I disagree that all sectors of our world must be demographically proportional to the population in general. Further, using the argument that any racial group is best understood by someone of the same group is also a false conclusion.
Regardless, for doctors, the future is AI, which in 20 years will likely replace most, if not all MD's and AI's don't have any race or color.
Hi Jojo, thanks for the comments. A few thoughts. The first point is a matter of preference. Some will have that preference, others not, and it's difficult to argue over preferences or values. The second point you raise is actually an empirical one, meaning it can be evaluated with actual data and can be debated on the merits. The study we refer to in the post was a randomized trial in Oakland, CA to assess this exact issue. It suggests that racial concordance between provider and patient may be beneficial in the context studied. It may not always be but other empirical work has reached similar conclusions. Third, AI may change medicine but lots of people have thought deeply about how biases (racial, cognitive, or otherwise) can creep into AI. It's important to be aware of those possibilities when making clinical decisions based on AI.
1. My first point was not meant to be considered a point of preference. It was an absolute. "Preference" was an argument used to support and advance segregation decades ago and is no more valid today than it was then.
2. I wouldn't quote a study out of Oakland/Berserkley, CA as being anything to take seriously. I personally have no preferences as to the ethnicity of medical people I interact with. I've encountered competent and incompetent MD's of all races.
Of course, I have heard this "someone like me" meme in the media as it is often used in support of elevating people into roles and places that they perhaps should not be solely because it makes some people more comfortable.
3. AI, hopefully, will not have any biases that aren't learned from actual experience although again, as always, there are a subset of people who would love to imbue it with limits that make sense to themselves. Similar to how the public health agencies did not want anyone to read anything related to Covid that was not THEIR truth.
Speaking of AI in medicine, here is a new article that you may find interesting:
------
Dr. ChatGPT Will Interface With You Now
Questioning the answers at the intersection of Big Data and Big Doctor
It would be interesting to see if there is any relationship that could be established (rather than just inferred) between the results you report and the existing studies that show that patients of minority status, particularly black men and even moreso women, are statistically more likely to have their concerns dismissed/not taken seriously than white patients.
I like the idea! It would be an undertaking to get all the data, particularly tracking down the patient experience data for doctors from these schools who spread across the country to practice. But you could theoretically follow the doctors included in this study out for many years, through residency and any subsequent training, and look for differences in patient experience/satisfaction between doctors who were part of a more diverse med school class just before the ban compared to those who studied just after the bans. I would also be interested to see, in addition to the potential impact on patients, the potential impact on their careers of studying alongside people from underrepresented backgrounds. Could it impact their choice of specialty? Where they practice? Where they live? Etc
Hi Andrew, great question. It could act in that direction but I'm not sure it will for at least two reasons. First, it depends on who the applicants are that are on "the margin" under a new evaluation system that does not explicitly prioritize race or legacy. If those students don't fall into under-represented groups, then there would be no offsetting effect. That's a theoretical point. Empirically, the study of ours that we referred to above focused on medical school admissions. I don't think (or least know of firm evidence) that legacy admissions play a large role there, so the reductions we observed in that setting might still hold. There is a caveat that medical school admissions are downstream of college admissions so to the extent that a small group of colleges have legacy admissions, part two of my answer might not hold (though in our study, as you can see from the plot above, we saw near immediate reductions in under-represented groups, suggesting that this was not simply pipeline driven). - Bapu
Thanks, Bapu. That is helpful! And, I'm really just trying to wrap my head around this complicated mess we're in. Really appreciate the medical example!
In addition to what Bapu said, in my opinion the idea of "legacy admissions" in medical school is not completely analogous to the undergraduate world. For example, I don't think anyone in my medical school class had parents who went to that same school, but it was common for them to have parents who were physicians. Given all of the hoops one has to jump through to be able to even apply to medical school, having parents who are physicians gives a legacy-like advantage to getting into medical school at all, even if it's not the same medical school. They can help a pre-med access many of the opportunities necessary to get in to medical school, such as volunteer, shadowing, or research opportunities that many applicants find hard to come by. They'll have a better sense of what admissions committees are looking for on an application. Since getting into medical school at all is such a competitive process, which specific medical school they go to may be a secondary concern for a parent with a child interested in medicine.
I disagree that all sectors of our world must be demographically proportional to the population in general. Further, using the argument that any racial group is best understood by someone of the same group is also a false conclusion.
Regardless, for doctors, the future is AI, which in 20 years will likely replace most, if not all MD's and AI's don't have any race or color.
Hi Jojo, thanks for the comments. A few thoughts. The first point is a matter of preference. Some will have that preference, others not, and it's difficult to argue over preferences or values. The second point you raise is actually an empirical one, meaning it can be evaluated with actual data and can be debated on the merits. The study we refer to in the post was a randomized trial in Oakland, CA to assess this exact issue. It suggests that racial concordance between provider and patient may be beneficial in the context studied. It may not always be but other empirical work has reached similar conclusions. Third, AI may change medicine but lots of people have thought deeply about how biases (racial, cognitive, or otherwise) can creep into AI. It's important to be aware of those possibilities when making clinical decisions based on AI.
1. My first point was not meant to be considered a point of preference. It was an absolute. "Preference" was an argument used to support and advance segregation decades ago and is no more valid today than it was then.
2. I wouldn't quote a study out of Oakland/Berserkley, CA as being anything to take seriously. I personally have no preferences as to the ethnicity of medical people I interact with. I've encountered competent and incompetent MD's of all races.
Of course, I have heard this "someone like me" meme in the media as it is often used in support of elevating people into roles and places that they perhaps should not be solely because it makes some people more comfortable.
3. AI, hopefully, will not have any biases that aren't learned from actual experience although again, as always, there are a subset of people who would love to imbue it with limits that make sense to themselves. Similar to how the public health agencies did not want anyone to read anything related to Covid that was not THEIR truth.
Speaking of AI in medicine, here is a new article that you may find interesting:
------
Dr. ChatGPT Will Interface With You Now
Questioning the answers at the intersection of Big Data and Big Doctor
Eliza Strickland
7 Jul 2023
https://spectrum.ieee.org/chatgpt-medical-exam
It would be interesting to see if there is any relationship that could be established (rather than just inferred) between the results you report and the existing studies that show that patients of minority status, particularly black men and even moreso women, are statistically more likely to have their concerns dismissed/not taken seriously than white patients.
I like the idea! It would be an undertaking to get all the data, particularly tracking down the patient experience data for doctors from these schools who spread across the country to practice. But you could theoretically follow the doctors included in this study out for many years, through residency and any subsequent training, and look for differences in patient experience/satisfaction between doctors who were part of a more diverse med school class just before the ban compared to those who studied just after the bans. I would also be interested to see, in addition to the potential impact on patients, the potential impact on their careers of studying alongside people from underrepresented backgrounds. Could it impact their choice of specialty? Where they practice? Where they live? Etc
Sidebar question for you: do you think the push to remove legacy admissions offsets this loss somewhat?
Hi Andrew, great question. It could act in that direction but I'm not sure it will for at least two reasons. First, it depends on who the applicants are that are on "the margin" under a new evaluation system that does not explicitly prioritize race or legacy. If those students don't fall into under-represented groups, then there would be no offsetting effect. That's a theoretical point. Empirically, the study of ours that we referred to above focused on medical school admissions. I don't think (or least know of firm evidence) that legacy admissions play a large role there, so the reductions we observed in that setting might still hold. There is a caveat that medical school admissions are downstream of college admissions so to the extent that a small group of colleges have legacy admissions, part two of my answer might not hold (though in our study, as you can see from the plot above, we saw near immediate reductions in under-represented groups, suggesting that this was not simply pipeline driven). - Bapu
Thanks, Bapu. That is helpful! And, I'm really just trying to wrap my head around this complicated mess we're in. Really appreciate the medical example!
In addition to what Bapu said, in my opinion the idea of "legacy admissions" in medical school is not completely analogous to the undergraduate world. For example, I don't think anyone in my medical school class had parents who went to that same school, but it was common for them to have parents who were physicians. Given all of the hoops one has to jump through to be able to even apply to medical school, having parents who are physicians gives a legacy-like advantage to getting into medical school at all, even if it's not the same medical school. They can help a pre-med access many of the opportunities necessary to get in to medical school, such as volunteer, shadowing, or research opportunities that many applicants find hard to come by. They'll have a better sense of what admissions committees are looking for on an application. Since getting into medical school at all is such a competitive process, which specific medical school they go to may be a secondary concern for a parent with a child interested in medicine.
Thanks, Chris! I love hearing the nuance.