Collaborating Authors

Addressing racial inequities in medicine


COVID-19 inequitably affects marginalized racial and ethnic populations across the world. A review of more than 17 million adult patients in the United Kingdom revealed a nearly twofold risk of death from COVID-19 among Black and Asian populations compared with white populations ([ 1 ][1]). Black people comprise 12.5% of the US population, yet they account for more than 18% of COVID-19 associated deaths ([ 2 ][2]). Although Black and Latinx populations in the US experience higher rates of infection, hospitalization, and deaths compared with white populations, they have similar case fatality rates, suggesting that there is no innate vulnerability or susceptibility to COVID-19 ([ 3 ][3]). Persistent COVID-19 racial and ethnic inequities are likely caused by structural racism that results in an increased risk of exposure and inadequate health care access in communities of color ([ 3 ][3]). Structural racism, or the discriminatory policies, practices, and systems that reinforce an unequal distribution of power and resources in social institutions, is a considerable driver of health inequities ([ 4 ][4]). The COVID-19 pandemic has exacerbated the socioeconomic disadvantages that have led to an overrepresentation of marginalized groups in service-industry jobs and excessive financial insecurities that affect how people live and their ability to access health care ([ 4 ][4]). Communities of color are more likely to reside in shared and congregate housing, use public transportation for commuting, and support themselves and their families with low-income jobs. These factors increase COVID-19 exposure, because of the likelihood for more person-to-person contact, and limit the ability to self-isolate. Additionally, these social conditions increase a person's risk for chronic conditions, including hypertension, diabetes, cardiovascular disease, and obesity, that are associated with more severe COVID-19 outcomes ([ 5 ][5]). A broad, multisector community-engaged response is required to address the social determinants of health and achieve health equity with COVID-19 and other diseases. The medical and scientific community must be a willing partner in dismantling structural racism in society, which includes an emphasis on addressing the socioeconomic, environmental, and behavioral factors that influence most health outcomes. Although these efforts are critical, there is also an urgent need and opportunity to dismantle structural racism within the traditional functions of the medical and scientific community (see the figure). This necessitates several actions in clinical practice, medical education, and research: acknowledging that race (group categorizations based on physical traits such as skin color) and ethnicity (groups defined by shared language, history, religion, and culture) are social and political constructs and are a poor proxy for ancestry (inherited genetic variations traced to geographical origins of a person's ancestors); uprooting the racism that is deeply embedded in perspectives, policies, and practices in health systems; and co-creating a new system that incorporates the voices of marginalized populations who are unjustly suffering with unmet social needs and adverse health outcomes related to COVID-19 and other diseases. Quality improvement activities are a core element of health care, but an equity lens has not routinely been included ([ 6 ][6]). Annual data from US hospitals and health systems show that substantial inequities remain in the quality of care between racial and ethnic groups ([ 7 ][7]). Marginalized racial and ethnic groups in the US are less likely to receive clinically necessary and routine treatments for several diseases, including cancer and cardiovascular disease, as well as mental health disorders, compared with their white counterparts ([ 7 ][7]). To begin effectively operationalizing an equity lens in quality improvement, clinics, hospitals, and health systems should collect and stratify data by self-identified racial and ethnic categories and make these data transparent to leaders and clinicians. This will allow previously unnoticed differences in quality of care, due to implicit or explicit bias, to be addressed with system-level changes that facilitate quality improvement. Providing quality, equitable care also requires a reexamination of how clinical algorithms are used in medicine. The use of race in clinical decision support tools that provide guidance on medications, treatments, or procedures can delay access to care and further exacerbate health inequities ([ 8 ][8]). For example, in pediatrics, a calculator has been developed to predict the likelihood of a urinary tract infection (UTI). If the child is Black, the calculator predicts a lower risk of UTI, which may delay diagnostic tests and treatments for Black patients ([ 8 ][8]). This tool provides neither an evidence-based biological nor social explanation for why race has been added. Additionally, some hospitals have removed race from the calculation of kidney function estimates, or glomerular filtration rate (GFR). In a recent study, removing race from the GFR could allow up to a third of Black patients to be reclassified to a more severe stage of kidney disease and prompt appropriate referrals for dialysis and transplant ([ 9 ][9]). There is evidence of variation in the GFR within racial categories, suggesting differences based on ancestry and not race ([ 10 ][10]). Within these algorithms, clinicians and scientists should consider opportunities to include more specificity, such as country of origin, an explanation of the genetic association from a geographic region, and the social factors linked to identifying as a particular race in various regions around the world. Without this context, the sole inclusion of race serves to reinforce the false narrative of genetic or biological racial differences and disregard the socioeconomic conditions that largely influence health outcomes. Beliefs about biological differences between races are often explicitly or implicitly embedded in medical education ([ 11 ][11]). Some medical students and residents believe that Black people are not as sensitive to pain or are more prone to blood clots ([ 12 ][12]). Other physicians incorrectly associate diseases with race instead of ancestry. For example, it is common to hear an association between sickle cell disease and Black people versus examining it as an ancestrally associated disease common in populations at risk for malaria ([ 11 ][11]). Challenging beliefs and assumptions about biological differences requires increased education about the implications of racism and racial categorizations. White Coats for Black Lives, a US medical student group, developed a racial justice report card to evaluate curricula, policies, and practices to dismantle racism in medical school. Their evaluation includes an analysis of the pedagogy on race as a social and political construct, rather than solely based in science ([ 13 ][13]). These students and other faculty champions across the US are helping to catalyze changes in medical school curricula. Historically accurate and transformative education for learners, physicians, and scientists should be made available to counteract the subtle and unfounded teachings that have persisted along the medical education continuum. Medical and scientific professionals should have a critical understanding of the unjust social conditions that contribute to inequities as well as their root in structural racism and other systems of oppression. Historical injustices, both in society and in medical institutions, should be taught and opportunities provided to participate in reversing oppressive systems-level policies and practices. Foundational and continuing education should also be made available on population genetics, migration patterns, and expression of genes due to environmental exposures related to their impact on diseases in the global community. Evidence-based education that describes both the social impact of race and genetic ancestry will help the medical and scientific community identify the underlying critical factors that cause health inequities. In research studies, race is often used as a variable with imprecise definitions and without questioning what the variables measure ([ 14 ][14]). For example, a research study comparing Black people to other racial and ethnic groups could include three people who self-identify as “Black” with varying backgrounds: someone who is biracial with a white parent while living in the US, someone who is multiracial with an Asian parent living in the Caribbean, or someone with two parents from and living in West Africa. Depending on their individual decisions and the communities they identify with, those three people may change which racial category they identify with over time. Grouping the individuals into the same analysis that studies a “Black” population may overlook extensive genetic variation between them. Research questions, analyses, and conclusions made solely on race should account for the complex and imprecise socially constructed categories of race ([ 14 ][14]). Analyses are incomplete if they do not also reflect the impact of unfair social advantage and disadvantage experienced by racial and ethnic groups worldwide. Several studies show an association between perceived racism and poor health indicators such as high blood pressure, coronary artery calcification, and breast cancer incidence ([ 15 ][15]). Research investigations should include defining the racial and ethnic categories and clarifying the impact of both social factors on what is being studied and whether conclusions should be based on ancestry instead of race. There is value in continuing to include race in research to elucidate root causes and helpful interventions to address health inequities if more specificity and rigor are added to what race is serving as a proxy for in the study. For this research to have greater impact and advance health equity, efforts should be made to diversify the biomedical and scientific workforce, include multidisciplinary scholars in study designs, and increase participation of marginalized populations in research studies. ![Figure][16] Remedying racial inequities Taking action in the medical and scientific community requires a rigorous analysis of the use of race in medical education, clinical care, and research; continuous institutional assessment using an equity lens; and partnering with local communities to advance health equity. GRAPHIC: N. CARY/ SCIENCE One of the most important lessons from the COVID-19 pandemic is that the medical community waited too long to become trustworthy partners with marginalized communities. The systems were not in place to hear their concerns, respond quickly in areas that were disproportionately affected by COVID-19 cases, or promptly build vaccine confidence. To address both the health inequities in local communities and the health care inequities within hospitals and clinics, the medical and scientific community must listen and adopt shared leadership models that prioritize communities in health interventions, including COVID-19 vaccine development and distribution. This begins with recruiting and retaining a diverse health care workforce that reflects the local community and ensuring broader access to care before, during, and after public health crises. Institutions must prioritize and support these efforts in clinical care, medical education, and research and embed their work through the lens of the communities served. Processes for continuous review and modification of systems-level policies and practices should be integrated into every aspect of health care to eliminate inequities. As global partners and leaders, the medical and scientific communities must make bold changes for the world to overcome two public health crises—racism and COVID-19. The greatest of these is to resist any further passivity and delay toward remedying social injustices and the structural racism that has proven beyond the COVID-19 pandemic that the burden is too heavy to further ignore it. 1. [↵][17]1. E. J. Williamson et al ., Nature 584, 430 (2020). [OpenUrl][18][CrossRef][19][PubMed][20] 2. [↵][21]1. J. A. W. Gold et al ., MMWR Morb. Mortal. Wkly. Rep. 69, 1517 (2020). [OpenUrl][22][CrossRef][23][PubMed][20] 3. [↵][24]1. K. Mackey et al ., Ann. Intern. Med. 174, 362 (2021). [OpenUrl][25] 4. [↵][26]1. L. E. Egede, 2. R. J. Walker , N. Engl. J. Med. 383, e77 (2020). [OpenUrl][27][CrossRef][28][PubMed][20] 5. [↵][29]1. S. Singu, 2. A. Acharya, 3. K. Challagundla, 4. S. N. Byrareddy , Front. Public Health 8, 406 (2020). [OpenUrl][30] 6. [↵][31]Institute of Medicine (US) Committee on Quality of Health Care in America, Crossing the Quality Chasm: A New Health System for the 21st Century (National Academies Press, 2001). 7. [↵][32]Agency for Healthcare Research and Quality, “2019 national healthcare quality and disparities report” (U.S. Department of Health and Human Services, 2020); [][33]. 8. [↵][34]1. D. A. Vyas, 2. L. G. Eisenstein, 3. D. S. Jones , N. Engl. J. Med. 383, 874 (2020). [OpenUrl][35][CrossRef][36][PubMed][20] 9. [↵][37]1. S. Ahmed et al ., J. Gen. Intern. Med. 36, 464 (2021). [OpenUrl][38][PubMed][39] 10. [↵][40]1. B. W. Teo et al ., Adv. Chronic Kidney Dis. 25, 41 (2018). [OpenUrl][41][PubMed][42] 11. [↵][43]1. C. Amutah et al ., N. Engl. J. Med. 384, 872 (2021). [OpenUrl][44][CrossRef][45][PubMed][46] 12. [↵][47]1. K. M. Hoffman, 2. S. Trawalter, 3. J. R. Axt, 4. M. N. Oliver , Proc. Natl. Acad. Sci. U.S.A. 113, 4296 (2016). [OpenUrl][48][Abstract/FREE Full Text][49] 13. [↵][50]White Coats for Black Lives, Racial justice report card (2019); . 14. [↵][51]1. R. R. Hardeman, 2. J. Karbeah , Health Serv. Res. 55, 777 (2020). [OpenUrl][52] 15. [↵][53]1. D. R. Williams, 2. S. A. Mohammed , J. Behav. Med. 32, 20 (2009). [OpenUrl][54][CrossRef][55][PubMed][56][Web of Science][57] [1]: #ref-1 [2]: #ref-2 [3]: #ref-3 [4]: #ref-4 [5]: #ref-5 [6]: #ref-6 [7]: #ref-7 [8]: #ref-8 [9]: #ref-9 [10]: #ref-10 [11]: #ref-11 [12]: #ref-12 [13]: #ref-13 [14]: #ref-14 [15]: #ref-15 [16]: pending:yes [17]: #xref-ref-1-1 "View reference 1 in text" [18]: {openurl}?query=rft.jtitle%253DNature%26rft.volume%253D584%26rft.spage%253D430%26rft_id%253Dinfo%253Adoi%252F10.1038%252Fs41586-020-2521-4%26rft_id%253Dinfo%253Apmid%252Fhttp%253A%252F%252Fwww.n%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [19]: /lookup/external-ref?access_num=10.1038/s41586-020-2521-4&link_type=DOI [20]: /lookup/external-ref?access_num=http://www.n&link_type=MED&atom=%2Fsci%2F372%2F6540%2F348.atom [21]: #xref-ref-2-1 "View reference 2 in text" [22]: {openurl}?query=rft.jtitle%253DMMWR%2BMorb.%2BMortal.%2BWkly.%2BRep.%26rft.volume%253D69%26rft.spage%253D1517%26rft_id%253Dinfo%253Adoi%252F10.15585%252Fmmwr.mm6942e1%26rft_id%253Dinfo%253Apmid%252Fhttp%253A%252F%252Fwww.n%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [23]: /lookup/external-ref?access_num=10.15585/mmwr.mm6942e1&link_type=DOI [24]: #xref-ref-3-1 "View reference 3 in text" [25]: {openurl}?query=rft.jtitle%253DAnn.%2BIntern.%2BMed.%26rft.volume%253D174%26rft.spage%253D362%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [26]: #xref-ref-4-1 "View reference 4 in text" [27]: {openurl}?query=rft.jtitle%253DN.%2BEngl.%2BJ.%2BMed.%26rft.volume%253D383%26rft.spage%253De77%26rft_id%253Dinfo%253Adoi%252F10.1056%252FNEJMp2023616%26rft_id%253Dinfo%253Apmid%252Fhttp%253A%252F%252Fwww.n%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [28]: /lookup/external-ref?access_num=10.1056/NEJMp2023616&link_type=DOI [29]: #xref-ref-5-1 "View reference 5 in text" [30]: {openurl}?query=rft.jtitle%253DFront.%2BPublic%2BHealth%26rft.volume%253D8%26rft.spage%253D406%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [31]: #xref-ref-6-1 "View reference 6 in text" [32]: #xref-ref-7-1 "View reference 7 in text" [33]: [34]: #xref-ref-8-1 "View reference 8 in text" [35]: {openurl}?query=rft.jtitle%253DN.%2BEngl.%2BJ.%2BMed.%26rft.volume%253D383%26rft.spage%253D874%26rft_id%253Dinfo%253Adoi%252F10.1056%252Fnejmms2004740%26rft_id%253Dinfo%253Apmid%252Fhttp%253A%252F%252Fwww.n%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [36]: /lookup/external-ref?access_num=10.1056/nejmms2004740&link_type=DOI [37]: #xref-ref-9-1 "View reference 9 in text" [38]: {openurl}?query=rft.jtitle%253DJ.%2BGen.%2BIntern.%2BMed.%26rft.volume%253D36%26rft.spage%253D464%26rft_id%253Dinfo%253Apmid%252F33063202%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [39]: /lookup/external-ref?access_num=33063202&link_type=MED&atom=%2Fsci%2F372%2F6540%2F348.atom [40]: #xref-ref-10-1 "View reference 10 in text" [41]: {openurl}?query=rft.jtitle%253DAdv.%2BChronic%2BKidney%2BDis.%26rft.volume%253D25%26rft.spage%253D41%26rft_id%253Dinfo%253Apmid%252F29499886%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [42]: /lookup/external-ref?access_num=29499886&link_type=MED&atom=%2Fsci%2F372%2F6540%2F348.atom [43]: #xref-ref-11-1 "View reference 11 in text" [44]: {openurl}?query=rft.jtitle%253DN.%2BEngl.%2BJ.%2BMed.%26rft.volume%253D384%26rft.spage%253D872%26rft_id%253Dinfo%253Adoi%252F10.1056%252FNEJMms2025768%26rft_id%253Dinfo%253Apmid%252F33406326%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [45]: /lookup/external-ref?access_num=10.1056/NEJMms2025768&link_type=DOI [46]: /lookup/external-ref?access_num=33406326&link_type=MED&atom=%2Fsci%2F372%2F6540%2F348.atom [47]: #xref-ref-12-1 "View reference 12 in text" [48]: {openurl}?query=rft.jtitle%253DProc.%2BNatl.%2BAcad.%2BSci.%2BU.S.A.%26rft_id%253Dinfo%253Adoi%252F10.1073%252Fpnas.1516047113%26rft_id%253Dinfo%253Apmid%252F27044069%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [49]: /lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NDoicG5hcyI7czo1OiJyZXNpZCI7czoxMToiMTEzLzE2LzQyOTYiO3M6NDoiYXRvbSI7czoyMjoiL3NjaS8zNzIvNjU0MC8zNDguYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9 [50]: #xref-ref-13-1 "View reference 13 in text" [51]: #xref-ref-14-1 "View reference 14 in text" [52]: {openurl}?query=rft.jtitle%253DHealth%2BServ.%2BRes.%26rft.volume%253D55%26rft.spage%253D777%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [53]: #xref-ref-15-1 "View reference 15 in text" [54]: {openurl}?query=rft.jtitle%253DJournal%2Bof%2Bbehavioral%2Bmedicine%26rft.stitle%253DJ%2BBehav%2BMed%26rft.aulast%253DWilliams%26rft.auinit1%253DD.%2BR.%26rft.volume%253D32%26rft.issue%253D1%26rft.spage%253D20%26rft.epage%253D47%26rft.atitle%253DDiscrimination%2Band%2Bracial%2Bdisparities%2Bin%2Bhealth%253A%2Bevidence%2Band%2Bneeded%2Bresearch.%26rft_id%253Dinfo%253Adoi%252F10.1007%252Fs10865-008-9185-0%26rft_id%253Dinfo%253Apmid%252F19030981%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [55]: /lookup/external-ref?access_num=10.1007/s10865-008-9185-0&link_type=DOI [56]: /lookup/external-ref?access_num=19030981&link_type=MED&atom=%2Fsci%2F372%2F6540%2F348.atom [57]: /lookup/external-ref?access_num=000262434000003&link_type=ISI

Unethical race-based healthcare proposal nothing more than 'political eugenics:' Professor

FOX News

College professor calls it'political eugenics' for hospitals offering care based on race on'Tucker Carlson Tonight' Some U.S. physicians appear to be embracing "political eugenics" with calls to normalize preferred health care for Black and Latino patients to allegedly address structural inequities, DePaul University professor Jason Hill told "Tucker Carlson Tonight" Tuesday. "I think for the first time, we are seeing hospitals that are being used by doctors as indoctrination centers and advancing their own politicized views that are using Critical Race Theory, which is a nefarious and quite racist kind of doctrine," Hill told host Tucker Carlson. Hill said doctors have been using hospitals as platforms to advance their political views on reparations and "bring it into play" by providing preferential health care based on the race of the patient. The effort "is really an inverted racial guilt system in the hospital centers and I think ... I would question whether or not they are violating the Hippocratic Oath under which they have sworn and pledged allegiance to build and construct their medical lives under," Hill said. Michelle Morse and Bram Wispelwey – both of whom have worked at a teaching hospital for Harvard University – called for "an antiracist approach to medicine."

Does a white doctor understand a black patient's pain?

Los Angeles Times

Does the blood of black people clot more readily than that of white people? Does a black person's skin generally have more collagen--is it thicker--than a white person's? Are black people better at detecting movement than white people, and do they age more slowly? If you are white and said yes -- or even maybe -- to any of the questions above, you are not alone in falling prey to false beliefs about physiological differences between white and black people. A new study reveals that in a group of 222 white medical students, half judged as possibly, probably or definitely true at least one of 11 false beliefs about racial differences.

AI Makes Strangely Accurate Predictions From Blurry Medical Scans, Alarming Researchers


New research has found that artificial intelligence (AI) analyzing medical scans can identify the race of patients with an astonishing degree of accuracy, while their human counterparts cannot. With the Food and Drug Administration (FDA) approving more algorithms for medical use, the researchers are concerned that AI could end up perpetuating racial biases. They are especially concerned that they could not figure out precisely how the machine-learning models were able to identify race, even from heavily corrupted and low-resolution images. In the study, published on pre-print service Arxiv, an international team of doctors investigated how deep learning models can detect race from medical images. Using private and public chest scans and self-reported data on race and ethnicity, they first assessed how accurate the algorithms were, before investigating the mechanism.

These Algorithms Look at X-Rays--and Somehow Detect Your Race


Millions of dollars are being spent to develop artificial intelligence software that reads x-rays and other medical scans in hopes it can spot things doctors look for but sometimes miss, such as lung cancers. A new study reports that these algorithms can also see something doctors don't look for on such scans: a patient's race. The study authors and other medical AI experts say the results make it more crucial than ever to check that health algorithms perform fairly on people with different racial identities. Complicating that task: The authors themselves aren't sure what cues the algorithms they created use to predict a person's race. Evidence that algorithms can read race from a person's medical scans emerged from tests on five types of imagery used in radiology research, including chest and hand x-rays and mammograms.