A new study comparing natural immunity, vaccines, boosters, and hybrid immunity for Covid dropped in NEJM yesterday.
If you haven't been paying close attention, the results might surprise you.
Details below...
Here's the key result.
In order, they found the combination of previous infection (🦠) and # of jabs (💉) had the following protection against symptomatic infection:
💉💉: -1.1%
🦠: 46%
💉💉💉: 52%
🦠💉💉: 55%
🦠💉💉💉: 77%
This is for BA2, though BA1 was similar.
Notably, 2-doses does exactly zero to stop transmission. 👀
Everything else confers some degree of protection between ~50% (making you 2x less likely to get Covid) and 75% (4x less likely).
Though durability is also important to look at.
Comparing 🦠 (46%) and 💉💉💉 (52%) - one notable is that the 💉💉💉 group had a median time between the last 💉 and infection of 43 days. Whereas for the 🦠 group it was almost a year. So (as we've already known), protection from infection appears far more durable than vax.
The authors hammer this home: "The analysis of the effectiveness of previous infection, 2-dose vaccination, and 3-dose vaccination as a function of time showed rapidly waning vaccine protection after the second and third doses but slowly waning protection from previous infection"
Take a closer look at that chart - and the x-axis.
Natural immunity appears to flatten out after 9 months. Meanwhile vaccine immunity looks to plummet within 6 months. The booster line ends at ">1 month". What do we think happens in 5 more?
Looking at hybrid immunity (that is 🦠+💉):
🦠 of 46% vs 🦠💉💉of 55% is unremarkable.
🦠 of 46% vs 🦠💉💉💉of 75% is better, but remember that's only 45 days of data.
The shots help a bit in the short term, not much in the long term.
The authors agree with this: "Protection conferred by hybrid immunity of previous infection and 2-dose vaccination was similar to that of previous infection alone, which suggests that this protection originated from the previous infection and not from vaccination."
In terms of severe outcomes they found that basically everything is still holding up well.
Every group is basically in the same boat (and above 70%) once you factor in error bars.
(Editorializing now...)
First, can anyone look at this data and claim that natural immunity shouldn't "count"?
Why are the majority of policies continuing to ignore previous infections? It's nonsensical.
Second, vaccine protection against infection wanes fast. If we want to keep using them for this purpose, we should expect people to need to be boosted every 6 months, possibly more frequently.
This doesn't feel sustainable.
Getting infected appears to be the only durable protection from future infections, and it isn't all that good.
We're all gonna get Covid many times in our lives. This is just unavoidable and we need to come to terms with it.
Finally, we need to take a closer look at our mandate policies.
Two doses does literally nothing to stop infection anymore. And yet it's exactly what is mandated by the majority of policies still in place.
Can anyone explain with a straight face what purpose that serves?
In my country (South Africa) it's estimated that over 90% of us have had Covid. But only the 30% who have 2-doses are allowed to enter many countries (including the US and Canada).
Why??
Literally, what is the purpose of that policy?
It's not stopping Covid, that's for sure.
I'm sorry for ranting. And for Tweeting about this at all. I'd like to move on from this but the world still isn't fully cooperating.
For the past several weeks I've been resurrecting a big, legacy Django project that was last meaningfully touched in 2013.
Here are a few observations from the effort.
1. Safety nets are absolutely critical.
What's a safety net? Something that will tell you if you borked something.
In this case, the project has a comprehensive test suite which is a fantastic starting position.
Without the tests this process would be ~impossible.
The second safety net I use is error monitoring (@getsentry). You can only really use this one in a low traffic/stakes situation, which thankfully, this is.
I wasn't able to set up sentry until upgrading from Django 1.3(!!) to 1.6, and felt so much better after that was done.