Vaccine Effectiveness Shows in Declining Case Rates

 

I have heard this week, in multiple media reports, that infection rates are falling everywhere in the U.S., and that it's because of vaccines. I've not seen any data on the latter claim. The data are available, so I decided to geek out and see for myself.

Let's start with the first claim, that infection rates are falling everywhere.  That is easy to check with data from CDC's Socrata API.  I queried it to get the new cases for the seven days ending May 18 (CDC data are a few days behind) and the same for the seven days ending April 19. 

I converted these to cases per 100K using population numbers, got an average for the current week and the week one month ago, and subtracted the two values.  Here are the results:

Case rates have not declined in every state, but close.  They have increased in Alabama by 7.35 and in Wyoming by 2.36.  There is essentially no change in Arkansas, Mississippi, Hawaii, Arizona and Missouri.  

In every other state, rates have decreased by more than 1 case per 100K over the last month.  Michigan and New Jersey have seen huge declines. Good thing too  because they were getting hammered a month ago.

Can we associate vaccinations with these declines?  I added the current percent fully vaccinated to the one-month change data.  Here is a map showing both values for the various states:

In this map, the circles represent the one-month change in new cases per 100K just described.  The blue circles represent increases, and the white circles with blue borders represent decreases.  The bigger the change over the last month, the bigger the circle. 

The purple shading represents the vaccination rate. The darker the shade, the higher the rate.  It goes from a low of 24.7% for Mississippi to a high of 45.6% for Connecticut.

If vaccines are affecting the case rate declines, we should expect to see the darker purple states with larger white circles, and the lighter purple states with smaller and/or blue circles.  

This seems to be true in the East and South, but less so in the West.  This may be because the West did not have the same surge in April that Eastern states did.  They tend to have relatively low infection rates at present.  Perhaps then, there is less "room" for declines in the Western states.

We can also look at a scatterplot of the vaccine percentage versus the one-month change:

The trend line shows that the more vaccinated a state is, the more its cases dropped over the last month.  We can get the slope of that line using a linear regression. It shows that for every additional 10% of population fully vaccinated, the cases per 100K drop by 8.94. Percentage vaccinated accounts for about 24% of the variation in the case rate, so vaccine percentage is influencing it, but it's not the only thing. 

There are outlier cases like Michigan, which is near the middle of the pack on vaccination percent but had a huge drop in their case rate.  There is also Hawaii, which despite having one of the highest vaccination percentages, had essentially no change in its case rate.  

So something besides vaccinations is influencing these outlier states. In any case it's reassuring that vaccinations are associated with falling infection rates when you look at it "by the numbers."

Header image by torstensimon from Pixabay