It’s always good to read the actual study instead of just the news article. Reporters aren’t always that good about accurately describing the results of a study.
They compared the outcomes of four separate groups:
unvaccinated with no prior infection
vaccinated with no prior infection
unvaccinated who recovered from a prior infection
vaccinated who recovered from a prior infection.
Per this main figure, people with no vaccine and no prior infection did far worse than any of the other groups. (Duh). But those two groups that had recovered from a prior infection both did better than those who had a vaccine alone.
So it’s not like this report suggests that natural immunity is worthless. If it’s accurate it means that natural immunity alone is better than vaccine immunity alone for preventing severe outcomes. Of course you need to survive covid first in order to get it, so there may be a fair amount of survivorship bias built into the results.
Four days before Christmas, my son and I drove to the DC Costco and parked, but instead of going shopping, we walked down the hill and went into a nondescript DC government warehouse. There, we walked around thousands of boxes of Covid testing supplies to join a few dozen other vaccinated and masked DC residents who’d responded to the city’s call for volunteers to assemble Covid PCR test kits that people could use at home. We spent a few hours on the assembly line, folding up Labcorp baggies and instruction leaflets for the volunteers at the next table to stuff into packages and vacuum seal.
Once they were quality checked, the kits went right out the door to city libraries, where residents could pick up two a day for free. After they swabbed their noses, residents could drop the kit into a library pickup box outside. After 8 pm, the city transported the tests to a lab for processing and, in a few days, results would be emailed. Over the course of the next two weeks or so, volunteers from all over the city assembled tens of thousands of test kits for distribution.
the text on that graph describing events is a bit odd. maybe something in the paper explains? it doesn’t seem to actually sort out first, second, and booster doses even though that text implies it.
definitely everything else ive heard has been inline with: vaccine acquired natural immunity is more protective than exposure based natural immunity. it’s thought to be more broad protection against variants as well.
i guess it will be interesting to see which holds up as more analysis rolls in…
(eta) the paper’s abstract is interesting:
Before Delta became the predominant variant in June, case rates were higher among persons who survived a previous infection than persons who were vaccinated alone. By early October, persons who survived a previous infection had lower case rates than persons who were vaccinated alone.
maybe seeing those preliminary numbers were why they really started pushing on boosting
Do you have any references to those studies? There’s one earlier one I saw from the CDC that compared prior infection with no vaccination against those with prior infection AND vaccination, with the latter group doing better. But I haven’t seen a study showing that prior infection alone does worse than vaccination alone. (Not that such a study doesn’t exist. I just haven’t seen it yet)
Here’s the study that a lot of people were citing several months back, but if you read it it doesn’t say what a lot of people assumed it said:
Whoa, that’s a leap given the time-base of that data set. Without factoring booster in at all, that’s accurate, but what this data set seems to really be showing is that both vaccine and prior infection immunity depletes over time. In order to be able to draw your conclusion, patients in each cohort would have to be roughly matched in time since immunization/infection. The analyzed data set cuts off just before the big booster push last fall/winter, and that would be when I would expect that comparison to hit an inflection point where the vaccine became more protective than prior infection again.
But, again, the time scale is different for each cohort. The main vaccination push for the two vaccination cohorts was 6 months prior to the largest peak in infection. So the functional comparison between cohorts is vaccination plus 6 months vs recent infection, not vaccination vs infection. There is no way to use that data set to come to the conclusion that any infection cohort has better protection than any vaccination cohort unless they can time-match the patients by vaccination date vs infection date.
A characteristic of coronavirus infections as a whole is the fairly rapid decline of immune protection. We never said “I had a cold already, so I won’t get another one.” It’s just that the previous coronavirus infections did not fill up the ICU’s!
(Trying to find a study on non-covid coronavirus immunity historically is remarkably difficult!)
What do you mean by “recent” infection? Maybe I misread it but my understanding is for the “prior infection” group that they were only including people who had been diagnosed prior to March 1. Are you reading it differently?
I missed that part. Good spot. So I still think to draw meaningful conclusions, I would pair patients by vaccination/infection date, but it looks like they did factor that in, at least by elimination(which has its own statistical problems but it’s usually better than leaving in directly confounding data).
Looking in that section, one thing that jumps out is the disparity in the size of the cohorts. The two big cohorts are vaccinated alone and unvaccinated w/o prior infection. The small cohorts are the vaccinated w prior infection and unvaccinated w prior infection. Comparison between either of the large groups and either of the small groups becomes problematic due to that disparity. Typically in medical research, when we’re looking at that kind of disparity in cohort size, I would rather see a Kaplan-Meier survival curve rather than full population hazard ratio curves. That way, we can see event-free survival over time and not just overall case/hospitalization rates.
where the paper’s abstract said the same… until october when they started to notice a difference )