It looks like I'll have to be the contrarian here.
I think Dr. Cifu's understanding of race is out of step with the biological and genetic literature. It is simply false to say that it is a social construct with little basis in reality. He undermines his argument by pointing out the life expectancy in the US by race; though he implies that this has something to do with racism, these charts and effects are seen all over the world, even in majority black countries.
IQ has been identified as a strong indicator of life outcomes and IQ is strongly hereditary. No matter where it is tested, we see the same type of group differences by race: blacks tend to score towards the bottom, hispanics slightly higher, then whites, and then asians at the top of the IQ range. No matter how these tests are conducted, the results are almost always the same.
The same things can be seen across a host of other outcomes across different races. Life expectancy, criminality, rate of STIs, economic growth, etc. we see the same pattern; blacks toward the bottom, then hispanics, then whites, and lastly asians at the top.
Not to mention that stereotype threat is among the most reproduced sociological findings ever. People's stereotypes of other groups tend to be more accurate than they are inaccurate.
Saying that two groups of humans are more similar than they are different is just a red herring. We have 98% genetic similarity with chimpanzees, would we seriously say that the difference between humans and chimpanzees is solely a social construct?
The idea that every human is a tabula rasa has been roundly disproven. While our genetics are not 100% determinative of our future, it is simply impossible and wrong to ignore that they have an outsized effect on our future and in ways that are not immediately obvious to us.
I mostly agree. Mostly, i say, primarily because of my bias towards avoiding absolutes. There may be times when someone’s race (or other social constructs) may be appropriate identifiers in a chief complaint or the HPI. Certainly relevant in the social history.
Great read, better to think of individual risks than collective undertones. This is what patients want as well. To the extent that racial constructs help patients to be treated as unique individuals, I’m all for it. To the extent that it perpetuates harmful stigma, we should seek to avoid it. Seems to me the problem is in detecting the difference… who’s to say? This is what makes the topic so pernicious. It can be used to traffic really harmful ideas, so good to expand (rather than mute) conversation and be clearer about instances of systemic racism (eGFR adjustments in Neph Transplant), instances of true risk stratification (HLA B*5801 testing in southeast Asian populations before prescribing allopurinol), and instances when it’s aimless, lazy, and unfortunately normative to make race ‘the identifier’ of a social history. I’d rather know what patients do for a living and how they spend their free time. Maybe we just need to be clearer about our intentions (when race is invoked) and better at appraising true vs perceived risks related to this complex construct.
Good stuff, interesting points. On the one hand, I don't think I disagree with your actual argument. On the other hand, I wonder if including/not including race might depend more on the clinical circumstances than a rule. For example, I'm thinking of cases where the details of the HPI are heavily "psychosocial", in which omitting the race of the patient would actually leave out important context and would seem out of place relegated to the social history. Not that these are necessarily the majority of cases, but there are certainly some, in my opinion. Perhaps in psych we see relative more of this?
Also—and this isn't an argument against your stance but just a concern—I worry about people taking the wrong message from this and viewing it primarily as an issue of "not being judgmental/stigmatizing". eg. I remember as PGY-4 teaching interns about the psychiatric evaluation and stressing the importance of observing the physical characteristics of the patient. In one case, an intern omitted reference to a patient's numerous face tattoos which were quite obviously gang-related. She omitted this because she was worried about "stigma". Obviously face tattoos and race are not the same since we can choose to get tattoos but can't choose our genes, but the general habit of "I'm not going to mention what I'm seeing in front of me due to fear" is not a good one to have in medicine.
Totally agree! I've practiced clinical medicine in the community for the past 30 years and have not dealt with resident training. I can say that when I talk to a colleague about a patient, I don't generally add race. But I'll add whatever physical or social details I'd think important to the situation or question that I have (where at times, I'd imagine race might fall into).
It would seem attendings in academic training programs could just decide that race need not be included. But to frame it in some regimented, data-driven admonition seems a bit overdone and as you point out, can have the opposite effect of making trainees afraid to mention anything. I don't think we need to view patients as uniform beings dressed in tan jumpsuits to practice compassionate and high level medicine.
I think race is used by different people for different reasons. Imagine asking each presenter, "Why did you use race?"
"It came natural, like if I were to describe a person to anyone."
"I want you to picture them."
"I think they have Disease X, and I read that Race Y is at higher risk."
"I think they have Disease X, and in my experience, Race Y is at higher risk."
"We rarely care for Race Y here, I suspect for Reason X, so I thought it was worth noting."
"There are health disparities for Race Y, so I thought it was worth nothing."
"Race Y has a history of receiving disparate care in this (micro or macro setting), so I thought it was worth noting."
"Race Y is a social construct, and though imprecise, I believe still valuable marker for historical, social, and environmental influences on this patient's health."
"I'm of Race Y, and I felt a connection for Reason X."
"You asked for a thorough history."
Imagine all the learning that can happen from further discussion.
As doctors, we pride ourselves on using science to search for perfect answers. As humans, we should lean into the complexity of messy questions and messy answers, as I'd argue that's where much magic and meaning happen.
It seems to me that race does have a role in certain diseases. Black people dx with MS often have a more aggressive course of disease, as do the Japanese. Should this inform tx decisions?
Blacks are less likely to respond to ACEs and do well with diuretics to control BP. Shouldn’t that be top of mind for a PC?
nice piece; i want to question a little your dismissal of the medical importance of race. Race may be mostly a social construct and its biological basis may be commonly overestimated; but it is unclear that the biologial basis of our concept of race is “vanishingly small”. The caution of Oni-Orisan et al. in NEJM from 2021 that genetic ancestry can be medically important seems appropriate here: What would you make, for example, of Do No Harm’s critique of the removal of the race correction in estimated GFR formulas? If I get it right, their case is that in fact there are ancestry-related differences between blacks and whites (for this purpose, self-identification as one or the other is pretty accurate) in the relation of creatinine levels to measured GFR, most likely mediated by group differences in muscle mass. The race correction is (properly) aimed at getting an eGFR that corresponds most closely to a mGFR. Removing the correction amounts to implying an ideal of identical treatment for blacks and non-blacks having the same creatinine concentrations, when these indicate differing GFRs in the two groups—amounting to diagnoses of CKD at higher mGFRs in blacks than in non-blacks and more treatment for the one group than for the other at identical levels of renal function. Is that really what we want?
I have stopped using race as an identifier of the patient. "65 yom ...." What IS interesting is my black counterparts ALWAYS begin an H&P, chart documentation,etc with age - gender - race. My patient is my patient no matter age, gender, race, etc, and I treat them as such, equally across the board.
Came across this:
In 2020, the AMA adopted new policies that encourage medical education programs to recognize the harmful effects of using race as a proxy for biology in medical education
Race bears upon both prevalence and prognosis, and therefore is pertinent in the medical history. Is Sarcoidosis not more likely to explain hilar adenopathy in an African American than a caucasian? Is Thalassemia not more likely to explain anemia in an Italian than a Jew?
The shared DNA is a specious argument: we share 99% of our DNA with chimps. That still allows for major differences in health and disease.
We don't bolster our antiracism bona fides by claiming to be colorblind in the exam room. We undermine our claim on common sense.
Here’s another way of looking at it. Do you know the positive likelihood ratio of race for sarcoidosis in that population? I’ll bet you don’t because it would be nearly impossible to define given the variation within each race. How often do use other tests with no idea what the test characteristics are? No real benefit and lots of harm.
You’re right. I don’t. But a higher pretest probability aids your Bayesian analysis. This way you can be approximately right instead of precisely wrong 😉
I can't remember a single case where race was a useful bit of information in the presentation of a case. Perhaps it could be mentioned in the family history if it is at all relevant. It is probably just an unthinking repetition of a mode of presentation that the student has heard in the initial phase of their training. When my training began (early 70s) the terms "well nourished and well developed" were also used in the introductory sentence---probably out of habit. I would never allow a verbal presentation or discussion of a case at the patient's bedside.
I am a hematologist and chasing down Italian/Mediterranean/Sicilian extraction leads to more rapid diagnosis of the bleeding and RBC diatheses these populations seem to have more often than others. I have been told that we cannot discuss sickle cell as a race-enhanced disease because ANYONE might have sickle cell. This kind of argument may be "true" but is wrong. Assuming that you should not use EVERY fact about a patient to the best of their care is ALWAYS bad medicine -- showing one's moral purity by not checking on race/ethnicity disadvantages the patient sitting in front of you. Here is a surprise: you cannot test EVERYONE for sickle cell disease or b-thal or whatever. Clinical insight is based on all facts at hand...for some diseases race/ethnicity matters and you should use it.
Race is much more consequential than your overall article implies Dr Cifu. From the article you admired and linked:
Nonproblematic uses of race: In contrast to the examples described above, nonproblematic uses of race or ethnicity in the Textbook of Pediatric Care referred to race in the context of the social determinants of health or appropriately mentioned ancestry without conflating genetic differences and social determinants. We also considered quotes indicating that there is no difference in disease prevalence among various racial or ethnic groups as appropriate uses of race or ethnicity. In general, if a statement notes differences in disease prevalence among racial or ethnic groups, it is essential to provide the relevant social, historical, and political context for the statement.
Race has powerful associations with occupation, education, economic opportunity and SES Setting aside more biologic conditions such as triple negative breast cancer in women under 40, race linked conditions with strong social roots are everywhere, at least in a decent and evidence-based medical curriculum.
For instance, using self-identified race B v W, B patients face 2x higher rates of infant mortality, 2-3x increased maternal mortality, 2-4x elevated risk of amputation in diabetes, and 2-3 fold higher mortality rates in uterine cancer. Adjusting for age sex Medicaid eligibility and other clinical confounders, black kids still have 2.5 times higher odds of asthma hospitalization than white kids.
With all due respect, I think you are missing my primary point.
Of course race has enormous impact on how people are treated and thus every part of people's health and their outcomes, see the mortality graph I included and the whole middle section of the article. Medical education would be failing if this was not taught.
But, these huge differences are not due to biologic differences dictating outcomes.
It looks like I'll have to be the contrarian here.
I think Dr. Cifu's understanding of race is out of step with the biological and genetic literature. It is simply false to say that it is a social construct with little basis in reality. He undermines his argument by pointing out the life expectancy in the US by race; though he implies that this has something to do with racism, these charts and effects are seen all over the world, even in majority black countries.
IQ has been identified as a strong indicator of life outcomes and IQ is strongly hereditary. No matter where it is tested, we see the same type of group differences by race: blacks tend to score towards the bottom, hispanics slightly higher, then whites, and then asians at the top of the IQ range. No matter how these tests are conducted, the results are almost always the same.
The same things can be seen across a host of other outcomes across different races. Life expectancy, criminality, rate of STIs, economic growth, etc. we see the same pattern; blacks toward the bottom, then hispanics, then whites, and lastly asians at the top.
Not to mention that stereotype threat is among the most reproduced sociological findings ever. People's stereotypes of other groups tend to be more accurate than they are inaccurate.
Saying that two groups of humans are more similar than they are different is just a red herring. We have 98% genetic similarity with chimpanzees, would we seriously say that the difference between humans and chimpanzees is solely a social construct?
The idea that every human is a tabula rasa has been roundly disproven. While our genetics are not 100% determinative of our future, it is simply impossible and wrong to ignore that they have an outsized effect on our future and in ways that are not immediately obvious to us.
Just seems that race is one of many pieces that define a patient. To remove that fact for righteous reasons or for any reason diminishes our picture.
I mostly agree. Mostly, i say, primarily because of my bias towards avoiding absolutes. There may be times when someone’s race (or other social constructs) may be appropriate identifiers in a chief complaint or the HPI. Certainly relevant in the social history.
I’ve also stopped putting the race of the patient in my own notes long ago and feel the same as you do, Dr. Cifu. Great piece.
Great read, better to think of individual risks than collective undertones. This is what patients want as well. To the extent that racial constructs help patients to be treated as unique individuals, I’m all for it. To the extent that it perpetuates harmful stigma, we should seek to avoid it. Seems to me the problem is in detecting the difference… who’s to say? This is what makes the topic so pernicious. It can be used to traffic really harmful ideas, so good to expand (rather than mute) conversation and be clearer about instances of systemic racism (eGFR adjustments in Neph Transplant), instances of true risk stratification (HLA B*5801 testing in southeast Asian populations before prescribing allopurinol), and instances when it’s aimless, lazy, and unfortunately normative to make race ‘the identifier’ of a social history. I’d rather know what patients do for a living and how they spend their free time. Maybe we just need to be clearer about our intentions (when race is invoked) and better at appraising true vs perceived risks related to this complex construct.
I like “collective undertones”!
Good stuff, interesting points. On the one hand, I don't think I disagree with your actual argument. On the other hand, I wonder if including/not including race might depend more on the clinical circumstances than a rule. For example, I'm thinking of cases where the details of the HPI are heavily "psychosocial", in which omitting the race of the patient would actually leave out important context and would seem out of place relegated to the social history. Not that these are necessarily the majority of cases, but there are certainly some, in my opinion. Perhaps in psych we see relative more of this?
Also—and this isn't an argument against your stance but just a concern—I worry about people taking the wrong message from this and viewing it primarily as an issue of "not being judgmental/stigmatizing". eg. I remember as PGY-4 teaching interns about the psychiatric evaluation and stressing the importance of observing the physical characteristics of the patient. In one case, an intern omitted reference to a patient's numerous face tattoos which were quite obviously gang-related. She omitted this because she was worried about "stigma". Obviously face tattoos and race are not the same since we can choose to get tattoos but can't choose our genes, but the general habit of "I'm not going to mention what I'm seeing in front of me due to fear" is not a good one to have in medicine.
Totally agree! I've practiced clinical medicine in the community for the past 30 years and have not dealt with resident training. I can say that when I talk to a colleague about a patient, I don't generally add race. But I'll add whatever physical or social details I'd think important to the situation or question that I have (where at times, I'd imagine race might fall into).
It would seem attendings in academic training programs could just decide that race need not be included. But to frame it in some regimented, data-driven admonition seems a bit overdone and as you point out, can have the opposite effect of making trainees afraid to mention anything. I don't think we need to view patients as uniform beings dressed in tan jumpsuits to practice compassionate and high level medicine.
Excellent points!
Always love reading your articles, Dr. Cifu.
I think race is used by different people for different reasons. Imagine asking each presenter, "Why did you use race?"
"It came natural, like if I were to describe a person to anyone."
"I want you to picture them."
"I think they have Disease X, and I read that Race Y is at higher risk."
"I think they have Disease X, and in my experience, Race Y is at higher risk."
"We rarely care for Race Y here, I suspect for Reason X, so I thought it was worth noting."
"There are health disparities for Race Y, so I thought it was worth nothing."
"Race Y has a history of receiving disparate care in this (micro or macro setting), so I thought it was worth noting."
"Race Y is a social construct, and though imprecise, I believe still valuable marker for historical, social, and environmental influences on this patient's health."
"I'm of Race Y, and I felt a connection for Reason X."
"You asked for a thorough history."
Imagine all the learning that can happen from further discussion.
As doctors, we pride ourselves on using science to search for perfect answers. As humans, we should lean into the complexity of messy questions and messy answers, as I'd argue that's where much magic and meaning happen.
So well put. Thanks.
Adam
It seems to me that race does have a role in certain diseases. Black people dx with MS often have a more aggressive course of disease, as do the Japanese. Should this inform tx decisions?
Blacks are less likely to respond to ACEs and do well with diuretics to control BP. Shouldn’t that be top of mind for a PC?
nice piece; i want to question a little your dismissal of the medical importance of race. Race may be mostly a social construct and its biological basis may be commonly overestimated; but it is unclear that the biologial basis of our concept of race is “vanishingly small”. The caution of Oni-Orisan et al. in NEJM from 2021 that genetic ancestry can be medically important seems appropriate here: What would you make, for example, of Do No Harm’s critique of the removal of the race correction in estimated GFR formulas? If I get it right, their case is that in fact there are ancestry-related differences between blacks and whites (for this purpose, self-identification as one or the other is pretty accurate) in the relation of creatinine levels to measured GFR, most likely mediated by group differences in muscle mass. The race correction is (properly) aimed at getting an eGFR that corresponds most closely to a mGFR. Removing the correction amounts to implying an ideal of identical treatment for blacks and non-blacks having the same creatinine concentrations, when these indicate differing GFRs in the two groups—amounting to diagnoses of CKD at higher mGFRs in blacks than in non-blacks and more treatment for the one group than for the other at identical levels of renal function. Is that really what we want?
I have stopped using race as an identifier of the patient. "65 yom ...." What IS interesting is my black counterparts ALWAYS begin an H&P, chart documentation,etc with age - gender - race. My patient is my patient no matter age, gender, race, etc, and I treat them as such, equally across the board.
Came across this:
In 2020, the AMA adopted new policies that encourage medical education programs to recognize the harmful effects of using race as a proxy for biology in medical education
I'm going to have to disagree.
Race bears upon both prevalence and prognosis, and therefore is pertinent in the medical history. Is Sarcoidosis not more likely to explain hilar adenopathy in an African American than a caucasian? Is Thalassemia not more likely to explain anemia in an Italian than a Jew?
The shared DNA is a specious argument: we share 99% of our DNA with chimps. That still allows for major differences in health and disease.
We don't bolster our antiracism bona fides by claiming to be colorblind in the exam room. We undermine our claim on common sense.
Here’s another way of looking at it. Do you know the positive likelihood ratio of race for sarcoidosis in that population? I’ll bet you don’t because it would be nearly impossible to define given the variation within each race. How often do use other tests with no idea what the test characteristics are? No real benefit and lots of harm.
You’re right. I don’t. But a higher pretest probability aids your Bayesian analysis. This way you can be approximately right instead of precisely wrong 😉
This feels like a personal attack... ;-)
If you see a McDougal with an MCV of 60, do you ask her her maiden name? Every time!
I can't remember a single case where race was a useful bit of information in the presentation of a case. Perhaps it could be mentioned in the family history if it is at all relevant. It is probably just an unthinking repetition of a mode of presentation that the student has heard in the initial phase of their training. When my training began (early 70s) the terms "well nourished and well developed" were also used in the introductory sentence---probably out of habit. I would never allow a verbal presentation or discussion of a case at the patient's bedside.
I am a hematologist and chasing down Italian/Mediterranean/Sicilian extraction leads to more rapid diagnosis of the bleeding and RBC diatheses these populations seem to have more often than others. I have been told that we cannot discuss sickle cell as a race-enhanced disease because ANYONE might have sickle cell. This kind of argument may be "true" but is wrong. Assuming that you should not use EVERY fact about a patient to the best of their care is ALWAYS bad medicine -- showing one's moral purity by not checking on race/ethnicity disadvantages the patient sitting in front of you. Here is a surprise: you cannot test EVERYONE for sickle cell disease or b-thal or whatever. Clinical insight is based on all facts at hand...for some diseases race/ethnicity matters and you should use it.
That is why a thorough history always includes family history.
A useful book for those interested
Classified: The Untold Story of Racial Classification in America
by David E Bernstein
Bombardier Books, 2022
Race is much more consequential than your overall article implies Dr Cifu. From the article you admired and linked:
Nonproblematic uses of race: In contrast to the examples described above, nonproblematic uses of race or ethnicity in the Textbook of Pediatric Care referred to race in the context of the social determinants of health or appropriately mentioned ancestry without conflating genetic differences and social determinants. We also considered quotes indicating that there is no difference in disease prevalence among various racial or ethnic groups as appropriate uses of race or ethnicity. In general, if a statement notes differences in disease prevalence among racial or ethnic groups, it is essential to provide the relevant social, historical, and political context for the statement.
Race has powerful associations with occupation, education, economic opportunity and SES Setting aside more biologic conditions such as triple negative breast cancer in women under 40, race linked conditions with strong social roots are everywhere, at least in a decent and evidence-based medical curriculum.
For instance, using self-identified race B v W, B patients face 2x higher rates of infant mortality, 2-3x increased maternal mortality, 2-4x elevated risk of amputation in diabetes, and 2-3 fold higher mortality rates in uterine cancer. Adjusting for age sex Medicaid eligibility and other clinical confounders, black kids still have 2.5 times higher odds of asthma hospitalization than white kids.
No role in medical education, sir??
With all due respect, I think you are missing my primary point.
Of course race has enormous impact on how people are treated and thus every part of people's health and their outcomes, see the mortality graph I included and the whole middle section of the article. Medical education would be failing if this was not taught.
But, these huge differences are not due to biologic differences dictating outcomes.
I know you know this.
I am a South Asian and race definitely plays a role in my Cardio Vascular risk.
Adam, u r always so grounded and thoughtful. Will share w colleagues