A few weeks back I wrote an article about how high the risk of dying from covid is. I mentioned that a senior representative of the World Health Organization (WHO) had recently said that the WHO’s best estimate was that roughly one in 750 people who get infected die of the disease. I also mentioned a study published by the WHO, authored by professor John Ioannidis at Stanford University, which was based on antibody data. That study estimated that the mortality rate for covid was around 0,23% overall, which would mean that roughly one in 430 people who are infected overall die of the disease, and 0,05% for people under 70, which would mean that if you’re under 70, the risk of dying of covid is about one in 2,000.
Since then, professor Ioannidis has updated his figures. The newer numbers have been published in The European Journal of Clinical Investigation. The modifications have been made to compensate for the fact that the earlier estimates were extrapolated from the countries that were hardest hit by covid. When this is accounted for, the new estimate is that covid kills around 0,15-0,20% of those infected, so around one in 600 infected people die of the disease overall. Among people under 70 years of age, the revised estimate is that 0,03-0,04% die, which is around one in 3,000.
However, professor Ioannidis also mentions that the fatality rate varies a lot between countries, related to varying levels of risk factors. As I mentioned in a previous article, the main risk factor for dying of covid is obesity. So countries with high levels of obesity will be hit harder than countries with low levels, which likely explains why the US has been hit so much harder by covid than Japan. Other health related factors that increase the risk of dying of covid are high age, organ transplantation, uncontrolled diabetes, chronic obstructive pulmonary disease, liver failure, kidney failure, and cancer. Basically, the things that predispose you to dying in the near future more generally, also predispose you to dying of covid.
In his article, professor Ioannidis mentions attempts that have been made to estimate the number of years of life lost, on average, when someone dies of covid, and I think this is interesting to explore further, because it is actually extremely important when trying to determine how severe covid is. When a small child dies, for whatever reason, that generally means around 80 years of life are lost. If a 90-year old dies, for whatever reason, that usually means at most a few years of life have been lost. Most people therefore reasonably think it’s much more tragic when a small child dies than when a very old person dies, because much more potential lifetime has been lost.
So, if covid results in 20 years of life lost on average, that’s reasonably about 20 times worse than if it results in one year of life lost on average. And that’s why it’s important to know how many years of life are lost to covid, when someone succumbs to the disease.
So, how many years are lost?
An article was published in The Proceedings of the Nationol Academy of Science (PNAS) in July that sought to estimate this. Using cohort life tables (the tables that insurance companies use to predict how many years of life someone has left based on how old they are), they calculated that the average person who dies of covid loses 12 years of life!
To me, this number seems implausibly high, because it doesn’t match what I am seeing here in Sweden. Half of Swedish covid deaths happened in nursing homes, where median life expectancy is less than a year. If half of all people who died of covid in Sweden would have been dead within a year even without covid, that would mean that the other half who died would have had to have twenty plus years of life left, in order for the average to end up being 12 years.
Considering that the average age of those who have died of covid in Sweden is 84, while the average age of death in Sweden more generally is 82, that seems extremely unlikely. Just looking at the data from what has actually happened in Sweden, it seems more likely that the average amount of lifetime lost to covid is very low, a few years at most.
In fact, I would go so far as to venture that covid can not possibly have resulted in an average loss of 12 years of life per person who dies, based on what the real world numbers actually show. So, how could the authors of the article in PNAS get the numbers so wrong?
Well, there is one thing that they should certainly have done, which they didn’t do. They didn’t take co-morbid conditions in to account. An 82 year old with type 2 diabetes, heart failure, chronic obstructive pulmonary disease and high blood pressure has a much shorter life expectancy than an 82-year old without any underlying conditions.
And we know that most of the people who die of covid have multiple underlying conditions. According to official data from the US Centers for Disease Control (CDC), 94% of people who have so far died of covid in the US had at least one underlying condition, and the average person who died of covid had three underlying conditions. You would think that the authors of the article in PNAS would have taken this factor in to account, since the number of underlying conditions a person has makes a big difference to how much longer they can expect to live.
A separate study was published in Wellcome Open Research in April that did try to correct for co-morbidities. Somehow, even when factoring in co-morbidities, this study still managed to arrive at an average of 12 years of life lost per person dying of covid. In other words, factoring in co-morbidities made zero difference to the projected years of life lost. Very strange.
If that was correct, it would mean that the average person dying of covid in Sweden, being 84 years old, would have lived to 96 if covid hadn’t happened. That is in spite of the fact that this average person has multiple underlying co-morbidities, and also in spite of the fact that the average 84 year old in Sweden can only expect to live seven more years, and will on average die around the age of 91.
So, if the authors of the study are right, this would mean that the average person dying of covid is healthier than the average person, since the average person dying of covid has 12 years of life left, and the average 84-year old only has seven years of life left. But we know the exact opposite is true – the people dying of covid are in general significantly less healthy than the average person. Admittedly, the data used in the study are taken from Italy and the UK, and I’m extrapolating to Sweden, but I find it implausible that the difference could possibly be that big.
So, something is fishy about the numbers. What?
Actually multiple things.
First, the authors only include eleven specific co-morbidities in their analysis, which means that all other co-morbidities that could affect years of life lost are excluded. As an example, cystic fibrosis is excluded from the analysis. The average life expectancy of a person with cystic fibrosis is 44 years. If a 40 year old with cystic fibrosis catches covid and dies, they would be considered completely healthy in the modeling done in this study, and would contribute around 45 years of life lost, when their real life expectancy is much lower.
Obesity, widely recognized as the biggest risk factor for severe covid, is not included in the modeling (probably because the authors didn’t consider it a “co-morbidity”), which means that people with obesity are considered to have the same life expectancy as non-obese people. Ignoring risk factors in this way will give the impression that people who die of covid are healthier than they really are, which will in turn lead to an over-estimation of their remaining years of life.
Second, the severity of co-morbidities is not factored in. Someone with end-stage chronic obstructive pulmonary disease (COPD) has a much shorter life expectancy than someone with mild COPD, and is probably much more likely to die if they catch covid, but this was also ignored in the modeling.
So, although the modeling in this second study attempted to factor in co-morbidities, it did so in an incomplete way, which likely resulted in a big overestimation of the number of years of life lost.
In the sub-group analysis, where the authors divide things up by age and number of co-morbidities, they find that a person age 80+ with three co-morbidities likely loses 6-7 years of life if they die of covid. Considering that the average person dying of covid is 80+ and has three co-morbidities, this seems like a much more reasonable number than the twelve years presented above. But it still seems high when compared with the real world data we have at this point in the pandemic.
Let’s say this number is right though, and the average person who dies of covid loses 6-7 years of life. Considering that roughly 1,4 million people have now died of covid, that would mean around 9 million years of life have been lost to covid so far.
To gain some perspective on the issue, let’s talk about measles. Measles is a disease that mostly kills children under the age of five. So, whenever someone dies of measles, the average number of life years lost is around 80. As a result of the global obsession with covid this year, measles vaccination programs have been paused in 26 countries. That means 94 million children are at risk of not getting their measles vaccines. It will only take about 110,000 children dying of measles because they didn’t get their vaccine, for the total number of life years lost to be the same as those lost due to covid.
And measles is just one disease. There’s also polio, tetanus, and yellow fever, to name a few. These are diseases that maim and kill children, but which can easily be prevented with vaccinations. How many years of life will be lost in total due to lockdown and to the world’s single-minded focus on covid this year?
Probably many times more than are lost to covid directly.
You might also be interested in my article about the Swedish response to covid, or my article about whether or not long covid is really a new disease.