More than the Sum of Its Parts

Figure: Ecological fallacies, racial inequities, error of inference, eGFR calculation, pediatric urinary tract infections, extreme skepticismA common method for measuring kidney function, the estimated glomerular filtration rate, or eGFR, was calculated for two decades by combining serum creatinine levels with certain other patient characteristics that included age, sex, height, weight—and race. Why was race a modifier for eGFR? The authors of the paper that proposed the calculation observed that Black patients had higher rates of creatinine excretion than non-Black patients. (Ann Intern Med. 1999;130[6]:461.) Based on these and similar findings in a report from the National Health and Nutrition Examination Survey from the year before the paper was published, the researchers concluded that Black race was an independent predictor of GFR, theorizing—without underlying biological support—that Black patients have more muscle mass than white patients. This method of calculating eGFR, which increases the estimate of Black patients' kidney function by more than 20 percent compared with non-Black patients, was widely adopted in U.S. hospitals and medical practices. The effect: artificially making the kidney function of Black patients seem healthier than it was, even though Black Americans are four times more likely to develop kidney failure than white patients. It wasn't until 2022 when a study by a fourth-year medical resident at Yale University pointed out the faults in this calculation that institutions began removing the race element from their eGFR calculations. This racially biased error is an example of a common epidemiological problem in medicine called the ecological fallacy. Treatment Implications “This error of inference attributes differences between populations to individuals,” explained Daniel Gingold, MD, MPH, an associate professor of emergency medicine and the assistant director of population health at the University of Maryland Medical Center. “Many medical teachings highlighting the importance of race in medical decision-making and treatment are susceptible to this error. This includes derived risk scores, calculation of physiologic parameters, and treatment recommendations. Not only are these tools inaccurate but also ingrain bias against historically marginalized groups and perpetuate health disparities.” The eGFR calculation, which Dr. Gingold noted was taught to him in medical school, did not consider that the higher average and distribution of creatinine in Black patients might simply reflect the fact that kidney disease is more prevalent in that population—which is the case—and the experts just corrected for it. “They took distributions that did overlap but were slightly different and made them match, which underestimates disease prevalence in Black populations,” he said. “That, of course, had other knock-on effects: goal-directed therapy and guideline-based treatment doesn't happen as often as it should, and Black patients with kidney disease receive fewer kidney transplants than patients of other races.” Indeed, the Yale student's paper estimated that removing the race modifier from eGFR calculations would allow 3.3 million Black Americans to cross the threshold for diagnosis of stage 3 chronic kidney disease; 300,000 more to qualify for beneficial nephrologist referral; and 31,000 more to become eligible for transplant evaluation and waitlist inclusion. That particular ecological fallacy has since been corrected, but other similar epidemiologic errors continue to be made at the general practice and individual level. “For example, recent trainees have told me that institutions are still teaching that calcium channel blockers work better in Black patients,” Dr. Gingold said. “While it is true that the mean response to blood pressure medications and different classes of these drugs does vary slightly across racial divisions, that doesn't mean you should take it to mean that we should give all Black patients the drug class in which the mean response of Black patients is 2 mm Hg different. The reality is that they overlap a great deal, and there are other factors that drive response to blood pressure medications, many of which are more impactful than race.” Baking in Inequities Another commonly employed clinical “rule” is a risk calculator for pediatric urinary tract infections aimed at decreasing the need for invasive catheter-based testing in very young children. “This one was recently revised, and the way it was changed is particularly interesting,” Dr. Gingold said. “Statistics showed differences in the prevalence of UTIs in children depending on their race. This, again, was a genuine finding, but it wasn't because of any sort of physiologic difference. Instead, it likely has more to do with health care utilization and access to care, especially emergency care. So, the calculator as it was put out had Black race as allegedly ‘protective’ against having a pediatric UTI with non-Black race as a ‘risk factor.’” This might mean that that a physician is more likely to request a urine sample in a young febrile child if he is not Black. “Again, the obvious implication would be the tendency to underrecognize, undertest, and undertreat UTIs in Black children, further exacerbating disparities in care,” he said. “When they went back and took race out of the model, they found that it worked just as well, especially if you allow for other clinically important variables. When you remove race, what pops out as significant is history of previous UTI and duration of fever. Race was masking the importance of those variables.” The authors of the article, “Hidden in Plain Sight—Reconsidering the Use of Race Correction in Clinical Algorithms,” listed the eGFR and UTI tools among more than a dozen such risk scores, calculators, and tools that employ “race correction” in clinical medicine, which could contribute to perpetuating or exacerbating race-based health inequities. (New Engl J Med. 2020;383[9]:874; https://tinyurl.com/yc2jp88e.) “Most race corrections implicitly, if not explicitly, operate on the assumption that genetic difference tracks reliably with race,” they wrote. “If the empirical differences seen between racial groups were actually due to genetic differences, then race adjustment might be justified: different coefficients for different bodies. Such situations, however, are exceedingly unlikely.” The authors noted that the genetic structure of human populations is more varied within racial groups than between them and racial differences found in large data sets most likely reflect effects of racism—“the experience of being Black in America rather than being Black itself—such as toxic stress and its physiological consequences.” Correcting for race in these cases would not address the cause of the disparity, they said, and risks “baking inequity into the system” if that deters physicians from offering services to patients. Ecological fallacy in medicine often manifests around race, but that's not the only categorization that can facilitate erroneous approaches to clinical decision-making. “We typically hear that men who have sex with men (MSM) are at higher risk for sexually transmitted diseases, and while there certainly are STDs that are more prevalent in MSM communities, that is not so much associated with male-on-male sex being inherently more dangerous,” Dr. Gingold said. “Rather, it is primarily associated with the number of partners. Anyone with a high number of partners has a similarly high risk for STDs.” He said it's important for physicians not to overly focus on orientation when the biggest factors are number of sexual partners and frequency of unprotected sex. “Being clear about the underlying cause that is the driver of population-based differences is important when it comes to individuals,” Dr. Gingold said. “That's what you as the clinician need to focus on with the individual in front of you, so that you don't miss factors that are likely physiologically and clinically important.” Extreme Skepticism That can be particularly challenging in the emergency department, said Patrick Croskerry, MD, PhD, a professor of emergency medicine at Dalhousie University in Halifax, Nova Scotia, Canada. “The conditions of the emergency department are often error-producing,” he said. “As with most of human performance, performance in the ED follows an inverted U function. As you increase the level of activity and busyness in the ED, you will improve performance, but then it plateaus and starts to decline. “And when stress is put on the situation, such as an extremely high number of patients or many difficult patients or the level of uncertainty goes up, then my tendency to make assumptions and be influenced by biases or fallacies goes up, even though I'm trying to be rational and objective in a clinical situation.” And in this environment, he noted, the emergency physician—particularly the overburdened emergency medicine resident—is more likely to rely on clinical prediction rules, calculators, and checklists that may have built-in ecological fallacies. “The conditions of an emergency department promote the use of heuristics—basically taking mental shortcuts,” Dr. Croskerry said. “And a lot of the time, the shortcuts work. But emergency physicians need to be aware that those heuristics need to be challenged occasionally. You can't afford to go along with them just because it's convenient. You have to examine them.” Dr. Gingold agreed. “The easiest thing to do is have extreme skepticism, if not dismissal, of the explicit bias that we may absorb in training,” he said. “I went to medical school a little more than 10 years ago, and even that recently, it's amazing how much I was taught that different diseases affect different groups because of X, and that a particular race is in and of itself a ‘risk factor’ for different diseases. “In general, that is scientifically incorrect and has ethical implications as well. It's very difficult to change implicit bias, but at least you can try to avoid perpetuating explicit differences in treatment based on factors that don't make any sense. It is helpful always to be checking our implicit and unconscious assumptions of the way people are based on the way they look. It's always worth trying to watch yourself to make sure you're not taking shortcuts in that particular way. That's a lifelong project, and it's not only based on race.” Ecological Fallacies in Practice American Heart Association's Get with the Guidelines-Heart Failure Program The Ecological Fallacy: This calculator adds three points to the risk score for heart failure if the patient is not Black. The Effect: This classifies Black patients as lower risk and may deter physicians from using resources for them. The STONE Score The Ecological Fallacy: A higher score denotes a higher risk of a ureteral stone for a patient with flank pain; three points are added for non-Black race. The Effect: A lower score for Black patients may prompt physicians not to evaluate them as thoroughly. The Urinary Tract Infection Calculator The Ecological Fallacy: This tool assesses the risk of UTI in children 2 to 23 months to determine whether to perform urine testing; it designates a lower score for a child who is Black. The Effect: This lower score for Black children may dissuade physicians from pursuing definitive diagnostic testing for them. Pulmonary Function Tests The Ecological Fallacy: U.S. spirometers employ correction factors for Black (10-15%) and Asian (4-6%) patients. The Effect: This may misclassify disease severity and impairment for conditions such as asthma and COPD. Source: Adapted from New Engl J Med. 2020;383[9]:874; https://tinyurl.com/yc2jp88e. MS. SHAW is a freelance writer with more than 20 years of experience writing about health and medicine. She is also the author of Having Children After Cancer, the only guide for cancer survivors hoping to build their families after a cancer diagnosis. You can find her work at www.writergina.com. Follow her on X @writergina. Read her past articles at http://tinyurl.com/EMN-Shaw. Share this article on X and Facebook. Access the links in EMN by reading this on our website: www.EM-News.com. Comments? Write to us at [email protected].

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.