Florida on my mind

 In early September, I had a post inspired by the Covid surge in Florida.  The conclusion was that there was a strong relationship between vaccination rates and Covid hospitalization rates, but no evidence that other measures made a difference.  Florida was an outlier, having very high hospitalization rates despite an average rate of vaccinations.  In late September, it occurred to me that I hadn't heard much about Florida laely, so I did an updated version of the analysis.  The vaccination/hospitalization relationship was still there, and there was still no visible relationship to other restrictions (measured in late August).  Florida was somewhat above its predicted value, but no longer an extreme outlier.  A few days ago, I saw that Florida now has the lowest rate of new Covid cases, and one of lower hospitalization rates.   So here's another update:

The relationship between vaccination and hospitalization rates is still clear, although the estimated regression slope is smaller (about half as large) as it was in the first two analyses.  Given the features of the data (e. g., non-independence of the cases) I wouldn't put much weight on the difference in estimates--the main thing is that we see the same pattern at all three points, despite many shifts in the position of individual states.*  As before, there's more variation in hospitalization rates at low vaccination rates--not all states with low vaccination rates have high hospitalization rates, but all states with high vaccination rates have low hospitalization rates.  

In the second post, I found that higher hospitalization rates in late August were associated with more new vaccinations in September, suggesting that people were influenced by actual conditions--when they saw a lot of cases around them, they got vaccinated.  That wasn't the case for vaccinations in October--there was actually a negative relationship between hospitalization rates in late September and new vaccinations in October.  So maybe there is no general relationship--both the first and second estimates were heavily influenced by a few states.  

The states averaged 15 new doses per hundred between late August and late October, but there was a lot of variation, ranging from a low of 4.4 to a high of 19.4.  Three states stood out as having a low rate of new vaccinations--West Virginia (4.4), Kentucky (8.3), and New Mexico (9.2).  In my last post, I said "given the clear importance of vaccination rates and the lack of evidence that anything else matters now, I think there should be more attention to understanding differences in the pace of new vaccinations."  In doing this post, it occurred to me that, unlike new case rates, information on new vaccinations isn't routinely reported in the media.  Overall state-level vaccination rates are widely reported, but they don't change rapidly (the correlation between state-level vaccination rates in August and October is 0.99).  The New York Times shows change in state-level  vaccination rates, but those rates are cumulative, so you can't really see recent trends.  For example, I don't think anyone looking at the figure for New Mexico would see that it's been slipping recently.  Merely giving the rate of new vaccinations as much prominence as the rate of new cases, hospitalizations, and deaths might make a difference.  If the rate of new vaccinations is low, state officials should be getting questions about why that is the case and what they are doing about it.  So here is the list:

New shots per 100 people, August-October 2021

Alaska          19.5
North Carolina  18.5
South Carolina  18.3
Virginia        18.1
Florida         18.0
Mississippi     17.9
DC              17.6
Tennessee       17.5
Oklahoma        17.4
Pennsylvania    17.4 

Arizona         17.2
New York        17.0
Maine           17.0
New Hampshire   16.8
Connecticut     16.5
Rhode Island    16.4
Wyoming         16.4
Alabama         16.4
Georgia         16.2
California      15.9
Colorado        15.9
Oregon          15.8
Texas           15.8
Vermont         15.8
Montana         15.7
Arkansas        15.7
Maryland        15.6
Wisconsin       15.2
Delaware        15.1
Louisiana       15.1
New Jersey      15.0
Kansas          14.9
Nevada          14.6
Minnesota       14.6
Idaho           14.6
Nebraska        14.5
South Dakota    14.5
Utah            14.2
Massachusetts   14.1
Hawaii          14.0
North Dakota    13.8
 

Missouri        13.8
Iowa            13.5
Michigan        12.9
Illinois        12.5
Indiana         12.1
Ohio            11.8
Washington      11.3
New Mexico       9.2
Kentucky         8.3
West Virginia    4.4

*The correlation between state-level hospitalization rates in late August and late October is essentially zero (-.06).