Read on Twitter
Had a chance to read the PNAS paper 
I believe we absolutely need more Black physicians, and more physicians from multiple under-represented groups!
Unfortunately, the paper is too weak to make the claim CNN is amplifying
Some challenges [thread] https://twitter.com/CNN/status/1295777514976153600
I believe we absolutely need more Black physicians, and more physicians from multiple under-represented groups!
Unfortunately, the paper is too weak to make the claim CNN is amplifying
Some challenges [thread] https://twitter.com/CNN/status/1295777514976153600
The first challenge is that the entire paper rests on the idea that physician - patient pairing is quasi-randomized.
This is a recent idea, but as far as I can tell there is no evidence that supports it
It is entirely possible that some people do choose their docs
This is a recent idea, but as far as I can tell there is no evidence that supports it
It is entirely possible that some people do choose their docs
Wealthier Black families may be more likely to seek out practices with Black physicians or specific Black docs, than poorer Black families.
The socioeconomics, and not the racial-pairing may then account for the differences seen
The socioeconomics, and not the racial-pairing may then account for the differences seen
The other core challenge is ascribing the outcome of the child to one of many doctors who plays a role in the care of that child.
Health care is a team sport. All of these papers that seek to link doctor characteristics to patient outcomes must make divisive pairings
Health care is a team sport. All of these papers that seek to link doctor characteristics to patient outcomes must make divisive pairings
Those divisive pairings are debatable.
Moreover, they are artificial.
In this paper, the pairings are unclear. I can't figure out how they make them despite reading the paper and supplement.
Moreover, they are artificial.
In this paper, the pairings are unclear. I can't figure out how they make them despite reading the paper and supplement.
The third challenge. I don't like what this paper and commenters insinuate.
Yes, we need more Black and URM physicians,
AND more physicians from lower SES,
AND marginalized groups (sexual orient. or gender identity)
AND more..
Our workforce must be as diverse as we r
But...
Yes, we need more Black and URM physicians,
AND more physicians from lower SES,
AND marginalized groups (sexual orient. or gender identity)
AND more..
Our workforce must be as diverse as we r
But...
But, we should not have prove that having this diverse workforce lowers mortality.
That is a bizzare standard.
We need a diverse workforce for reasons of justice and equality, and whether or not Black doctors have the same outcome is not germane to the discussion
That is a bizzare standard.
We need a diverse workforce for reasons of justice and equality, and whether or not Black doctors have the same outcome is not germane to the discussion
Trying to prove that point is a distraction from the moral need for diversity
It may or may not be true, but certainly is not relevant to the pressing need.
It may or may not be true, but certainly is not relevant to the pressing need.
The amplification of this paper, which has serious deficiencies, concerns me.
I am committed to two things.
Racial justice, and strong evidence in policy & science
The world would be better if we commit to both
I am committed to two things.
Racial justice, and strong evidence in policy & science
The world would be better if we commit to both
I am aware of dozens of people who agree with my assessment of this paper and are scared to comment.
I have tried my best in this thread to show you my strong moral commitment to the cause, but also to highlight concerns with the specific research study.
I have tried my best in this thread to show you my strong moral commitment to the cause, but also to highlight concerns with the specific research study.
PS I did not mention the way in which authors ascribe racial identity by having research associates look at doctor photos and judge surnames.
Not a big fan of this practice as a general rule.
Not a big fan of this practice as a general rule.
Indeed. This is a big problem https://twitter.com/mikejohansenmd/status/1296175856625356802?s=19
https://twitter.com/mikejohansenmd/status/1296175856625356802?s=19
Since my thread, the "Physician of Record" was clarified to me
It turns out this is a field in Florida for which there is absolutely no consistency in coding
It varies widely https://twitter.com/hhburris/status/1297006264891904008?s=20
It turns out this is a field in Florida for which there is absolutely no consistency in coding
It varies widely https://twitter.com/hhburris/status/1297006264891904008?s=20
There is potential that it is different in encounters where the baby unfortunately dies vs. ones where the baby lives
And the racial make up of doctors may vary by specialty https://twitter.com/BrianKingNeo/status/1297019274410840064?s=20
And the racial make up of doctors may vary by specialty https://twitter.com/BrianKingNeo/status/1297019274410840064?s=20
I will summarize by saying that a non-standard, and potentially systematically different (in cases of alive v. dead) attribution of the doctor of record is a problem that no analysis can overcome. It is a data limitation.
Interested folks can listen to this week's episode. Recorded before I knew this last pt https://twitter.com/Plenary_Session/status/1297006382105784320?s=19
