Data insight

Excess mortality and vaccination

The later part of 2021 showed a negative relation, but only for richer countries

Note:  Charts current as of date above; text updated on January 6, 2021.

The relationship between excess mortality and vaccination appears to have tilted in the later part of 2021. But only for the richer countries which managed to get a large share of their population vaccinated.

The visualization below illustrates the evolving relationship across countries between excess mortality and vaccine coverage. Let’s unpack it:

  • On the horizontal axis, the vaccine coverage ratio: doses administered per 100 people, with all doses converted into 2-dose equivalents (read here why that is necessary). 
  • On the vertical axis, the daily excess mortality rate:  daily 7-day averaged estimates of excess deaths per 100,000 people, where the estimates reflect the mid-point values of the excess death model by The Economist (read more about that here). 
  • Individual countries are represented by dots, where countries are grouped according to World Bank income groups: high-income (HIC), upper-middle-income (UMIC), lower-middle-income (LMIC) and low-income (LIC) countries.
  • The size of each bubble reflects population size (measured by the medium variant of the WPP’s projection for 2021).
  • The line is a linear weighted regression between excess mortality and vaccination, grouped by income category and weighted by population size.

High excess mortality in developing countries

Before we get into discussing the evolving relationship between excess mortality and vaccination, let’s have a look first at the excess mortality estimates shown in the visualization to appreciate how the differ across income groups.

  • Estimated excess death rates have been high in the developing world (the countries that are not high income). In fact, if we were to look at the cumulative excess mortality rate (shown in this post), we would obtain the result that they are the highest in LMICs, then UMICs and then only HICs. For LICs, they have been low throughout relative to the other income groups, but a lot higher than reported COVID-19 deaths still.
  • Turning back to the visualization which shows 7-day daily averages, check out the panel of LMICs and notice the large bubble. That is India. Notice how much it jumps up during the crisis of May 2021, which is entirely unprecedented among populous countries, let alone a country with the population size of India.
  • Notice as well how high the excess mortality rates are for several UMICs, which is primarily driven by the countries in Latin America & the Caribbean. The large bubble here is China, which unlike India is estimated to have registered much lower excess mortality relative to its population. 

A tilting relation between excess mortality and vaccination
 
The visualization traces out how the relationship between excess mortality and vaccination across countries changes over time and how these changes differ across World Bank income groups. 
 

The general pattern is for the relation to be first positive or flat and then turn negative, the latter being suggestive of a beneficial effect of vaccination on excess mortality.  

The full cycle is observed however only in HICs and UMICs (click on the links to get the dynamic visualizations).

  • In HICs, we detect a positive slope until about April 2021 when the relationship starts to flatten. The slope becomes negative around October. This is also the time when excess mortality started to rise again given the onset of wintry weather in most of the Northern Hemisphere (where most of the HICs are located). During that period, however, we notice that excess mortality did not rise as much in the well-vaccinated HICs, whereas in others it did rise more considerably contributing to the negative slope.
  • In UMICs, the flat part of the curve occurred around the same time (around April 2021) but it immediately turned negative. This is the result of China (the large bubble), which managed to keep estimated excess mortality down while rapidly vaccinating a large share of its population. The fact that many UMICs are located in the Southern Hemisphere meant that the excess mortality went up more considerably during the months of July-August 2021, which contributed to the negative slope. 

For LMICs and LICs, we see a different pattern:

  • In LMICs, excess mortality saw a huge spike in May 2021, due to India. Vaccination had by then not advanced sufficiently to prevent the rise in mortality. Later in the year around September, the relationship between excess mortality and vaccination flattened. 
  • In LICs, vaccination progress has been most limited among all country groups. We are therefore not able to detect a noticeable relationship between excess mortality and vaccination. 

Snapshots as of the latest date

In what follows, we show a gallery of the latest snapshots of the relationship between excess mortality and vaccination for each income group. Notice again the differences between HICs and UMICs on the one hand and LMICs and LICs on the other. 

While these cross-country comparisons suggest that vaccine coverage above a certain threshold may impart of a positive effect on excess mortality, there are of course many confounding factors. These include the quality and strictness of containment and suppression policies adopted by a country, a country’s age and health profile as well as the quality of a country’s health system.

Some of these factors are picked up in the variation across income groups, but even so the correlations here should be interpreted with caution. Yet, as tentative as they may be, the results do complement the finding from micro studies that clearly show the beneficial effects of vaccination on mortality.

Conclusion

While these cross-country comparisons suggest that vaccine coverage above a certain threshold may impart of a positive effect on excess mortality, there are of course many confounding factors. These include the quality and strictness of containment and suppression policies adopted by a country, a country’s age and health profile as well as the quality of a country’s health system.

Some of these factors are picked up in the variation across income groups, but even so the correlations here should be interpreted with caution. Yet, as tentative as they may be, the results do complement the finding from micro studies that clearly show the beneficial effects of vaccination on mortality.