In a surprising twist, researchers have found that including race in colon cancer prediction algorithms may actually help reduce health disparities. A study published in the Proceedings of the National Academy of Sciences challenges the growing trend of removing race from clinical algorithms, revealing that race-adjusted models can lead to more accurate cancer risk predictions for Black patients.
The research team analyzed data from over 77,000 participants in the Southern Community Cohort. They discovered that Black participants were more likely to report unknown family histories of cancer, a crucial factor in predicting colorectal cancer risk. This data discrepancy resulted in family history being less reliable for predicting cancer risk in Black individuals.
When comparing race-blind and race-adjusted algorithms, the study found that race-blind models underpredicted cancer risk for Black participants and overpredicted risk for white participants. In contrast, the race-adjusted algorithm more accurately predicted risk for both groups. Notably, the race-adjusted model identified 74.4% of Black participants in the highest risk category, compared to only 66.1% with the race-blind approach.
Emma Pierson, senior author of the study, emphasizes the importance of carefully evaluating the impact of removing race from medical algorithms. While acknowledging the need to reconsider race in these tools due to potential biases and outdated data, Pierson views race-adjusted algorithms as a temporary solution.
The researchers stress the importance of improving medical data quality in the long term. As the medical community grapples with these issues, the findings suggest that a nuanced approach to race in clinical algorithms may be necessary to ensure equitable healthcare for all patients.
See “Considering race in colon cancer prediction algorithms reduces disparities, researchers find” (September 19, 2024)
