The COVID-19 vaccine caused almost 290,000 US deaths in 2022 and 2023

CDC data from over 3,000 US counties, with a population of about 330 million people, show that the COVID-19 vaccine in 2022 caused 138,229 (95% CI: 90,178-186,279) deaths, and 150,602 (95% CI: 118,606-182,599) in 2023. Altogether, almost 290,000 deaths. The percentage increase in deaths caused by the vaccine was 4.40 (95% CI: 2.80-6.00) in 2022 and 5.12 (95% CI: 3.98-6.26) in 2023.

In a previous post, I showed that it is unlikely that COVID-19 vaccination reduced US mortality in 2021. Hence, there was an almost 290,000 net increase in US deaths between 2021 and 2023 due to the vaccine (Note 1). I plan to report on the 2024 data when they become available.

Assuming a small US county in 2022 experienced 120 deaths, compared to a baseline of 100 expected, the all-cause excess mortality rate was 20% (Note 2). In the following, I apply the concept to investigate whether COVID-19 vaccination had a genuine impact on 2022 and 2023 mortality in US counties. All data are from the US Centers for Disease Control (CDC), publicly available (Note 3).

The table below presents regression models that report on the association between US county-level per capita COVID-19 vaccine uptake by the end of 2021 (Model 1) and 2022 (Model 2), respectively, as independent variables (Note 4), and all-cause excess mortality in 2022 and 2023, as dependent variables. All analyses were weighted for counties’ population size (Note 5).

I included the lagged dependent variables as controls, i.e., all-cause excess mortality in 2020 (DV20) and 2021 (DV21), when testing the vaccine effect on 2022 all-cause mortality (Model 1), and adding all-cause mortality in 2022 (DV22) when testing the vaccine effect on 2023 all-cause mortality (Note 6).

Regressions weighted for counties’ population size with robust standard errors.

2022

Model 1 shows a strongly significant positive association between per capita vaccine uptake by the end of 2021 and 2022 all-cause excess mortality (Note 5). I.e., the higher a county’s vaccine uptake by the end of 2021, the higher its 2022 all-cause excess mortality.

Model 1 further reports that, by the end of 2021, the average per capita vaccine uptake was 144.8 doses per 100 people among the 3,068 counties included, with a total population exceeding 330 million (Note 7).

Using the average per capita vaccine uptake score (144.8 doses per 100 people) as an input value in Stata’s margins command algorithm returned a value of 113.7. I.e., average vaccine uptake by the end of 2022 corresponded with 13.7% 2023 all-cause excess mortality. The same exercise, with zero vaccine uptake as the input value, returned a value of 108.9. I.e., counties with estimated zero vaccination by the end of 2021 corresponded with 8.9% 2022 excess mortality (Note 8). These numbers show that counties with average vaccine uptake, corresponding to the overall US vaccine uptake, experienced a 4.40% higher all-cause excess mortality compared to counties estimated to have zero vaccine uptake.

As the US reported 3,279,857 deaths in 2022, a change, that is, an increase, was estimated to be 138,229. I reached that number by subtracting from 1 the ratio between “Mortality if vaccine uptake=0” and “Mortality if vaccine uptake=average”, and then multiplying it by the total US deaths.

Taken together, the average COVID-19 vaccine uptake by the end of 2021 resulted in 138,229 additional deaths in 2022 (Table 1 reports relevant 95% CIs).

2023

Model 2, similar to the previous one, reveals a strongly significant positive association between per capita vaccine uptake by the end of 2022 and 2023 and all-cause excess mortality (Note 5). I.e., the higher a county’s vaccine uptake by the end of 2022, the higher its 2023 all-cause excess mortality.

Model 2 further reports that, by the end of 2022, the average per capita vaccine uptake was 195.0 doses per 100 people among the 3,075 counties included, with a total population exceeding 330 million (Note 7).

Using the average per capita vaccine uptake score (195.0 doses per 100 people) as an input value in Stata’s margins command algorithm returned a value of 106.8. I.e., average vaccine uptake by the end of 2022 corresponded with 6.8% 2023 all-cause excess mortality. The same exercise, with zero vaccine uptake as the input value, returned a value of 101.6. I.e., counties with estimated zero vaccination by the end of 2022 corresponded with 1.6% 2023 excess mortality (Note 9). These numbers show that counties with average vaccine uptake, corresponding to the overall US vaccine uptake, experienced a 5.12% higher all-cause excess mortality compared to counties estimated to have zero vaccine uptake.

As the US reported 3,090,964 deaths in 2023, a change, that is, an increase, was estimated to be 150,602. I reached that number by subtracting from 1 the ratio between “Mortality if vaccine uptake=0” and “Mortality if vaccine uptake=average”, and then multiplying it by the total US deaths.

Taken together, the average COVID-19 vaccine uptake by the end of 2021 resulted in 150,602 additional deaths in 2023 (Table 1 reports relevant 95% CIs).

Conclusion

CDC data from over 3,000 US counties, with a population of about 330 million people, show that the COVID-19 vaccine in 2022 caused 138,229 deaths, and 150,602 in 2023. Altogether, almost 290,000 deaths. The increase in deaths caused by the vaccine was 4.40% in 2022 and 5.12% in 2023.

In a previous post, I showed that it is unlikely that COVID-19 vaccination reduced US mortality in 2021. Hence, the almost 290,000 net increase in US deaths between 2021 and 2023 was due to the vaccine (Note 1). I plan to report on the 2024 data when they become available.

Notes

1. The arguments used in the previous post concerning non-causal association in 2021 do not apply to 2022 and 2023. The reason is that a detrimental vaccine effect can cause both COVID-19-related and -unrelated deaths. That is why I only applied all-cause excess mortality as the dependent variable in 2022 and 2023. Also, vaccine uptake by the end of 2021 and 2022 was considerably higher than the 1.02% by the end of 2020.

2. The baseline for this study is the county-level deaths in 2018 and 2019, divided by the population size in those years. For example, if 100 people died in a county both years and the population was 10,000 both years, the baseline was (100+100)/(10,000+10,000)=.01. If 120 died in 2022 and the population was the same, the county’s 2022 all-cause excess mortality was 20% (.012/.01=1.2).

3. The CDC does not report data if a county’s deaths are between 1 and 9, coded as missing. A very small number of cases with zero reported deaths were also coded as missing. If data for either 2018 or 2019 were missing, data for both years, used as a baseline, were further coded as missing. The US has a little over 3,200 counties, and this study has baseline data from 3,096 of them.

4. Vaccine data were included from counties reporting positive values on Completeness_pct. To model a proxy for doses per capita, I first summarized the number of doses administered in each county concerning Administered_Dose1_Recip, Series_Complete_Yes (as it typically includes two doses), Booster_Doses, Second_Booster_50Plus, and Bivalent_Booster_5Plus. Next, I divided the number by the population sizes in 2021 and 2022, respectively, and multiplied the result by 100. I.e., per capita vaccine uptake refers to the number of doses administered per 100 people at a given period.

5. Models 1 and 2 report robust standard errors and two-tailed tests of significance concerning the regression coefficients; † p < .10; * p < .05; ** p < .01; *** p< .001. 95% confidence intervals are reported in parentheses.

6. The motive is that the approach “provides a simple way to account for historical factors that cause current differences in the dependent variable that are difficult to account for in other ways”, according to Wooldrige, probably among the most authoritative voices in econometrics today. Including more than one lagged dependent variable, if available, is useful. Issues causing differences in all-cause excess mortality in 2022 or 2023, of which we are unaware, may have also contributed to similar differences in previous years. For example, counties experiencing an abnormally low (high) number of deaths in 2018 and 2019 have led to estimating a relatively high (low) number of excess deaths in the following years (cf. Note 1). Additionally, county-level interventions related to COVID-19 in 2020 may have had lasting effects on mortality over the years. Including all-cause excess mortality in the years preceding 2022 or 2023 as control variables, this approach at least partly controls for other issues we may be unaware of. Another way to explain the inclusion of lagged dependent variables is that it keeps the all-cause excess mortality in years before the estimation year fixed. That is, a model estimates the vaccine effect on all-cause excess mortality in, for instance, 2022, while canceling out differences between counties in 2020 and 2021. It helps isolate the vaccine effect we are interested in. Not doing that would instead increase noise in the data, as counties with high or low all-cause excess mortality in 2020 and 2021 may experience the same in 2022. That would induce overestimations (underestimations) concerning counties with high (low) 2022 values. For a further discussion of the pros of including the lagged dependent variable and cons in a few instances, see, for instance, Rönkkö.

7. The reason we have fewer observations than the 3,096 counties from which we have 2018-2019 baseline data (Note 1) is chiefly due to missing vaccine data (Note 3).

8. Notice that these are the numbers obtained after controlling for the lagged dependent variables, 2020 and 2021 all-cause excess mortality. According to a previous argument, the 2021 lagged dependent variable was positively associated with the 2022 all-cause excess mortality. The effect of the 2020 lagged dependent variable, on the other hand, was non-significant, as the effect was absorbed by the 2021 lagged dependent variable.

9. The negative effect of the 2020 lagged dependent variable is likely due to an oscillatory pattern or mortality deficit, along with other reasons for lagged effects I have discussed (Note 6).