I'm not sure what you think that thread means, @JanglyBeads?
Her first tweet:
A sketch to explain how a new variant may appear milder even with no change in underlying virulence.
The sketches are explaining why. What she is saying is that is that even if a smaller percentage of people become severely ill with omicron, you can't conclude that means it is less virulent than other variants.
Say you have 1000 people exposed to delta. If 500 of them have already had alpha and that stops them being reinfected with delta, only 500 people will be infected with delta.
Of the 500 hundred who are infected:
100 will get severe symptoms
200 will get moderate symptoms
200 with mild symptoms
So 20% (100/500 x 100%) of people infected with delta get severe symptoms.
Now say, you have 1000 people exposed to omicron. Again, 500 of them have had alpha. This time all 1000 people become infected because omicron is better at reinfecting people.
If omicron has the same virulence as delta.
Of the 500 hundred who are infected and have never had the virus before, so they have no immunity:
100 will get severe symptoms
200 will get moderate symptoms
200 with mild symptoms
Of the 500 who had alpha and therefore have some immunity to omicron:
0 will get severe symptoms
0 have moderate symptoms
500 have mild symptoms
So only 10% (100/1000 x 100%) of people infected get severe symptoms. It's not because omicron is less virulent, it's because people with immunity are being reinfected who wouldn't have been reinfected by delta.
Those are totally made up numbers just to explain what her sketches are showing. I'm not sure why she chose to have 0 reinfections in her first example but you can see the principle.
Her last tweet in the thread after explaining her first tweet:
So when I look at data out of South Africa or elsewhere, I am thinking not just about the proportion of cases that are severe, but also the sheer number of severe cases, which demonstrates the public health impact of the variant.
She's making the point that it isn't only the percentage of people who get severe symptoms that can have a serious public health impact, the number of severe cases is an issue. 10% of 10,000 is a lot less than 10% of 1 million.
Furthermore, although she doesn't actually state it, if you have 1000 people exposed to an infectious person but only 500 (50%) become infected, then each of those people infect 50% of their contacts (let's say each peson has 10 contacts. You go from 500 cases to 2500 to 12,500.
If you have 1000 people exposed to an infectious person and all 1000 (100%) become infected, then each of those people infect 100% of their 10 contacts, you go from 1000 cases to 10,000 to 100,000.
20% of 2500 is a lot less than 10% of 100,000.