In my previous post, I highlighted the 3-year drop in US life expectancy from 2019 to 2021. It contrasts with European and other high-income countries where the availability of vaccines and continued implementation of other public health measures resulted in an increase in life expectancy in 2021 after the Covid-related 1-year drop in 2020.

A World Health Organization (WHO) expert group has estimated excess deaths by country and month during the pandemic years 2020 and 2021. The plot below compares excess death rates for the USA with those for the aggregate of the five largest European countries: Germany, France, Italy, Spain and the United Kingdom (“Europe 5”).

Unlike Europe, the USA had an excess death rate in the second half of each year that was around double the death rate in the first half of the year. Covid-19 vaccines became generally available in Europe in early 2021 and in the USA around April 2021. The much higher death rate in the USA in the second half of 2021 is likely associated with slower and lower uptake of vaccination.

The Seattle-based Institute for Health Metrics and Evaluation (IHME) has also estimated excess deaths by country for the period 2020-2021 compared with 2019. I used their estimates for US states (available here) to examine the differences in excess deaths associated with higher and lower vaccination rates and with support for the Republican candidate in the 2020 presidential election.

I sorted US states into ascending order of the per cent of the total state-level votes for Trump and Biden in the 2020 presidential election that were votes for Trump. I then grouped the sorted states into three groups (tertiles) of roughly equal populations. I similarly produced tertiles of states sorted in descending order according to the per cent of the population fully vaccinated in August 2022.

The following plot shows a scatterplot of US states vaccination status versus per cent voting for Trump. Ignoring the outlier point with 5.5% voting for Trump (District of Columbia), these two variables are strongly correlated across states, with a correlation coefficient of -0.87.

IHME estimates of excess deaths for 2020-2021 by state were grouped into voting tertiles and vaccination tertiles, and the excess death rates per 100,000 population are shown in the Table below.  If the second and third tertiles had had the excess death rate of the first tertile, there would have been 141,000 fewer deaths (based on voting tertiles) or 152,000 fewer deaths based vaccination tertiles. The numbers are quite similar when states are grouped by vaccination or voting behaviour, reflecting the close correlation between voting behaviour and vaccination levels.

While the evidence I’ve reviewed above linking political partisanship and covid-19 deaths is based on group-level correlation, which by does not necessarily imply causation, in this case the conclusion seems obvious.  Fortunately, we now have a study which took 577,659 death records from Ohio and Florida between January 2018 and December 2021 and matched the deaths to a 2017 voter file. In other words, they were able to identify the partisanship not only of the places those people lived but of the people themselves. The plot below shows the excess death rates according to the political affiliations of the people who died. The first solid vertical line marks the time of the beginning of pandemic in the USA and the second solid line shows the point at which all adults became eligible for covid vaccines.

The death rate was a little higher for Republicans in 2020, but started to diverge substantially after vaccines became generally available. The well-documented differences in vaccine attitudes and uptake between Republicans and Democrats directly resulted in higher excess deaths among Republicans.

Complete state-level data on deaths by age and sex for 2021 are not yet available, but I have used GBD2019 estimates of US state death rates by age and sex (IHME 2020) together with CDC information on state-level deaths up to 2020 and cause-specific increases in deaths in 2021 to impute the age-sex specific deaths data needed to compute average life expectancy at birth by US voting tertiles.  The estimated life expectancies for the 1^st voting tertile (labelled USA-Dem) and the 3^rd voting tertile (labelled USA-Rep) are included in the graph below. The curve for the 2^nd tertile is not shown as it is quite close to the line on the graph for the USA as a whole.

The gap between the 1^st and 3^rd tertiles widened from just under 2 years in 2000 to 3 years by 2014. This gap was fairly stable at 3.2 years in 2018 and 2019 and only widened by 0.1 year in 2020. However, in 2022, when vaccines became available it jumped to 4.1 years.

Another recent paper has looked at the difference between the most-vaccinated and least-vaccinated states for the more recent period of June 2021 to March 2022.  The table below compares the 10 most-vaccinated states and the 10 least-vaccinated states with the overall USA covid-19 mortality and that for Europe-5, Australia and Switzerland. Total covid-19 mortality in the least-vaccinated states was three times higher than in the most-vaccinated states and US covid-mortality overall was 2.7 times higher than for the Europe-5 countries.

The final columns of the table show the potential covid deaths that the full USA would have averted if it had had covid-19 death rate of the specified country/states during the period of interest. Whether it is assumed the USA could have had the excess covid death rate of the 10 most vaccinated states or the Europe-5 countries, the potential deaths averted would have been close to 300,000.

Putting together this analysis with my earlier analysis for the period 2020-2021, it seems reasonably to conclude that differences in vaccine uptake due to political partisanship have likely cost the USA at least 350,000 additional covid-19 deaths to date than would have otherwise occurred.