Hacker News Reader

Comment by gfodor | Hacker News Reader

There is presently to my understanding a lot of claims of severity reduction, but I would like to see evidence and an explanation for it. My working model is that a consistent alternative explanation is the vaccine reduces infection risk dramatically but once infected severity can be the same.

View on HN

That isn't how the immune system works. It's not a binary gate that's either open or closed. It's more like a war between two armies, and it depends on which side musters the larger army and how creatively the armies fight. If the vaccine reduces chance of infection, it will also reduce severity, because both operate via the same mechanism: antibodies killing the virus.

I've been trying to find accurate signals of this and haven't found much. I've only really spent time looking into the <50 yo cohort. I think the confounders for older folks are too significant to glean much truth from the type of data we have.

Data:

Graph from the CDC: https://i.imgur.com/Ql25IjB.png Source: https://www.cdc.gov/mmwr/volumes/70/wr/mm7034e1.htm

Spreadsheet based on weekly UK data: https://i.imgur.com/yySYFyd.png Source (PDF warning): https://assets.publishing.service.gov.uk/government/uploads/...

And a Twitter thread on Israel for good measure: https://twitter.com/gummibear737/status/1434237628258344961

All of this data says the same thing to me. Basically if you're younger there is no significant evidence these vaccines reduce your risk.

In the CDC case: They show 1-2 unvaccinated hospitalized cases out of 100k and <1 for vaccinated (<50 cohort). It's important to remember in all of this data the unvaccinated do not represent a strict control group which is analogous to one found in an RCT.

In the UK case: ~0.5% less incidence of hospitalization for <50 cohort and effectively none wrt mortality.

In the Israeli thread: Author comes to the opposite conclusion and decides to call risk reduction 90% even though they're comparing 2 extremely tiny numbers. Both less than 1 severe case out of 100,000.

If the improvement over severe outcomes is so stark, as many claim, then where should I look to see this data such that I can come to agreement?

I have seen this "alternative explanation" or variation on it used as an equivocation to justify the choice to go without being vaccinated.

I recently saw a social media post of a person in an ICU with Covid expressing regret about not getting vaccinated, and urging others to get vaccinated. One of their "friends" commented in response to the post with parent's "alternative explanation."

Would someone please share a link to source data showing that infection severity between vaccinated and unvaccinated is not similar?

The alternative explanation does not change P(severe infection|vaccinated), so doesn't change the nature of the choice to get vaccinated. But it would imply, if proven, that the claim that vaccines reduce severity is false.

If the alternative explanation is proven, it implies that people who are vaccinated and develop an infection should react to that news the same as an unvaccinated person. In other words, even if the alternative explanation doesn't influence the choice to get vaccinated, if it is true and we are saying it's false it would mean vaccinated people are more likely to die once infected because they've been led to believe it is not as serious a concern to be infected and will be less likely to, for example, notify their doctor as promptly or go to the ER given they have been informed their risk of severe infection is low.

Lower hospitalization rates amongst the vaccinated seem to suggest less people are getting seriously ill.

No it suggests P(severe|vaccinated) is much lower than P(severe|unvaccinated) which isn't the same thing as saying the vaccines reduce P(severe|infected). It's funny I've been downvoted tho because it shows how tribalist people are about vaccines that they can't even understand an argument that has no influence on ex-ante vaccine risk assessment.

Which is an important point. While the vaccinated are still getting COVID, and even sometimes ending up in the ICU, they very often aren't ending up hospitalized or in the ICU. The vast majority of ICU patients are unvaccinated [1].

[1] https://www.ncdhhs.gov/news/press-releases/2021/08/27/adult-...

> https://www.ncdhhs.gov/news/press-releases/2021/08/27/adult-...

It's frankly irresponsible to say this. We are talking about what's essentially an exploratory study that needs a LOT of followup before taking action and pushing it in a way that's going to cause people to run scared.

It's tempting to look at vaccinated/unvaccinated as a perfect RCT and the unvaccinated are a control group. This clearly isn't the case. There are surely many confounders at play.

Looking at this release's underlying report they mention:

"The age-adjusted mortality rate among unvaccinated individuals was 5.00 cases per 100,000 unvaccinated population and the age-adjusted mortality rate among vaccinated individuals was 0.33 per 100,000 vaccinated individuals for the four-week period ending August 28, 2021."

So the difference in mortality rate is roughly ~5 in 100,000 of the population. This is a very low number. Does it obviously override any possible confounders to indicate a strong signal? Also remember this is age-adjusted. I would imagine this difference is substantially smaller for cohorts omitting the elderly. Is there a material benefit to younger reasonably healthy people? It seems the question is muddier than is typically made out.

Generally when judging how some intervention does compared to not intervening, I believe you would divide the rates not subtract the rates.

I guess you're unable to respond to the substance of what I'm saying. Nominal risk reduction of already insanely low risk is not compelling especially when confounders are present and cannot be accounted for.

The evidence definitely points the other way, see [0]. The vaccines dramatically reduce the rate of severe infection and hospitalization. The vaccines do not prevent infection, they prime your immune system to fight the infection. That is why the endpoint for the clinical trials is not infection, which would have to be determined by testing, but rather symptomatic infection.

[0] https://www.covid-datascience.com/post/israeli-data-how-can-...

There’s also population skew since the vaccination rate is much higher among older Americans, and older people are more at risk of severe illness in general.

Reply on news.ycombinator.com