search

LFOD / real-blog

13 stars 0 forks source link

2021/07/21/denominators-matter/ #95

Open utterances-bot opened 3 years ago

utterances-bot commented 3 years ago

Live Free or Dichotomize - Denominators Matter

https://livefreeordichotomize.com/2021/07/21/denominators-matter/

DaRedPenguin commented 3 years ago

Interesting post. My question is:

Do we have real world figures to show cases where a 'majority vaccinated' population group had cases in which, 100% of the unvaccinated were infected vs ~11% of the vaccinated population were infected?

Wreny commented 3 years ago

Thanks Lucy. I love the name of your blog.

I see the point in comparing things relative to the right thing.

I came here from the ABC / RMIT fact check site (https://www.abc.net.au/news/2021-08-06/coronacheck-israel-higher-numbers-vaccinated-people-infected/100353540) which used your example in comparison to the outbreak in Barnstable (https://www.cdc.gov/mmwr/volumes/70/wr/mm7031e2.htm?s_cid=mm7031e2_w).

Would you please be able to run the numbers from the CDC through your methodology? I've tried myself, but I'm not an expert. I did a spreadsheet based on your example.

Example

Population Vaccinated 18 Population Unvaccinated 2 Total Population 20

Vaccinated Sick 2 Unvaccinated Sick 2 Total Sick 4

% of sick being vaccinated 50% (i.e. big scary number, but the wrong way to look at it)

% Population Vaccinated 90%

% of Vaccinated getting sick 11% % of Unvaccinated getting sick 100%

Vaccine efficacy 89%

The only info I didn't have directly from the CDC report was the population numbers. I ran it a few times with population data from wikipedia. First with the full population of the town of Barnstable. Once with the median town population plus allowance for 'thousands of tourists,' and once with the smallest town population plus the tourist allowance. Although this did affect the % of vaccinated and % of unvaccinated getting sick - it didn't change the Vaccine efficacy rate.

Based on my calculation, the vaccine efficacy for cases is -26%. For serious cases (i.e. hospitalisation) it was -80%. This kind of makes sense as there were more cases and hospitalisations as a %age than the population as a %age. Which indicates that the vaccines are less than effective? How could that be? Antibody Dependent Enhancement?

Cases in Median Town

Population Vaccinated 10971 Population Unvaccinated 4929 Total Population 15900

Vaccinated Sick 346 Unvaccinated Sick 123 Total Sick 469

% of sick being vaccinated 74% (i.e. big scary number, but the wrong way to look at it)

% Population Vaccinated 69%

% of Vaccinated getting sick 3.15% % of Unvaccinated getting sick 2.50%

Vaccine efficacy -26%

Hospitalised

Population Vaccinated 10971 Population Unvaccinated 4929 Total Population 15900

Vaccinated Sick 4 Unvaccinated Sick 1 Total Sick 5

% of sick being vaccinated 80% (i.e. big scary number, but the wrong way to look at it)

% Population Vaccinated 69%

% of Vaccinated getting sick 0.04% % of Unvaccinated getting sick 0.02%

Vaccine efficacy -80%

Maybe I'm wrong. Hopefully I'm wrong. If I am, can you please show me where.

Obviously this is a small sample. Maybe an outlier? Too many other factors? Of course it won't be a perfect analysis, but what kind of analysis could you make from the CDC data?

It would be great to have someone like yourself assist ABC/RMIT factcheck with this kind of assessment rather than them simply showing your theoretical graphic showing 89% efficacy.

LucyMcGowan commented 3 years ago

The % of the exposed population that was vaccinated is an important piece here that we don’t have a precise estimate of. I’ve seen it a few places that ~60,000 were in attendance. The NY Times mentions one estimate “Provincetown had one of the highest vaccination rates in the country, upward of 95 percent among permanent residents”. If this is closer to reality, it would change your numbers to:

Population Vaccinated 57,000 Population Unvaccinated 3,000 Total Population 60,000

Vaccinated Sick 346 Unvaccinated Sick 123 Total Sick 469

% of sick being vaccinated 74% (i.e. big scary number, but the wrong way to look at it)

% Population Vaccinated 95%

% of Vaccinated getting sick 0.6% % of Unvaccinated getting sick 4.1%

Vaccine efficacy 85%

The other issue that this post doesn’t address is confounding, that is, if the population have characteristics that make them different, for example if the vaccinated population was more vulnerable than the unvaccinated (perhaps older, more likely to be immunocompromised) or vice versa if the unvaccinated were more vulnerable (perhaps more willing to take risks) it could skew these numbers in a way that makes it difficult to ascertain what happened due to the vaccine and what happened due to underlying characteristics in the population. The information we have publicly available right now doesn’t allow us to look at that.

a81j commented 3 years ago

Hi,

If the 60,000 visitors to Provincetown (population 3000) had a population distribution like that of the rest of Massachusetts or Barnstable, about 15% of the population would be 12 or under meaning they would be unvaccinated. So we would have a % of population vaccinated of about 80% for that 95% vaccination rate.

Likely it would be lower, as Barnstable county (where the visitors to the town would most likely be from) has a vaccination rate of 77% of people over 12 as of August 5 ( 75% as of July 1):

The population under 12 can't be excluded because the cases are counting infections under 12 years too:

Using 80% vaccinated of total population:

Population Vaccinated 48,000 Population Unvaccinated 12,000 Total Population 60,000

Vaccinated Hospitalized 4 Unvaccinated Hospitalized 1 Total Hospitalized 5

Vaccinated Infected 346 Unvaccinated Infected 123 Total Infected 469

% of Vaccinated infected 0.72% % of Unvaccinated infected 1.03% Efficacy 30%

% of Vaccinated hospitalized 0.0083% % of Unvaccinated hospitalized 0.0083% Efficacy 0% ?

The hospitalization figure is the more interesting - because it is what the vaccine promises to reduce. Plus we can be certain we are counting all of the hospitalizations, but can't be sure we are counting all of the infected.

The efficacy doesn't look good there - 0% ? Have I calculated right

https://www.capecodtimes.com/story/news/2021/08/09/more-than-three-quarters-cape-cod-vaccinated-against-covid-19/5531197001/

https://www.cdc.gov/mmwr/volumes/70/wr/mm7031e2.htm

B1ppp commented 3 years ago

Very well explained. I wonder if this method would be the same for Bulgaria, for example. 20% vaccinated, new cases per day 6000, 80% of them unvaccinated. Deaths - 300 / day, 90% of them unvaccinated. So the ratio of risk for vaccinated is much higher that the one is your example. What do you think is the factor influencing these figures?

refreshro commented 3 years ago

Several factors to be considered in Bulgaria's data, similar to Romania;