Weekly COVID-19 Round-up - 5/20/21
River City Data No. 32
Tonight’s newsletter has my usual breakdown of our top-line numbers and details on trends in infections in Missouri. In the first “COVID-19 by the Numbers” section, you'll notice that the vaccination numbers are both much lower than they were last week. This is due to changes that I’ve made in how I calculate our progress. So, after my overview of infection trends, I have a deep dive into how vaccination progress is being reported by DHSS and discuss what is being both included and left out. - Chris
COVID-19 by the Numbers
Total cases in MO: 610,209 (+2,940 from last Thursday)
7-day average of new cases per day in MO: 420 (-27 from last Thursday)
Counties with the highest per capita rates (per 1,000) of new cases per day this past week and the only counties with per capita rates over 13:
Putnam (47.47 per 100,000), Livingston (41.69), Shelby (40.07), Linn (23.45), Lewis (19.95), Madison (17.56), and Polk (13.13)
Total deaths in MO: 9,550 (+183 from last Thursday)
7-day average of new deaths per day in MO: 26.14 (+19.29 from last Thursday)
Percent of people initiating vaccination: at least 36%
Percent of people completing vaccination: at least 30%
Case and mortality numbers are current as of Wednesday, May 20th. Vaccination numbers are current as of Thursday, May 21st. Additional statistics, maps, and plots are available on my COVID-19 tracking site.
Illness Trends in the Past Week
Our overall trends in Missouri have leveled off in the past week and even increased slightly. Late last week, our statewide 7-day average briefly dipped below 400 new cases per day. However, as you can see below, it has risen ever so slightly since my last newsletter on Saturday.
In “outstate” Missouri, however, the number of new cases has remained largely static. One reason for this is that rural counties continue to experience sharp spikes in their per capita rates. However, they are short-lived, and so the overall 7-day average does not move all that much. For example, take Shelby County in Northeastern Missouri, where we saw a large spike in their per capita rate followed by a just as quick decrease.
One thing to note about Shelby County and the other two counties with higher per capita 7-day averages right now (Putnam and Livingston) is that the number of new cases being added in absolute terms is pretty small - three to six cases per day in each of these counties.
Hospitalizations continue to be a bright spot as well. For example, in St. Louis, where new case rates are declining, hospitalizations are at all-time lows in the daily count of in-patients and the 7-day average of this metric.
Our less recent statewide data, which reflects conditions in early May, suggests that these in-patient patterns are broadly occurring across Missouri. There were a few slight upticks in several regions, but the overall trends are towards far lower hospitalization numbers than we saw between November 2020 and February 2021.
Vaccination Trends in the Past Week
Vaccination trends continue to be concerning. First, the pace of vaccinations keeps slowing. For example, the same map below had a top per capita rate of 43 vaccinations per 1,000 residents in my last newsletter. This week, the top recent vaccination rate per 1,000 has dropped to 33.5.
The general geographic trend remains the same, however. Again, the counties in and around St. Louis have the most notable cluster of high rates of new vaccinations. Boone County in Mid-Missouri, Joplin, and then around Kansas City round out the areas where the pace of vaccinations is the highest (in relative terms).
Vaccination Deep Dive
In the past week, I have been exploring data DHSS publishes on their vaccination progress. Some of these data are summarized on their COVID vaccine dashboard. For example, one of the top-line statistics is that 40.3% of “the population” has initiated vaccination, and 33.4% of “the population” has completed vaccination.
If you hover over these details, there is a note saying that some people may have listed an address outside of Missouri. However, it has not been possible to quantify what share of initiated or completed vaccinations went to out-of-state residents. For the first time tonight, thanks to newly released data today from DHSS, it is possible to do so.
For both initiated and completed vaccinations, the breakdown is roughly the same. About 84% of each group of vaccinations went to someone living in Missouri whose jurisdiction was known (a county, or the cities of Kansas City and Joplin). Another 4% of each group went to someone living in Missouri whose address was unclear, and so they cannot be concretely located in a jurisdiction. Finally, about 5.5% of each category were vaccinations to out-of-state residents, and another 6.5% went to people who had no address reported.
I do not want to suggest that people from outside of Missouri should not receive vaccinations here. Health care workers, college and graduate students, and many others work in one state but live in another and may have had access to vaccination through their school or employer. Likewise, people may have found a CVS or Walmart closer to their home but across state lines. All reasonable, understandable situations.
What I do want to suggest is that out-of-state vaccinations, and those with no known address, should not be included in Missouri’s calculation of the total percent of the population that has been vaccinated. The denominator in this calculation is the total number of people in Missouri (just over 6.1 million people). If we add in out-of-state individuals, we artificially increase the percentage of the population that appears to be vaccinated.
The difference is not small, either. You can compare DHSS’s reported percentages of 40.3% initiated and 33.4% completed to my revised tracking plot:
When I remove those out-of-state and unknown address individuals, our progress is about four percentage points behind on initiation and three points behind on completion. This equates to nearly three hundred thousand individuals in terms of initiated vaccinations, for example.
It is also worth noting that there are people left out of Missouri’s numerator, too. These are broadly two groups - individuals vaccinated at a Federal facility like a VA Hospital and individuals who live in Missouri but crossed into another state (or were attending school in another state) to be vaccinated. In their fine print, DHSS does note that the second of these two groups are missing from their data.
It is, therefore, possible that this is all a wash, and the percentages DHSS reports are roughly reflective of the “true” vaccination coverage. However, we do not have the data to assess whether this is actually the case. My view, then, is that I want to report lower, more conservative estimates of vaccination progress. Therefore, I have added the “at least” language to the image above to convey the uncertainty around both unknown addresses and individuals we know are being left out of DHSS’s numbers.
As for DHSS, I do not think they should be reporting percentages calculated using Missouri’s population where they knowingly add more than a hundred thousand out-of-state residents. I also think they report too many significant digits - 40.3% initiation vaccination sounds so certain, but we know that this percentage is wrong. As you continue to digest our progress with vaccinations, keep this uncertainty in mind. The reality is we have no concrete idea what proportion of the population is vaccinated. It is certainly over 30%, but we are just not sure by how much.
If you like what you see here and don’t already, please subscribe!