Dsew vaccination

[Ananya-Joshi]

Booster indicators at the state + level

Of note:

• Currently only pulling one file; potential problem with listing.
• Need the receiving directory!
• Did not add any new tests (did not see the need).

[Ananya-Joshi]

Converted to draft because safegraph is stalling..

[krivard]

would you merge in main please?

[krivard]

Is this a debugging leftover? Usually we try and use the structured logger to avoid mucking up the kibana dashboards

[Ananya-Joshi]

Ah, I'll make it a logger command (but maybe @nmdefries can weigh in ) - this is because only a pdf instead of an excel file was uploaded on Jan 25/26th to the CDC site, and thus returned None for dataset items). Is it worth keeping in and investigating?

[nmdefries]

If we want to fail in that case with a useful message, we can add an assert. But it would be reasonable to check if a given listing is empty in the for sig, lst in datasets.items(): loop below. If it is empty, continue to the next listing.

[Ananya-Joshi]

I think because it's not happened again, I'm going to just ignore it for now! That makes sense if it ever comes up again.

[nmdefries]

Since the booster field is under a new section heading ("overheader"), we need to include the vaccinations/boosters overheader, e.g. "COVID-19 VACCINATION DATA: LAST WEEK (January 15-21)", in parsing via including patterns for it in skip_overheader()

[nmdefries]

I was looking at all the vaccination-related fields CPR provides, including here for future reference.

We want total & new doses administered, + boosters. It's currently under discussion if total should be only initial doses, or include boosters. It's also unclear to me exactly what is included in CPR's "Doses administered".

Overheaders:
• (1) COVID-19 VACCINATION DATA: LAST WEEK (January 25-31)
• (exclude) COVID-19 VACCINATION DATA: % CHANGE FROM PREVIOUS WEEK
• (2) COVID-19 VACCINATION DATA: CUMULATIVE (January 31)
• (exclude) COVID-19 VACCINATION DATA: DEMOGRAPHIC DATA LAST WEEK
• (exclude) COVID-19 VACCINATION DATA: DEMOGRAPHIC DATA % CHANGE FROM PREVIOUS WEEK
• (exclude) COVID-19 VACCINATION DATA: DEMOGRAPHIC DATA CUMULATIVE

Headers:
• (1) Doses administered - last 7 days
• (1) (maybe) Doses administered per 100k population - last 7 days
• (1) Booster doses administered - last 7 days **** this PR
• (2) Doses administered
• (2) (maybe) Doses administered per 100k population
• (2) People with at least 1 dose
• (2) People who are fully vaccinated
• (2) People who have received a booster dose since August 13, 2021

From the data dictionary in the CPR spreadhsheets,

Doses distributed, people with at least one dose, and people who are fully vaccinated include the Pfizer-BioNTech, Moderna, and J&J/Janssen COVID-19 vaccines... People fully vaccinated includes those who have received two doses of the Pfizer-BioNTech or Moderna vaccine and those who have received one dose of the J&J/Janssen vaccine... The count of people who received a booster dose includes anyone who is fully vaccinated and has received another dose of COVID-19 vaccine since August 13, 2021. This includes people who received booster doses and people who received additional doses.

[Ananya-Joshi]

Hi Nat, I just saw your review, so I'll update the indicator shortly with your edits.

[krivard]

The count of people who received a booster dose includes anyone who is fully vaccinated and has received another dose of COVID-19 vaccine since August 13, 2021. This includes people who received booster doses and people who received additional doses.

Oh nooooo

This is a different definition than expected. Typically we'd consider an immunocompromised person who has received an additional dose as part of their initial series (ie 3 doses total) to be "fully vaccinated" and not "boosted", but this report seems to be tossing them into the "boosted" category so long as that 3rd dose occurred after their arbitrary August 13 cutoff.

I don't think there's anything we can do about that in our code, we'll just have to make it super clear in the documentation that this is how the CPR is counting it. cc @capnrefsmmat

[capnrefsmmat]

This is a different definition than expected. Typically we'd consider an immunocompromised person who has received an additional dose as part of their initial series (ie 3 doses total) to be "fully vaccinated" and not "boosted", but this report seems to be tossing them into the "boosted" category so long as that 3rd dose occurred after their arbitrary August 13 cutoff.

I don't think there's anything we can do about that in our code, we'll just have to make it super clear in the documentation that this is how the CPR is counting it. cc @capnrefsmmat

Not surprising that they're conflating the categories like this, but yes, that's definitely a problem we should clearly document. (I'd expect many of our users not to know the additional dose/booster dose distinction either.)

[nmdefries]

Looks good!

Note: when we're ready to publish these in the API, the new signal keys from SIGNALS will need to be added to the exported_signals setting in the Ansible production params template.

[Ananya-Joshi]

Visual and statistical checks: Comment from slack:

There are some counties that have "000" as the last 3 numbers. This is because until 11/15/2021, the state levels for people_fully_vaccinated were not NA. I have removed it in the signal_prechecks pdf, but I'm not sure how we want to handle them in the future. We also stop having data for some fips codes after December 16th.
I also don't know how to rotate the x-values on the legend for choropleth maps, so the values are hard to read, but the maps look okay in the signal_prechecks.pdf

I'm not sure why only < 1/2 of the first choropleth map is filled out at the county level the rest of the values I assume being NA.

signal_prechecks.pdf
correlations_analysis-1.html.zip

[Ananya-Joshi]

Before 03-09-2021, "People with full course administered" was used instead of "fully vaccinated". Should this be a separate signal or might there be a better way to combine the two explicitly in the code.
[EDIT: Addressed in Slack]

[nmdefries]

🥳

[krivard]

Great! Holding off on merging until we have the remaining approvals and API documentation ready to go.

[Ananya-Joshi]

Made a PR for the documentation, I think the visual and stat checks should be done - anything else left?

For the PR, looks like I'm running into an error I don't quite understand (because I only edited the markdown).

[nmdefries]

Added a note about the failure.