right, a demographic discount could explain why this "corrected" rate (as suggested in my other reply) is still not even higher.
In fact, this could be the cause of the "rise" - the immortal time bias that you suggested in your earlier post is evaporating, and now the true rate is showing in the real time (last month) math.
I fear you are correct . I have tried to make sense of the data and it feels off. Much like the NEJM study some time ago; I think they conflate different time periods and status.
Not sure I can prove it (yet) but I think they may take overall % vaccinated at end of Dec and rate the outcomes against that %, rather than chart a moving picture. A pregnant and jabbed women at first tri is not the same as at final tri. And nor can they give the same data as to full term etc etc
What might be the tell is looking at the reports over the last few months and trends.
For example, if I recall correctly, % low birthweight was higher for those initially vaccinated ~ 12 weeks ago yet is now lower/the same as for unvaccinated. This does not compute if everything else is equal.
What stuck me (apart from your main finding which was well made) was the proportion of births in each category. We’re women who were getting pregnant staying away from the jab due to fears, or is fertility being somewhat effected by the jab?
Picking up from here: "Now, I take the 74 stillborns among the unvaccinated and divide by the 19,029 births in December and I get a rate of 3.89 stillborns per 1,000 births… basically the same as the 3.90 we have been seeing all year for the unvaccinated. For the vaccinated, I take the 96 stillborns and divide by the 22,312 December births and I get a rate of 4.30 (~11% higher than unvaccinated)."
...I calculated the test statistic for a hypothesis test for the difference between the two proportions, 74/19029 and 96/22312 as -0.6553, which is not statistically significant using any usual threshold.
I admit to jumping down to these four supplied numbers (74, ...) and only quickly skimming the rest of this post, and am happy to be corrected.
The purpose of my previous 3 posts on this topic was to show that we shouldn’t expect parity in these rates, because of survivorship bias, parity is a bad sign…. The vaccinated should have lower rates, therefore December rates look concerning to me.
This is fascinating. I will need to go back and read more about your argument on timing, tho I’m guessing for now it would be whether the vaccine was administered earlier vs later on in pregnancy. If raw data were accessible, it would be interesting to know whether there would be a correlation of vaccine with people who have either had covid and/or have underlying health ailments.
Thanks for the update.
While the math that leads to "somewhere in the 33% lower range" might not have changed much for the overall timespan, it probably doesn't apply to the discrete rate for December - it seems unlikely that as many pregnant women would have put off injection until October or later rather than having already made their decision by then.
I see your point. On the other hand, my previous calculations arriving at the 33% number did not include an adjustment for socio-economic status. I've not written here about this, but I estimate based on economic status alone (vaccinated women are much less deprived), the vaccinated women should have stillborns ~10% less often than unvaccinated women, so this pushes us more in the area of concern again. Let me see if we have the data to estimate what fraction of Dec vaccinated births were vaccinations in 3rd trimester.
The horrible study from Scotland suggests a drop-off in uptake after September (https://www.nature.com/articles/s41591-021-01666-2 Fig 3). Technically September is still in the window, but only a fraction of it.
Right, a demographic discount could explain why this "corrected" rate (as suggested in my other reply) is still not even higher.
In fact, this could be the cause of the "rise" - the immortal time bias that you suggested in your earlier post is evaporating, and now the true rate is showing in the real time (last month) math.
I fear you are correct . I have tried to make sense of the data and it feels off. Much like the NEJM study some time ago; I think they conflate different time periods and status.
Not sure I can prove it (yet) but I think they may take overall % vaccinated at end of Dec and rate the outcomes against that %, rather than chart a moving picture. A pregnant and jabbed women at first tri is not the same as at final tri. And nor can they give the same data as to full term etc etc
What might be the tell is looking at the reports over the last few months and trends.
For example, if I recall correctly, % low birthweight was higher for those initially vaccinated ~ 12 weeks ago yet is now lower/the same as for unvaccinated. This does not compute if everything else is equal.
What stuck me (apart from your main finding which was well made) was the proportion of births in each category. We’re women who were getting pregnant staying away from the jab due to fears, or is fertility being somewhat effected by the jab?
Picking up from here: "Now, I take the 74 stillborns among the unvaccinated and divide by the 19,029 births in December and I get a rate of 3.89 stillborns per 1,000 births… basically the same as the 3.90 we have been seeing all year for the unvaccinated. For the vaccinated, I take the 96 stillborns and divide by the 22,312 December births and I get a rate of 4.30 (~11% higher than unvaccinated)."
...I calculated the test statistic for a hypothesis test for the difference between the two proportions, 74/19029 and 96/22312 as -0.6553, which is not statistically significant using any usual threshold.
I admit to jumping down to these four supplied numbers (74, ...) and only quickly skimming the rest of this post, and am happy to be corrected.
The purpose of my previous 3 posts on this topic was to show that we shouldn’t expect parity in these rates, because of survivorship bias, parity is a bad sign…. The vaccinated should have lower rates, therefore December rates look concerning to me.
This is fascinating. I will need to go back and read more about your argument on timing, tho I’m guessing for now it would be whether the vaccine was administered earlier vs later on in pregnancy. If raw data were accessible, it would be interesting to know whether there would be a correlation of vaccine with people who have either had covid and/or have underlying health ailments.