"The Infection Fatality Ratio (IFR) of COVID-19 is very low - 99.5% of people survive it. The entire world is drastically overreacting to what doesn't kill the vast, vast majority."

1. It's still a lot of death. Let's assume the "99.5% survival rate" number is correct (although it's not - see point 5 below). This translates to an Infection Fatality Rate (IFR) of 0.5%. If we just didn't suppress the virus, this would kill up to 300,000 people in a country the size of the UK, or up to 1.55 million in the US. Here's how we worked this out:

Because of specific characteristics of the coronavirus, the infection spreads far more rapidly than a typical influenza virus. It would infect—and kill—far more people if no interventions were in place. The total impact of the disease, if no action was taken, is known as the final attack rate of the virus, and is calculated from the R0 number, which is the number of further virus cases you'd expect to be spread by a single infected person.

As Covid is estimated to have an R0 between 2.5–3.5, the final attack rate is estimated at around 80-95% - that's the proportion of the population that would be infected if nothing was done to curb its spread. It would result in 0.4–0.475% of the entire population dying from the disease using an IFR estimate of 0.5% (266,000 to 316,000 people in the United Kingdom, 1.31 to 1.55 million people in the United States). By the way, this would make it around 66-175 times more deadly than, say, the 2009 H1N1 (swine flu) pandemic, which was estimated to have an IFR of 0.0076% and a final attack rate of 30-80%.

1. The IFR is kept low because the disease is being suppressed. At the moment, most people who catch Covid and become seriously ill are treated in hospital. That keeps the fatality rate low - lower than it would be if fewer of those people received adequate treatment. Without efforts to suppress the virus—if the virus was allowed to run rampant—hospitals would quickly become full, and many of those who currently survive a Covid infection would not receive treatment and would die. This means that the IFR has a kind of "observer effect" - it's dependent on the sorts of measures society takes to suppress the virus.

2. A single IFR number is oversimplified. Again, let's assume the 99.5% survival rate number is correct, so the Infection Fatality Ratio is 0.5%. That's an average, but we know that the survival rate differs dramatically by age. If you're under 35, the survival rate is more like 99.996% (though see point 3 below for an important point about this). But if you're over 85, the survival rate is only around 72% - that is, 28% of people in this age bracket will die of the disease if they catch it. When people talk about the "99.5% survival rate", they mean "on average" - and that average helps them forget the dramatically higher fatality rate in older people. Even for the not-so-elderly—for example, 60-year-olds with on average 24 years of life ahead of them—their Covid fatality rate, which is 1%, is more than double their usual annual chance of dying.

3. Death isn't the only thing that matters. By now there are many credible reports of long-term after-effects of Covid-19: even if it doesn't kill you, it could still affect your heart, lungs, or even your brain. There's even a name for the longer-term conditions: "Long Covid". And that's leaving aside the experience of getting the disease itself, which for many people ranges from an extremely unpleasant few days or weeks to a harrowing experience on a ventilator in an intensive care unit. Focusing only on deaths ignores all of the additional suffering Covid causes - even to the young.

4. It's not the right number anyway. Certain researchers have underestimated the IFR of Covid-19, mainly because they didn't include all the relevant data in their studies. The initial "seroprevalence" studies were known to have serious problems with bias: they didn't collect random samples, and instead asked people more likely to have had the disease to take part. This made it seem like the disease was far more prevalent than it really was, and led to an underestimate of how many people the disease killed (it wrongly increased the denominator in the "no. of deaths divided by no. of infections" calculation). Better estimates report that the IFR is more like 0.68%, or a 99.33% survival rate, if you prefer (though remember, as per point 1, above, this translates to a lot of death). In any case, as we've seen, it's questionable how useful a single IFR estimate is, anyway.

5. In some circumstances, the IFR will be even higher than this. This is an add-on to point 2, above. Because the IFR varies so much by age, the number of elderly people in a population and their likelihood of catching the virus can affect the ratio significantly. The 0.5% estimate we're discussing here assumes a globally average population structure and assumes that young people are more likely to catch the virus. In countries with more elderly people, like the U.K. or where elderly people catch it at similar rates to younger people—for example, in situations of unmitigated spread—the IFR will likely be even higher.

Page added on 19 January 2021