Q&A: Reexamining racial factors in clinical algorithms

Key takeaways:

• Inappropriately factoring race into clinical decisions risks worse care for patients from underrepresented ethnic and racial groups.
• PCPs should consider if the tools they use include race as a factor.

A reexamination of factoring race into clinical equations for patient care will require changes at each step of the algorithm life cycle, according to an expert.

The Council for Medical Specialty Societies (CMSS) recently released a white paper that provides recommendations on how to reexamine the use of race in clinical equations.

The report came as a result of a June meeting in which the CMSS, alongside the Doris Duke foundation, the National Academy of Medicine and the Gordon and Betty Moore Foundation, invited leaders in medicine, policy, research and patient advocacy to “explore current field efforts, challenges, and timely opportunities to bolster actions for more rigorous consideration of race in the design, implementation, and monitoring of clinical algorithms,” according to the CMSS.

Healio spoke with Helen Burstin, MD, MPH, chief executive officer of the CMSS, to learn more about the report and factoring race into clinical decision-making.

Healio: Historically speaking, can you give examples of clinical algorithms that may have caused harm because they included race?

Burstin: There are several examples of clinical algorithms that may have caused harm because they included race. The most well-known example is the estimated glomerular filtration rate (eGFR) — the equation of kidney function. By including Black race in the equation, Black patients were more likely to be diagnosed with kidney failure later and, as a result, more likely to get on kidney transplant waitlists later. For pulmonary function tests (PFTs), the inclusion of race led Black patients to be underdiagnosed and diagnosed later with more severe pulmonary disease. And for children, the inclusion of Black race in a guideline for UTIs for young girls led to fewer Black girls getting the full evaluation they needed. Subsequent research has shown that race does not belong in the UTI algorithm and the factors that matter are a history of prior UTIs and the duration of fever that predicted UTIs in kids. Race is not a biological construct and should not be included in these guidelines and algorithms. By inappropriately including race, we risk worse care for patients of color and we potentially miss the opportunity to identify the clinical factors that drive risk.

Healio: Can you give examples of when the inclusion of race in an algorithm is beneficial? What makes it so?

Burstin: Inclusion of race in an algorithm is acceptable if it is intentionally included to reduce known inequities. The best example is the current effort to move Black patients up the kidney transplant waitlists after years of inappropriately being at the bottom of the transplant list due to the eGFR formula that included race.

Healio: In light of the conclusions of the report, how should practicing physicians think about the tools they use to make diagnoses and treatment decisions?

Burstin: First and foremost, we need practicing physicians and physicians-in-training to consider whether the tools they are using include race. I suspect many physicians would be surprised to know how many clinical guidelines and clinical algorithms inappropriately include race. It is very important that we teach medical students, residents and practicing physicians the historical bases for the inappropriate inclusion of race that often goes back hundreds of years to the time of slavery. If you see race in one of your clinical guidelines or algorithms, ask your institution or specialty society why it is included. If your labs are still using tests that include race, ask about how to accelerate implementation of revised algorithms. We will work with our specialty societies to ensure that updated guidelines and algorithms are disseminated and changed as fast as possible as the health of our patients is at stake.

Healio: What are the new recommendations moving forward here?

Burstin: To move forward, we need to change the whole algorithm life cycle — from the evaluation of evidence by specialty societies to develop guidelines to monitoring algorithms for bias by tech and health systems. With this broad-based approach, we can identify problematic guidelines and algorithms, update or adapt them, and implement updated guidelines and algorithms. We also need to build guidelines and algorithms with more representative data sets that include more diverse populations. And even if algorithms are changed, implementation does not happen overnight. For example, the changes to PFTs will require a major implementation lift. We need to offer specific recommendations about how to make change happen faster for physicians, labs and others.

Healio: What are some concrete steps that medical societies can take to address this issue?

Burstin: Specialty societies can review their clinical guidelines and algorithms and identify which ones include race. With funding from the Doris Duke Foundation, the American Academy of Pediatrics is laying out a methodologic approach that others will be able to replicate. Societies can also follow the example of some, including the American Society of Nephrology and the American Thoracic Society, and appoint a task force to examine the inclusion of race in the clinical guideline or algorithm and consider whether other variables would be more appropriate in its place. We also plan to bring specialty societies together as we did at our recent annual meeting to share best practices and learn how best to approach this issue. The report we recently released offers a series of steps for specialty societies.

Healio: What else should physicians know about factoring race into clinical decision making?

Burstin: Although some of the algorithms are being updated, the implementation fixes are still lagging. If you think your patient should be moving more quickly up the transplant waitlist once race is removed from the eGFR equation, ask questions and keep advocating for them. Physicians should keep in mind that payers may also factor race into their clinical decision-making algorithms. And some of these algorithms may not be very transparent. Don’t just accept the current algorithms. Please ask questions when patients are being denied services. Are they factoring in your patient’s race in their determinations? Push back and ask to see the factors they are using to make determinations.

Reference:

• Reconsidering race in clinical algorithms: driving equity through new models in research and implementation. https://cmss.org/reconsidering-race-in-clinical-algorithms-driving-equity-through-new-models-in-research-and-implementation/. Published Nov. 1, 2023. Accessed Nov. 17, 2023.