The New England Journal of Medicine recently published an article entitled “Hidden in Plain Sight—Reconsidering the Use of Race Correction in Clinical Algorithms” about race and racism in medicine, specifically looking at algorithms and calculators that include race (or more often, Black v. non-Black) in their inputs. (June 17, 2020; https://bit.ly/3iqGYiT.)
We thought long and hard about how MDCalc (https://www.mdcalc.com/) should respond, and we eventually came up with a patient care-based policy statement that would support our goals: Help clinicians deliver the best quality care to patients and use evidence in their practices. (Read it at www.mdcalc.com/race.)
We decided to summarize briefly how race affects the results of an algorithm or score so clinicians can be better informed about the score they're using and make race an optional input when possible, allowing clinicians to opt in or out of including it. We will also specifically draw attention to race when it exists in a calculator in the instructions sections and evidence content to provide clearer, transparent information for our users.
As I looked deeper, I realized the whole thing is really, really messy. You've got:
- A history of race, racism, and racial stereotypes being used in medicine in the past and present;
- “Race” being a really blunt instrument to group patients together when what we really want is an individual's genetic makeup;
- A current lack of ability to assess genetic makeup easily;
- Very little medical knowledge that looks at ancestry or lineage to inform care, including diagnosis, prognosis, and treatment;
- “Non-Black” being the default in these studies because they mostly took place in the United States, which assumes Black and non-Black are the only two categories that matter;
- No way to address multiracial patients;
- The fact that some diseases do seem to be more or less prevalent depending on ancestry, which we often less accurately label as “race.”
While I won't cover all the specific examples in our statement, this got me wondering again about our heuristics in medicine. How and why have we decided to see certain trends and correlations and use those but not others? It's like my dog: I've trained her to sit on command, and she knows she's going for a walk when I put on my socks. She somehow also knows when I'm about to give her a bath, and she hides. There must be something I'm doing that she's picked up on as a pattern, but I'm completely oblivious.
Why is it “fat fertile female 40” for cholecystitis when there are so many men with it? Or the “worst headache of your life” for subarachnoid hemorrhage, not “sudden-onset headache?” Or “reheated rice” for Bacillus cereus or “clindamycin” for Clostridioides difficile diarrhea? (Some of these may ring true in your experience, while others don't at all. I reheat my rice all the time and happily risk every bite.)
This is the challenge of our craft, which is part science, part art, part profession, and part trade. The problem is that some of these teachings and mantras are true, while others are stereotypes, some are old wives' tales, and some are based on one doctor's case report disseminated by a medical journal and taken as medical fact. What any doctor has in his head is a uniquely human mix of medical school textbook training, mnemonics, residency training, edicts issued by the attendings we respected and trusted, cases from residency and beyond that went well or poorly, trends and patterns our brains consciously and subconsciously absorb, heuristics and hunches, and probably least of all, evidence-based medicine. Often our brains turn that into a gut feeling, Spidey sense, or gestalt so we can generate a diagnosis and differential in less than a second. Pretty amazing but not without fault. Just like in any data analysis, including our brains, if it's garbage in, it's garbage out.
Take the disease-causing cystic fibrosis (CF) resulting from a mutation in the CFTR protein. CFTR mutations that can cause CF are present in up to one in 25 people of Northern European ancestry. One single mutation is the cause of about half of CF cases, but they have found more than 1500 other mutations in the gene. I definitely think of CF as a Northern European disease, to the extent that I probably don't consider it when I see certain children in the ED who don't look Northern European to me. That's my gestalt brain talking. The factual brain, however, knows that CF can present in people who don't look Northern European and that I've met people with CF who don't fit the stereotype. Yet my factual brain has to work really hard to correct the gestalt brain's error.
It takes a lot of effort to fix our gestalt brains, especially once they have formed. We need more evidence and better evidence in them as foundations in medical school and residency because once they get cemented, they are set. Like cement.
Similarly, we need to move away from race and toward ancestry as much as we can. The promise of cracking the genetic code is decades away, and most patients can't wait that long. Ancestry, for now, might help us while we wait for a bedside genetic decoder test. Race and racism will unfortunately be with us for a very long time, but we can do our part by trying to move research and evidence away from blunt pigeonholes like race.
Share this article on Twitter and Facebook.
Access the links in EMN by reading this on our website, www.EM-News.com.
Comments? Write to us at firstname.lastname@example.org.
Dr. Walkeris an emergency physician at Kaiser San Francisco. He is the developer and co-creator of MDCalc (www.mdcalc.com), a medical calculator for clinical scores, equations, and risk stratifications, which also has an app (http://apps.mdcalc.com/), and The NNT (www.thennt.com), a number-needed-to-treat tool to communicate benefit and harm. Follow him on Twitter@grahamwalker, and read his past columns athttp://bit.ly/EMN-Emergentology.