>There was a study earlier this year which had the opposite findings [1]. This study addressed the discrepancy

They did address the one study from Israel. They did not address the nearly 20 other studies also done which share its conclusion that natural immunity is at least as powerful as a jab.

So this study only looks at covid-19 patients that were hospitalized and comparing that to the vaccinated population? How is it an improvement to introduce a selection bias that wasn’t there in a previous study? I guess more data points are always good but then the conclusion they reach here about natural immunity vs vaccinations needs to be qualified.

Well to be fair the other bias is against the prior infection group: those who didn’t survive the infection weren’t included in the study and were sure to drag down the average immune response for that group.

The study confirmed that covid vaccines are marginally better than natural immunity against severe covid within 0-6 months since the immunization event. "Laboratory-confirmed SARS-CoV-2 infection was identified among 324 (5.1%) of 6,328 fully vaccinated [hospitalized] persons and among 89 of 1,020 (8.7%) unvaccinated, previously infected [hospitalized] persons."

Independent question: are covid vaccines effective farther away from the immunization event? Could we perhaps plot a chart showing VE performance over time? Time will tell.

So best case scenario, vaccines only win short term?

No, that's an unsound conclusion. That assumes that all the unobserved kinds of cases are like the those in the Israeli study and not like those in this.

The sound conclusion is: "This seems to be complicated; further study needed."

Yes, but in a bayesian sense, what are we updating toward? So far everything indicates that disease has better imunity long term.

I think you're confusing apples with trees here...

Your choice now is whether to vaccinate or not. It's not whether to vaccinate or be infected. Being infected or not is another choice that luck chooses for you at some point in the future. Being infected again or not is another choice that luck chooses for you at some point still further in the future, assuming that you luck decides to infect you in the previous step.

So: Your resistance against the first infection is near zero without the vaccine, much better with. Your resistance against the second infection is what's being discussed in this study.

I'm not sure how to compare vaccines with natural resistance fairly without that off-by-one infection number. People say it's better than the vaccine, but that disregards the nonzero chance of dying while acquiring the immunity. Some five million people have died in the process of acquiring that immunity — it seems difficult to call that "better than" anything.

No, I'm not :)

I'm comparing immunity acquired by infection with immunity acquired by vaccine, post-factum. I'm definitely not saying that one should use infection as a substitute for vaccine, and honestly I'm a bit miffed that you're suggesting I am.

Also this is far from being a useless fact to know. For example in my country prior infection doesn't count for getting your green certificate. If infection is indeed better than vaccine, than you have a very strong argument that the government is making a pretty big mistake. If it offers the same order of magnitude of protection, that you have a weak argument. If it's significantly worse than vaccine, than you have no argument against this policy, and in fact it's a pretty good policy.

OK, here's a comparison.

For the first infection, immunity acquired by infection doesn't exist yet, so for the first infection, that strategy offers an increased risk of death of 20× or more compared to vaccination.

For the second infection, the two strategies are roughly equal. Not exact, but the details don't matter, because...

For these two on average, immunity acquired by infection offers about 10.5× the chance of death. If the effectiveness during the second infection varies very much, then that number will be 9 or 13 or something, so who cares about that detail. The number 20 dwarfs everything else. The only way to make the two strategies comparable is to ignore the outcome of the first infection.

Note that 20 isn't meant as a serious estimate, it's a number I picked out of a hat. MUCH higher estimates are available. You may have read that 99.5% of deaths in the US are unvaccinated. If you start with that number and that two-thirds of the USians are vaccinated, then the number 20 becomes 400, and immunity acquired by infection is then roughly 200× as dangerous as the vaccines they use in the US.

I. AM. NOT. SUGGESTING. TO. USE. INFECTIONS. TO. GET. IMMUNITY.

:)

I'm only considering already acquired immunity from past infections. And I've given an example on how that's relevant. It's probably another example of polarization that you're assuming almost by default that I'm an enemy anti-vaxxer (I'm not). Oh well, nothing else to do that to keep talking, and hoping civilized conversation helps.

I didn't assume that you did.

Rather, I thought you were interested comparing kinds of immunity. To to that fairly and well, you need to include the most common case: People's first infection.

Remember that most people's immune system is able to learn quickly enough that natural immunity takes care of the first infection without external help. The immune reaction doesn't start with the second infection, it starts as soon as the immune system starts learning. Excluding the first infection in an analysis seems very, very strange to me.