Buzz Hollander MD Profile picture
Oct 31, 2021 9 tweets 4 min read Read on X
The reaction to the recent CDC MMWR on immunity after covid infection being 5X less protective than vaccination shows off everything that has been wrong about science communication through this pandemic.
cdc.gov/mmwr/volumes/7…
“Told ya so” -
Seriously?
Public Health voices got their Gotchas in for anyone in anti-vax, anti-mandate, or just anti-establishment camps who ever dared to argue that we should consider immunity from prior infection on par with vaccination in terms of protection from disease.
But what DID the study show?
A: Not much.
Anyone claiming this study shows that unvaccinated people with prior infection are 5X more likely to be hospitalized than fully vaccinated people, probably either did not read the study, or did not understand it.
And should not be tweeting about it.
Put briefly, to do a study like this, you’d want a lg watershed of people for whom you have solid data on infection/vaccination history w/in a time frame (our denominators), and then complete data on that same watershed on who got sick later w/ covid (our numerators).
Like this:
That was our earlier Israeli pre-print:
medrxiv.org/content/10.110…
It checked all the boxes: large HMO, reliable records, sound study design.
It found prior infection far more protective vs symptomatic infection (7X) and also more protective vs hospitalization.
So what gives?
What did this CDC MMWR measure?
Numerator was covid hospitalizations;
Denominator was… people w/ vaccination w/in 6 mo vs people w/ prior infection who were hospitalized for respiratory infections. Not the whole watershed. Just those inpatients. 90+% of whom DID NOT HAVE COVID.
Hmmm. If about 50-60% of ppl are fully vaccinated in those areas; and maybe 15-20% of ppl are unvaxed but had prior infection; you might conclude: ppl w/ prior infection got really sick less than vaccinated ppl. And had less severe covid, too.
But you would just be guessing.
One thing that is conclusive, though, is that the study’s headline is not true.
You cannot possibly infer that vaccinations out-protected prior infection from this study.
Amazingly these limitations did not make it into the authors’ *discussion* of study limitations. 🤔
And apparently didn’t make into the heads of those who proclaimed these results as if they were decisive.
We wonder why the vax-hesitant don’t trust public health voices?
We’ve got to start by being honest.
I’m still quite happy my first covid exposure will come after a vaccine.

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Buzz Hollander MD

Buzz Hollander MD Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @buzzhollandermd

Feb 28, 2022
Crawling out from a week in which the War in Ukraine has felt a lot more compelling than Covid...
but I thought I'd review the big stories from the past week.
There actually was some interesting news. 1/7
#1: CDC managed to release mask recommendations based on case & hospitalization rates that pissed EVERYONE off. Maskers distressed that US overnight became low risk; non-maskers vs implication we will keep doing this; everyone noted that when hospitals fill, it's too late. 2/7 Image
#2: a favorite follow, @jsm2334, had study published showing mask mandates dropped case rates, comparing similar masked counties to unmasked. Study/discussion worth a read.
covid-datascience.com/post/masks-pap…
I suspect confounding responsible, but best study w/ this conclusion I've seen. 3/7
Read 7 tweets
Nov 26, 2021
Since most virologists I follow are being pretty technical about the B.1.1.529/“Omicron” variant that is feeding a MSM headline frenzy, I will add my non-technical thoughts.
I don’t usually get too fussed over new variants, as they mostly flame out.
Am I concerned?
Yes. 1/9
Omicron is loaded with mutations that *might* make it more resistant to immunity of all types (bearing in mind that S AFrica where it is spreading prob has a little more immunity from infections than vaccinations). 2/9
So - we’d probably rather it not take over.
The big question is: can it outcompete delta and do so?
We simply lack the answer to that question. The numbers in SA are so small still because their pandemic was in a deep lull that it’s too hard to say. 3/9
Read 9 tweets
Oct 30, 2021
This Lancet study from the UK has fueled a lot of “gotchas,” only some of which are appropriate. It’s interesting, though:
thelancet.com/journals/lanin…
The take-away: vaccinated ppl with “breakthrough” infections were as likely to spread to household members as unvaccinated ppl.
Take w/ a grain of salt, as even a smart, prospective study like this cannot avoid major confounders, esp when this small (163 cases, 232 household contacts). The index cases captured in this study might not represent average folks.
Still - this is as good as we’ve got.
I feel like some public health voices looked at the data from Singapore and N Illinois with rose-colored glasses and decided infected vaccinated ppl have lesser viral loads or less viable virus to spread; that’s not how I saw the same data.
ITRW, this study suggests the contrary.
Read 7 tweets
Oct 29, 2021
I appreciated the kind words of Prof Morris:
And, more reluctantly, the observation of one of his numerically-literate followers that I’d mis-characterized a CDC study on odds of getting myocarditis from Covid-19.
The “you’re 16X more likely to get myocarditis from covid” study.
I have seen that figure referenced twice in the last 24 hr, despite the study being 2 mo old.
It’s relevant now, in that lots of people are trying to assess the risk of myocarditis in kids if they choose/are mandated to have the vaccine vs if they do not.
My mistake? Image
I missed that this “hospital-based” study included hospital-based OUTPATIENT clinics where people were diagnosed with covid:
cdc.gov/mmwr/volumes/7…
I had assumed that the denominator was all hospitalized patients, when really it was an unstated mix of inpatients and outpatients.
Read 8 tweets
Jul 10, 2021
#ivermectin for covid has been in heavy-play mode in the media again.
Why?
A new trial report, and 3 meta-analyses.
As usual, half the world thinks these are proof of the #ivermectincure and half thinks it's bogus.
Why can't we all get along?

foxnews.com/media/biologis…
@DrAndrewHill released his team's meta-analysis right on the heels of 2 other meta-analyses, one positive & one negative. Dr Hill, who has been a bit of an ivermectin cheerleader, had a positive conclusion:
academic.oup.com/ofid/advance-a…
A huge concern w/ any meta-analysis is "Garbage In, Garbage Out" - and we've plenty of garbage in the IM world.
Whether a study rates inclusion and how it gets weighted in a meta-analysis can swing the conclusion.
These meta-analyses don't move the needle.
Read 13 tweets
Mar 14, 2021
Another look at long covid out in a pre-print, from @mlipsitch&co:
medrxiv.org/content/10.110…
I'm not surprised, but at last someone had the notion to compare folks post-covid to folks post-viral-infection, non-covid.
The headline is "14% had sequelae" but they buried the lede! 1/6
First, the limitations:
1) totally based, not on survey/exam, but on MD-reported ICD-10 codes; so these are "long," possibly random, symptoms bad enough to see an MD and have them code it.
2) retrospective cohort, so all the usual bias & matching concerns.
3) only age 18-65. 2/6
But it was HUGE (>200K matched cases), which helps;
and fairly well-matched;
and, at last, well-chosen cohorts (2020 no covid, 2019, pre-2020 viral).
WHAT DID THEY FIND???
Post-covid, you had a 14.6% chance of seeing an MD for post-viral problems.
Pre-2020 virus? 13%. 3/6
Read 8 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(