We now have fatality data for 2021 which should be mostly complete, and the results look very similar to 2020.  Once again, fatalities were about a half million more than would be expected based on pre-pandemic experience.

Prior to the pandemic, there were about 2.8 million fatalities each year in the US.  For many years, fatalities have grown consistently at a little over 1% each year.  There were slight variations, mostly based on the severity of the flu seasons.

But in 2020, fatalities jumped by nearly 20% to just under 3.4 million.  In 2021, fatalities increased to a little over 3.4 million, which is about 535,000 above the number we would have had if fatalities had continued to grow at 1% after 2019.

According to the CDC, about 460,000 people died in 2021 after being diagnosed with COVID.  So, we had about 75,000 more deaths than expected, which apparently were not related to COVID.  In 2020, the number of excess fatalities not attributed to COVID was about double that at 150,000.  I suspect that is mostly due to how little testing there was in the early days of the pandemic and that many fatalities related to COVID were not diagnosed.

But there is clearly something else going on in addition to COVID’s toll.  There has been a lot of speculation on this issue including increases in suicides, overdoses and the postponement of routine screenings for other diseases like cancer.  But we really do not have very good statistics on any of those yet.

“Mortality Displacement” & “Years of Life Lost”

There are two aspects of the pandemic that will ultimately affect our understanding of the fatality toll of COVID.

The first is mortality displacement.  Mortality displacement is an epidemiological/demographic concept that describes the phenomenon where an extraordinary event delays or accelerates fatalities that would otherwise have occurred in another time period.

Because the vast majority of COVID victims have been elderly and/or ill, there likely has been a significant displacement of fatalities that would have otherwise occurred over the next several years into 2020 and 2021.  This is particularly likely with regard to the fatalities which occurred in nursing homes.  The average life expectancy in a skilled nursing facility is less than two years.

Also, as you can see from the chart, there was an unusually small increase in 2019 from 2018.  Therefore, we probably went into 2020 with a larger amount of more vulnerable cohort than normal.

Last year I thought we might start seeing some effect of displacement in 2021.  But Delta and Omicron crushed that hope.  Also, excess deaths from Omicron spilled over into 2022 as well.  The peak of COVID fatalities probably occurred around the end of January but have been rapidly declining since then.  Once Omicron has fully abated, it will be interesting to compare each month’s total to previous years.  I still believe we are going to see some significant displacement at some point.

Second, epidemiologists and actuaries distinguish between “lives lost” and “years of life lost.”  Years of life lost is a calculation which compares the age of the decedent to their life expectancy.  For example, if an 84-year-old person died of COVID, the assumption would be that they lost 6.1 years of life because that is the life expectancy of the average 84-year-old person based on the currently life tables.  Conversely, the life expectancy of a 5-year-old is 79.8 years, so a death at that age results in years of life lost equal to that amount.

Again, because COVID fatalities have been very concentrated among the elderly, its toll in terms of years of life lost has been relatively less than other pandemics.  Most pandemics, and especially flu pandemics, tend to have a high fatality rate among children, as well as the elderly.  Thankfully, COVID almost entirely spared children.  So, we would expect the years-of-life-lost to be something substantially lower than the absolute increase in fatalities.

A team at Northwest University recently looked at this issue in a study of about 2 million patients in the Midwest over the course of the pandemic.  The study found that the fatality rate among their entire population studied was .23% (23/100 of one percent).  There is far more information in the study than I can share here.  But to me, the most significant takeaway was the stark difference between the effect on the overall population and the effect on specific individuals.

This is best seen in a comparison of the years of –life lost suffered at a population level compared to the years of life lost by individual decedents.  The study found that the mean of the years of life lost by everyone who died from COVID was 13.2 years.¹  However, the life expectancy loss for that entire group was only 11 days.  That startling difference suggests is that the vast majority of the population was not particularly impacted by COVID but a very small percentage were dramatically and catastrophically affected.

I think this is one of the reasons that public health officials and the medical community had such difficulty communicating with the public about COVID.  They were seeing that a relatively small part of the population (like the .23% in the study), which was dramatically affected by COVID, while the rest of us were mostly seeing people getting over it in a few days or someone who was already very old and/or ill succumb to it.  The two worlds were living in different realities.

There has been an almost ubiquitous media narrative that many of the COVID deaths could have been avoided had we just pursued the right public health policy.  This recent op-ed in the New York Times is a prime example.  This narrative almost always relies on comparing the US to other countries that either have highly unreliable fatality data or very different circumstances relating to epidemic outbreaks.  I think most of the differences in fatalities between geographic areas and jurisdictions can be adequately explained by differences in demographics, overall population health and reporting systems disparities.

Other than the unprecedented development vaccines during the pandemic phase and some improved  treatments and therapeutics as the pandemic progressed, I am pretty skeptical that anything was done, could have been done, that would have made a significant difference in the ultimate death toll.  I know that notion unsettles our vaulted anthropocentric view of the world.  But the reality is that there is still much in world beyond our control.

Note 1 –  It is highly doubtful that the average years of life lost for COVID decedents was really 13 years.  That is because most COVID decedents were not in “average” health.  The CDC has consistently found that over 90% of those who died from COVID had multiple co-morbidities.  If it were possible to look at each COVID decedent individually and calculate a life expectancy base on their clinical condition, I suspect the years –of –life lost would be dramatically lower.

‍