Study: Omicron 40-45% less likely than Delta to lead to 1+ day hospitalisation | Hacker News Reader

Study: Omicron 40-45% less likely than Delta to lead to 1+ day hospitalisation

imperial.ac.uk

67 pointsflipbrad55 comments

The hospitalizations in South Africa went from 19% to 1.7% but trust Imperial (whose highly inaccurate lockdown modelling last year caused who knows how much damage) to produce a study that gives significantly less favourable numbers. If anyone has the patience to read it, it would be interesting to know what statistical/definitional trickery they've used to make omicron maximally menacing.

How is 40% less likely to lead to 1+ day hospitalization “maximally menacing”?

I haven’t read this imperial college paper yet, but yesterdays preprint from SA has it at 80% less likely to lead to hospitalization, with some uncertainty as to whether that’s due to intrinsic lower virulence vs underestimated prior immunity

https://doi.org/10.1101/2021.12.21.21268116

Edit: read through a good portion of it, the methodology seems pretty reasonable, including definitions for infection/reinfection, inclusion criteria, the covariates included in the model. I’d say the biggest uncertainty factors are (1) it’s still early and there aren’t that many hospitalizations yet (the study period ended like a week ago) and (2) for their correction accounting for undercounted prior infections, they assume undercounting of 1/3 based on some other papers. Maybe that’s reasonable, I’m not that familiar with testing in the UK. But they report both corrected and incorrect results

Unfortunately, Imperial's predictions for the UK government have become politicised, so it can be difficult to discuss them rationally.

I'd be pretty concerned at the really small date range for this study, but otherwise it looks reasonable. The corrections of hazard ratios is the part that might, if you needed to, be hiding some trickery.

If Omicron turns out to have better health outcomes than Delta, even with the increased transmissibility, will we see countries welcoming Omicron-infected travelers so that it might spread and out-compete Delta?

Omicron is coming either way.

You forget the biggest issue with a highly-contagious RNA virus - the more people it infects, the more mutations we'll see!

True, but everyone's going to get exposed to Omicron anyway, unless something more even transmissible takes it's place. So, why not skip straight to Omicron and maybe have better health outcomes by eliminating Delta?

Because this means tens of millions of more dead bodies. So far, we have less than 300M infected on record with 5.3M dead. The real numbers are probably 5 times higher, i.e. a lot of people will die. Also, with the immune escape, those 300M are not greatly protected either.

I’m talking about a scenario where Omicron is strictly “better” than Delta. Why would that necessarily lead to more dead people than we’d get with Delta?

If it's faster spreading, it could "overwhelm the healthcare system" in areas where Delta would have been manageable.

But again, that’s not the scenario I’m talking about

> If Omicron turns out to have better health outcomes than Delta, even with the increased transmissibility

The problem is that sentence is still too vague. If you make a more quantitative statement it would be easier to engage with.

Requiring hospitalization but 10X higher probability compared to delta for surviving hospitalization can be described as "to have better health outcomes than Delta". But if transmissibility is higher it can still overwhelm the hospital infrastructure and cause more deaths by making beds a scarce commodity not only for those affected by covid but also others who need a hospitalization bed.

Note that a multiplicative factor reduction in the probability of surviving, becomes just a time shift during exponential growth phase of the spread. With a doubling rate of 2 days, a 1/16 factor is a delay of 8 days. It shifts the curve, does not flatten it, at least not at the off-peak regime.

Does anyone actually still buy into the "overwhelm the health system" narrative? I mean, that was the original reasoning for initiating lockdowns (two weeks to stop the spread, flatten the curve) and it's repeatedly been threatened but never actually come to fruition. Indeed, healthcare workers are busy, dedicated and hard-working, and I'm sure they would like Covid to end, but the idea that scores of people are dying untreated because hospitals are overwhelmed with Covid cases seems not to have ever materialized. (Except maybe NY under Cuomo's watch)

Anecdote and all, but... Look, my family members work in healthcare all over the country. There is no "overwhelm the the healthcare system" crap. There is, however, locked down departments, shut down services, and reduced amount of surgeries. Those are due to policies, and have nothing to do with COVID-19.

>Because this means tens of millions of more dead bodies.

That implies there are tens of million dead bodies already but there aren't. There are maybe a few millions at most. From which only a part it accrual excess death and from that only a part is excess death that actually reduced average age below life expectancy by a signification margin i.e. shortened the life by several month. If the virus would go away now we should see less death in the coming years which further diminishes the actual effect.

Just to put in in perspective there may be 5 million COVID associated deaths in a time frame (~2 years) where 120 million death are expected And 280 million new people are expect to have been born (unclear if COVID has a positive or negative effect on this number). Needless to say COVID does not even make a dent in the growth of the global population. It may slow growth down by like 1-2% that means 2 years of COVID made the global population lag behind is a few days compared to no COVID.

Excess deaths estimates are in excess of 10million. Ie totals of 18.4 with 95 confidence..

https://www.economist.com/graphic-detail/coronavirus-excess-...

Thats still not "tens of millions" and a huge portion of these will be caused by the measurement done in the name of fighting COVID.

Either way smoking has killed more people each year for many years now than COVID + all pandemic related excess death in one year together. That should give a context of how minuscule it is on a global scale. Like we are still an order of magnitudes away from making the global population stop growing.

If you know a person who died due to COVID and everyone you know also know a person who died due to COVID then everyone subjectively things many many people died but most people know 100+ people and every person who dies was know by many people so in reality only a fraction of a percent of all people would need to die for this to be statically true.

> Because this means tens of millions of more dead bodies.

What gives? Many countries have >80% vaccination rate (doubled vaccinated). Actually, where are we expecting these dead bodies to be, and why? If we answer that, maybe we can solve it.

I can only admire your Christmas spirit!

Huh? Sigh... I suppose you misunderstood it.

Wouldn't there be fewer mutation if everyone gets infected fast with a similar virus? Versus mutations building up while people are infected slowly.

Basically a wide tree of infections vs. a deep one.

Mutations happen all the time at each replication - within a single individual as well; they just get propagated with every newly infected person.

So wouldn't they build up less if everyone is infected with the same strain? And not the strain after 100 previous infections.

Yes because if everyone has variant A and variant A provides some kind of immunity to variant B then there will be less reproduction of B and therefore the chance of mutation becomes lower even if B infect everyone at the same time too.

But this is only true if you expect both A and B to reach everyone eventually. Currently we expect that a lot people never had any variant and therefore no mutation could happen in these people. But if they get it all eventually then it just delays mutation.

No.

Yes

Omicron will become the dominate strain worldwide without anything so insane as that

Why is it insane?

deliberately encouraging infection is almost a kind of murder...

Depends. I know coughs can kill, but it is pretty unlikely.

That's morally indefensible. Millions of people worldwide will die from omicron as you reach herd immunity. No one can deliberately choose to do that.

I’m not talking about trying to achieve herd immunity, though. I’m wondering if countries (maybe just the well-vaccinated ones?) might prefer Omicron to Delta given that they’ve all got Delta and are all going to get Omicron eventually.

The big scary omnicron boogie monster is coming and you can’t stop it from coming. We can fight it or

Why not let it come, especially if your country has high vaccination rates, so it spreads rather than the delta variant.

Or do what I do shrug and assume some vaccinated person will have almost no symptoms and they will spread it to the unvaccinated like me, and then I’ll either get a cold or die.

At least us unvaccinated are more likely to know we have the sickness and will stay home. All these vaccinated folk probably have it and are just spreading it because the vaccine protects them from symptoms.

Lol - I’m probably not going to be able to sell the story that the vaccinated are the dirty ones huh?

So if everyone was vaccinated almost no one would die? What’s your point?

That we don’t need mask mandates, lockdowns, vaccine mandates, or any restrictions to anything regarding covid, other than maybe an insistence that you stay home if you feel sick. (And sure because flying already sucks require a covid test and a flu test to fly)

This may not mean Omicron is less severe itself. It may just be indicative of 70-80% of populations being prior infected or vaccinated like South Africa was. Unvaccinated individuals still have a tough time clearing this.

Have you seen any solid data about how unvaccinated individuals fare with omicron? I have been curious and it’s been hard to come by.

This preprint from yesterday suggests basically what you’re saying, that underestimated prior immunity could possibly be a confounding factor in the models used to assess relative severity. But even that is still unclear at present

https://doi.org/10.1101/2021.12.21.21268116

Edit: table 3 in the imperial college preprint addresses this, they estimate hazard ratio of like 0.6-0.7 for unvaccinated, relative to delta

That's exactly the conflating factor that worries me. We know Omicron is much better at evading prior immunity, so we should expect a higher proportion of people with prior immunity among Omicron infections, which (all other things being equal) will make Omicron look less severe. But this would be illusory - the absolute number of severe infections wouldn't be less, they would just be diluted by mild infections that otherwise wouldn't have happened at all.

What I was taught in Virology 575 at UW-Madison: new viruses face intense selective pressure, and this tends to make them less pathogenic over time as they make their way through the new host population.

That seems to make sense.

Except that you forgot the key element - this is in the long run and not so applicable to RNA viruses! A lot of people will die until covid becomes the new seasonal flu!

> and not so applicable to RNA viruses

Are you sure about that? RdRps have quite the error rate, which means faster evolution. (And yes, I'm aware that the error rate in these vary based on the RdRp's sequence, but in either case, these are both higher than DNA transcription.)

I do agree with the rest of your comment. Go get vaxxed, people!

Many will die long after people forget about 2020 and 2021.

Not much we can do to stop that.

A lot of people will continue to die until we get a robust vaccine.

Yea and you'll hear the guys on TWiV saying it too. But it's not quite that simple. It works out that way in the short term if the virus kills you right away, but there is not quite as much evolutionary pressure if the virus spreads during its presymptomatic period, as is the case with SARS-CoV-2.

And there's the fact that Omicron mutated a whole lot without spreading (it doesn't share a lineage with Delta) which makes me think that this naive form of evolutionary pressure cannot be responsible for its reduced lethality.

Sars-Cov2 will happily live in animals without killing them but will still kill humans.

> not quite as much evolutionary pressure if the virus spreads during its presymptomatic period

That isn’t the only form of evolutionary pressure. The more your host population builds up immunity to the viral bag of tricks the more evolutionary trade offs there must be to remain infectious and over time they tend to harm the virulence. For example, Omicron seems to be very good at replicating in the upper respiratory but not so good in the lungs. Destructive mutations can still occur but they become much less probable over time.

> Omicron mutated a whole lot without spreading

One theory is that it did mutate through spread but in mice not humans. [1]

1: https://www.biorxiv.org/content/10.1101/2021.12.14.472632v1....

You should ask for your money back because you were miseducated.

Two questions: why so rude, and what do you base that on?

Given it's more than twice as contagious, this is not such great news - and this is ignoring the immune escape, which makes it much, much worse than Delta!

This is what I'm amazed that nobody seems to get - if the population of people infected by Delta is 80% unvaccinated (I'm making that up for the sake of an example), and the population of people infected by Omicron is only 20% unvaccinated because it's so much better at breaking through vaccination, then of course Omicron's going to have a lower hospitalization rate (unless it's vastly more severe, which it clearly is not). That's offset by the fact that it's infecting a much larger number of people because it can infect vaccinated folks much more effectively.

The much shorter doubling time makes sure it will overwhelm the much taxed intensivists. Death rates are much higher in triage.

Imperial college finally getting their contagion predictions lol

Whatever it takes to reduce the population, as the crash needs to happen to save the planet

I'm still happy we are seeing a overreaction rather than a underreaction with new variants.

Imagine a variant that spreads as fast as Omicron, but is more lethal than Delta.

With a underreaction we'd be back to January 2020. We clearly aren't out of the danger zone yet.

"The woods." We must continue using the phrase "out of the woods."

Reply on news.ycombinator.com