This is magnificent collection of things to think about, thank you. I cannot quite grasp why you think SA data is so incompatible with the rest, though. Here I do not mean cited paper, just raw data showing wave (possibly!) peaking without huge numbers of dead people.
Let´s say that omicron is 40 % less virulent than delta for someone unvaxed and without previous infection. Then let´s assume that previous infection, while not providing as much protection from reinfection by omicron, still provides good protection from hospitalization. Also, let´s assume South Africa had giant numbers of unrecorded previous infections, as many people think, with good reasons. Still doesn´t add up?
The NSW evidence is pushing me somewhat toward the conclusion of innate mildness, but I'm a bit unclear on some of the others. The latest Imperial College result of 40% reduced severity might be controlling for as much as possible, but...is controlling for everything the same as reducing the result to the innate virulence? I'm not sure it is. To what hypothetical person or population of people does the 40% reduction apply?
I might be confused about what "controlling" for variables actually means in practice, I really should look up the maths of how this works at some point. But it seems that in world where innate virulence were the same, you'd expect unvaccinated/not-previously-infected cases to have equal hospitalisation rates for omicron vs delta, and you'd expect vaccinated/previously-infected to have a lower hospitalisation rate per case for omicron vs delta. Which group are you referring to when you give the result of the study? The imperial study says the 40% figure is "averaging over all cases in the study period". But for innateness we don't want that, we want to look at the unvaccinated/not-previously infected only.
So your judgement on "Chance that Omicron is vastly (75%+ in the same person) less virulent than Delta" is ambiguous. Which person? If a vaccinated/previously-infected person, then conditional on infection, evidence is strong that you expect omicron to be milder because of the immune escape causing more breakthrough infections for omicron than delta - whether it's milder intrinsically or not. If unvaccinated/not-previously-infected - then we're looking at the innate virulence, and the evidence there is weaker. Some seem to be interpreting results like the Imperial study as applying to this, but I'm not sure that's right.
While you’re refining the questions, you might want to be more specific about who you have in mind for “boosters modified for Omicron within 6 months of our previous booster shot.” I got my booster back in August, so if they wanted to get me one in this time frame they’d only have 2 months left to do it.
"If severe disease risk among the hospitalized was the same, I’m not sure why the 80% becomes 70%? Are some severe cases not getting hospitalized?"
The study says "Among hospitalised individuals, after controlling for factors associated with severe disease... compared to earlier Delta infections, after controlling for factors associated with severe disease, SGTF-infected individuals had a lower odds of severe disease (aOR 0.3, 95% CI 0.2-0.5)."
So the 70% reduction in severe cases they are citing is *conditional* on hospitalisation. So an 80% reduction in hospitalisation, and then a 70% reduction in severe on top of that. (Of course, the 70% reduction is the one they flag as pretty unclear because it's "compared to Delta previously" whereas when they "control for time" / match the time periods they only get a 30% decline).
2. For the NSW data, the 45% decline is a lower bound (on the point estimate, at least). The aggregate CHR is down ~45% (with appropriate time lags). But some of the cases (and a decent chunk of the hospitalisations) are presumably Delta still - the case numbers of Delta had been flat leading up to Omicron arriving, and the rest of Lilley's thread uses that fact to try to estimate the decline for Omicron formally. https://twitter.com/andrewlilley_au/status/1473098231412404225/photo/1
I'm glad you and your family are feeling better!
This is magnificent collection of things to think about, thank you. I cannot quite grasp why you think SA data is so incompatible with the rest, though. Here I do not mean cited paper, just raw data showing wave (possibly!) peaking without huge numbers of dead people.
Let´s say that omicron is 40 % less virulent than delta for someone unvaxed and without previous infection. Then let´s assume that previous infection, while not providing as much protection from reinfection by omicron, still provides good protection from hospitalization. Also, let´s assume South Africa had giant numbers of unrecorded previous infections, as many people think, with good reasons. Still doesn´t add up?
The NSW evidence is pushing me somewhat toward the conclusion of innate mildness, but I'm a bit unclear on some of the others. The latest Imperial College result of 40% reduced severity might be controlling for as much as possible, but...is controlling for everything the same as reducing the result to the innate virulence? I'm not sure it is. To what hypothetical person or population of people does the 40% reduction apply?
I might be confused about what "controlling" for variables actually means in practice, I really should look up the maths of how this works at some point. But it seems that in world where innate virulence were the same, you'd expect unvaccinated/not-previously-infected cases to have equal hospitalisation rates for omicron vs delta, and you'd expect vaccinated/previously-infected to have a lower hospitalisation rate per case for omicron vs delta. Which group are you referring to when you give the result of the study? The imperial study says the 40% figure is "averaging over all cases in the study period". But for innateness we don't want that, we want to look at the unvaccinated/not-previously infected only.
So your judgement on "Chance that Omicron is vastly (75%+ in the same person) less virulent than Delta" is ambiguous. Which person? If a vaccinated/previously-infected person, then conditional on infection, evidence is strong that you expect omicron to be milder because of the immune escape causing more breakthrough infections for omicron than delta - whether it's milder intrinsically or not. If unvaccinated/not-previously-infected - then we're looking at the innate virulence, and the evidence there is weaker. Some seem to be interpreting results like the Imperial study as applying to this, but I'm not sure that's right.
Glad to hear you're feeling better! Happy Holidays!
While you’re refining the questions, you might want to be more specific about who you have in mind for “boosters modified for Omicron within 6 months of our previous booster shot.” I got my booster back in August, so if they wanted to get me one in this time frame they’d only have 2 months left to do it.
1. Regarding the South Africa study, you write:
"If severe disease risk among the hospitalized was the same, I’m not sure why the 80% becomes 70%? Are some severe cases not getting hospitalized?"
The study says "Among hospitalised individuals, after controlling for factors associated with severe disease... compared to earlier Delta infections, after controlling for factors associated with severe disease, SGTF-infected individuals had a lower odds of severe disease (aOR 0.3, 95% CI 0.2-0.5)."
So the 70% reduction in severe cases they are citing is *conditional* on hospitalisation. So an 80% reduction in hospitalisation, and then a 70% reduction in severe on top of that. (Of course, the 70% reduction is the one they flag as pretty unclear because it's "compared to Delta previously" whereas when they "control for time" / match the time periods they only get a 30% decline).
2. For the NSW data, the 45% decline is a lower bound (on the point estimate, at least). The aggregate CHR is down ~45% (with appropriate time lags). But some of the cases (and a decent chunk of the hospitalisations) are presumably Delta still - the case numbers of Delta had been flat leading up to Omicron arriving, and the rest of Lilley's thread uses that fact to try to estimate the decline for Omicron formally. https://twitter.com/andrewlilley_au/status/1473098231412404225/photo/1