Thread by @michaelzlin

Some Omicron trends that we can now definitively see:
• Case hospitalization and fatality rates (CHR, CFR) in SA, Denmark, UK, Australia, US
• A guess at true infection hospitalization and fatality rates (IHR, IFR)
• Omicron peaking in many regions of the US

This follows a previous thread where we inferred from early SA data that the CHR was lower for Omicron than earlier waves due to prior immunity and lower virulence, and it would be similar in the West. (But still bad news due to very high case # s)

https://twitter.com/michaelzlin/status/1471749391585214465

The SA data as of 2021/12/28 showed Omicron CHR and CFR at ~50% and ~15% of the 2021/01 peak.

The CHR is clearly ~50% of the year-ago peak. Deaths however are still creeping up and CFR now 20% of year-ago. Others observed older people got Omicron later in SA; may relate to that.

A 1-month fatality delay was not seen in earlier waves. This could be due to two reasons. First, even if the same age-time shift occurred a year ago, a broader range of ages succumbed due to lack of immunity. The smaller number of delayed cases and deaths in older patients...

... may have just added on to the wave, not be the peak. Second, anecdotal reports say it's Omicron exacerbating preexisting systemic health conditions that are driving ICU admits. Both weak immunity and other issues are much more common in older people.

https://twitter.com/michaelzlin/status/1479350583026806787

Okay let's move on to other countries. Next up is Denmark. Denmark seems to be seeing a double peak in cases. First peak in cases was around Jan 1. That's long enough ago to expect a hospitalization peak, and we see that. Based on peak/peak, CHR is about 20% of last time.

For CFR, if we count current fatalities vs cases a generous 1 month ago, it looks like 0.2%, vs 1% from 1 year ago in Denmark. This is highly speculative; the number can go down if the delay is less than 1mo, or up if fatalities are delayed even more than 1mo from cases.

Onto UK. We'll use London since Omicron hit London earlier than the UK in general. Hospitalizations clearly peaked a week ago. CHR is about 27% of a year ago. Fatalities haven't budged, but I expect fatalities will go up too eventually.

Vent use is only up a little. That's consistent with patients being ICU'ed not primarily for respiratory failure but for other issues exacerbated by Omicron.

But another possibility is that most vulnerable may already have died in a previous wave. UK was among the hardest hit.

Onto Australia. At first you wonder how they avoided a Jan 2021 peak, then you realize Jan is summer in Australia. So their winter peak began in June and peaked in September. Those Aussies, always doing things upside-down.

Australia cases haven't peaked but we can get a speculative CHR by using today's hospitalizations and cases from 10 days ago on 2021/1/1 as the denominator, and comparing to the Sept peak. But in 2021/9 many people were vaxxed, not sure if that was reducing CHR already.

In any case, we get 2.7% very rough CHR vs 5.2% CHR in Sept, so that's about a 50% reduction. Again Sept CHR may be lowered by vaccination; Australia was up to 30% already and the most vulnerable got it first.

Too early to compute CFR in Australia.

Finally the US. We'll look at NY as it's peaking early and its CFR and IFR have been pretty representative of the US in general.

NY cases peaked 1/8; hospitalizations 1/7. Case peak is biased by a recent wave of testing. Actual infection peak likely ~1/1 when testing was low.

In any case, NY CHR is about 28% of the year-ago peak. Fatalities are going up and that's ominous. If we use the SA experience and assign deaths to 1mo-ago cases, then you get a similar CFR to the year-ago peak.

Let's hope the delay isn't so long and deaths stop going up soon.

So the overall picture we get so far is this: Omicron case-hospitalization rate (CHR) ranges from 20% to 50% of a year ago, when most cases were among the non-immune. Clearly immunity is playing a huge role here; cases are now mostly in vaccinated and these are mild.

CFRs range much more widely, and it's not possible to conclude anything solid yet. All fatality trends are still going up and it's hard to know what time period to use for the case denominator. Omicron causing deaths partly by exacerbating other conditions complicates things.

However there is one other major factor that we've left out of the discussion so far that is a cause for optimism, and it is the relative degree of undercounting of Omicron cases vs previous cases. I think we are now undercounting 2.5x relative to previous peaks.

You expect more undercounting this time. This wave is faster and bigger so testing lines are longer, people expect milder disease so there's less incentive to test, and many are testing at home and not reporting positives.

And if you look at wastewater measurements, Omicron peaked this week at 10x relative to year-ago, but the case numbers are only 4x higher. This suggests 2.5x more undercounting compared to before.

Also discussed at boston.com/news/coronavir…

If those wastewater measurements are representative of other locations (and fairly represent Omicron levels) then the actual infection hospitalization rate (IHR) may not be 4x lower than the previous peak (as measured for CHR), but 10x lower.

Presumably then the actual infection fatality rate (IFR) would at least follow the IHR and then also be 10x lower. Since SARSCoV2 IFR was ~1% in the US by most estimates, future IFR, if our vaccinated state and future variants are no worse than Omicron, could be 0.1%.

That would be good news as 0.1% is similar to flu. But 3 caveats:
1. To maintain that rate, we'd likely need annual boosters.
2. We must avoid more immunoevasive or virulent strains, something that molnupiravir abuse may make more likely.

https://twitter.com/michaelzlin/status/1475518939413184518

3. The fatality trend in NY being higher than other countries is cause for concern and could refute this optimistic outlook if it continues.

Will keep a close watch on that, and meanwhile if you know of other wastewater data feel free to share.

But in a sense this is how we expected things to go all along. If vax or prior infection prevent deaths by 90% then IFR should drop from 1% to 0.1% once at least the high-risk population is vaccinated or infected. In the West we got there mostly with vax and some past infection.

The @washingtonpost had an article making the same points for US and SA. So far US overall is showing CHR at 33% of the rate of 1 year ago.
washingtonpost.com/health/interac…