This is a quick update to catch up on the work Oregon is doing to improve their public breakthrough data — first we need:
#include
We’re continuing from there…
So Oregon has added two more weeks of data since we last updated, and has in their breakthrough reports1 now nailed down the case vaccination status for most of the omicron outbreak. In each subsequent release they’ve gone back and corrected data for the prior weeks they include in the reports.
We have seen two improvements in the presentation of Oregon breakthrough data in recent weeks:
They no longer lump unknown vaccination status with the unvaccinated
They have re-released numbers back several weeks with updated knowledge of the vaccination status of cases, now approaching 100% known.
Below I’ve cooked up a plot to illustrate where things sit at this point and the process it looks like they’ve been going through. I hope we see other states improve (or maybe even provide) data more transparently like this!
This plot will take some explaining — there is a lot happening here.
First, this is plotting
Calculated from each week’s breakthrough cases and vaccinated populations. There are a bunch of curves in the plot. This is trying to illustrate the situation after each of Oregon’s published breakthrough reports — the color of the curve belonging to the date of the report shown in the legend. The end of each colored line is also a way to visualize that. Obviously you’re going to want to see a curve end before the date of its report.
I have a solid black line covering over a lot of it. That is extracted using the numbers in the most recent report last week. That’s the current best knowledge of the week to week effective Ev in Oregon. At least from public data. There are error bars on some of the points — I’ll get to that in a bit.
You also see two classes of colored lines in the plot — dotted and solid. The dotted lines are from earlier reports, where we only saw “unvaccinated” and “vaccinated” categories of case counts. From the Feb 10 report onwards though we began to see an additional “known vaccination status” in their table. The solid colored lines then are calculations of Ev only based on the known status, ignoring those with unknown status (which is logical since that status is what we’re using to land people in either half of the fraction we’re calculating). In prior weeks Oregon had been incorrectly including unknown status with the unvaccinated, which would bias an Ev calculation higher than it should.
There are also fluctuations in the solid lines as we progress. Until a few weeks ago, only around 70% of cases had known status. Oregon has been updating numbers as they determine the vaccination status of people and updating past weeks.
Which then brings us to the solid black line. This is based on the latest data. But note I have a couple error bars added in early January. Each week’s breakthrough report includes the past 6 weeks worth of data. As Oregon has caught up in its understanding of case vaccination status, bringing that to nearly 100% in the last couple weeks, those points also unfortunately dropped out of the report. I am then left with data as of a couple weeks ago, that only had 70% known status. So for these points I am showing an error bar swinging the remaining ~30% unknown between the vaccinated and unvaccinated. That defines the error bars — though its really a maximum possible envelope given the data we have. Currently the data from the majority of cases we do have known status from provides the solid line.
Some quick observations:
We still see a drop brushing negative efficacy in the beginning of Jan, on the leading edge of the omicron outbreak.
Though I show a large error bar here, if things work out statistically that error bar should be a huge overestimate of the uncertainty there.
We seem to be converging to the 30-40% vaccine efficacy range for omicron after that initial spike.
It is refreshing to see Oregon trying to present a clearer picture here — it would be very interesting to see the numbers from earlier weeks with the vaccination status nailed down. It would also be interesting to see case counts from the now considerable boosted population called out — that does seem to be available, given the plots included in the reports. Providing us the numbers along with the reports would be excellent!
https://www.oregon.gov/oha/erd/pages/covid-19-news.aspx. — particularly seek out the dated “Breakthrough Cases Report … “ entries under “COVID-19 Data Reports and Projections”
44% boosted with 75% vaccinated I would think means 58% of the vaccinated people in Oregon being boosted. https://public.tableau.com/app/profile/oregon.health.authority.covid.19/viz/OregonCOVID-19VaccineEffortMetrics/StatewideProgress