Share this page!

Farzad Mostashari Profile picture
Twitter logo
26 Sep, 4 tweets, 1 min read
This is actually an answerable question.

Does case severity drop during times when there are lots of open hospital beds?

Seems like a good episode of @DrBapuPod @AnupamBJena
Can we measure "likely unnecessary admissions" somehow?
...and to be clear, I am not sympathetic to the idea that we should somehow minimize the terrible negative impact of covid on hospitals and ICUs being full
(from a former policymaker/ hospital exec) "when IP census lags, do we see a rise in the % of ED patients admitted" might be the question.

• • •

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

Keep Current with Farzad Mostashari

Farzad Mostashari 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 @Farzad_MD

Farzad Mostashari Profile picture
15 Aug
1) Private Equity and Health Care Delivery by @brianwpowers @WillShrank @AmolNavathe

💯 Agree private equity isn't inherently good or bad "acts to amplify the response to extant financial incentives"

A couple of further points though....

ja.ma/3m6K6FU @JAMA_current
2) it's true that good policy can align private profit with public good, but if we are going to rely on that, need tight surveillance and fast response from regulators to close arbitrage opportunities where short-term profit maximizers will gather. c/f surprise billing
3/ if you wait too long, then entrenched profits become normalized, powerful incumbents are formed, and they can, and will, exert political influence to keep the "status quo" in place.

Many health policy examples (facility fees, drug pricing). But also...Medicare Advantage 👀
Read 12 tweets
Farzad Mostashari Profile picture
31 Jul
1/ Here's another quiz for budding epidemiologists

The Guandong Delta outbreak was meticulously investigated. Initial report of 1000x viral load came from there

A finding that hasn't been discussed much is the seeming shorter incubation period (panel b)

virological.org/t/viral-infect…
2/ In 2021 delta it was only 3.7 days (vs 5.6 days for 2020 outbreak).

This would have an impact on a key transmission dynamic factor we often look for: "serial interval periods" (time between symptom onset for index case vs subsequent case in a contact tracing investigation)
3/ What you are trying to estimate from observable symptom intervals is underlying mean generation time.

tangent: If you find negative serial intervals as in COVID, it's a sign of asymptomatic and presymptomatic spread.
Read 6 tweets
Farzad Mostashari Profile picture
31 Jul
Without looking it up,

What is Rt in Arkansas right now?
Rt is actually 1.2 (0.98-1.5). It is taking 12 days (4 generations) to double

here's some math

If Rt (aka Reff) was 5, it would mean that it would go up 5*5*5*5 = 625 TIMES in 12 days, not 2x

epiforecasts.io/covid/posts/su…
3/ Now, we aren't living in an R0 world. Even though it may feel like it sometimes

People are vaccinated (despite breakthrough infections)
People are infected (despite reinfection risk)
People have changed behaviors

But "Delta variant is as transmissible as Chicken Pox"(R0~9)??
Read 8 tweets
Farzad Mostashari Profile picture
30 Jul
1/ Like many others, I've been frustrated that we haven't seen the data behind the CDC's new recommendations. Parsing exact wording in transcripts is🤦

We're told there will be a release tomorrow but thanks to the @washingtonpost there's an internal CDC document to parse tonight
2/ The article by @yabutaleb7 @carolynyjohnson @JoelAchenbach is here:
washingtonpost.com/health/2021/07…

TY @bijans for spotting the "full pdf" download button.
3/ what do we learn?

The mysterious "other data" for high viral load in breakthrough cases came from a 4th of July outbreak in Provincetown (Barnstable, Mass) where the “vast majority” of the new cases were among fully vaccinated individuals

cc @zeynep
thehill.com/homenews/state…
Read 16 tweets
Farzad Mostashari Profile picture
28 Jul
1/ What percent of vaccinated people who get infected with covid will have long-term symptoms?

I don't know, but I highly doubt that it's 19%.

Here's why I say that, and what the *right* study design would be for answering that question.
2/ when I was an Epidemic Intelligence Service Officer at the CDC, I investigated lots of outbreaks, and every one would have a symptom checklist

A surprisingly high number of people with a diagnosed infection will say YES to a variety of symptoms

Fatigue?
Headache?
Fogginess?
3/ many of these reported complaints would have absolutely nothing to do with their infection.

But recall bias is super powerful, and respondents often feel like they're supposed to say "yes". So they do.
Read 7 tweets
Farzad Mostashari Profile picture
7 Jul
1/ Our national quality/value program (MACRA) is broken

Most people just complain about its shortcomings, but @Travis_Broome comes up with an elegant, grounded way to fix its biggest flaws

Apply behavioral economics, use virtual groups, lay path to APMs ajmc.com/view/macra-has…
2/ MACRA was a true milestone, and a concept that I still support- instead of artificially capping medical inflation (and then not having the guts to actually see doc pay cuts) lets create 2 paths- a "pay for performance" base and an incentivized alternative payment model track.
3/ But 3 seemingly technical details fundamentally sapped the potential impact of this huge bill.

classic behavioral economics- the impact of an incentive is not just proportional to its size, but also its cost, uncertainty, and delay

MACRA stunk on all counts here.
Read 14 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

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

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!

Like this author's thread?

Get emails when @Farzad_MD has new unrolls!

Check out other threads from @Farzad_MD!

Read all threads by @Farzad_MD

:(