Smaller than expected benefit partly due to most individuals receiving same decision, regardless of race being included in model<\/p>\n<\/h3>\n
By Elana Gotkine HealthDay Reporter<\/b>
\n<\/b><\/p>\n
TUESDAY, Dec. 3, 2024 (HealthDay News) — The clinical benefit of race-aware over race-unaware disease risk predictions is smaller than expected, according to a study published online Dec. 3 in the Annals of Internal Medicine<\/em>.<\/p>\n Madison Coots, from Harvard University in Cambridge, Massachusetts, and colleagues present a decision analytic framework for considering the potential benefits of race-aware over race-unaware risk predictions. A cross-sectional study involving U.S. adults was conducted using cardiovascular disease, breast cancer, and lung cancer as case studies. Race-unaware predictions were generated via statistical marginalization, starting with risk predictions from clinically recommended race-aware models. The utility gains of the race-aware over the race-unaware models were estimated based on a simple utility function assuming constant costs of screening and constant benefit of disease detection.<\/p>\n The researchers found that compared with the race-aware predictions as the benchmark, the race-unaware predictions were substantially miscalibrated across racial and ethnic groups. At the population level, the clinical net benefit of race-aware over race-unaware predictions was smaller than anticipated. This result stemmed from two patterns: (1) 95 percent or more of individuals would receive the same decision across all three diseases, irrespective of whether race and ethnicity are included in the risk model; and (2) the net benefit of screening or treatment was relatively small for those who received different decisions because these patients have disease risks close to the decision threshold. Race-aware models may have a more substantial net benefit when used to inform rationing.<\/p>\n “Our results suggest that race-aware risk models yield smaller gains in net benefit over race-unaware models than the improvement in predictions might suggest,” the authors write.<\/p>\n Abstract\/Full Text (subscription or payment may be required)<\/a><\/p>\n Editorial (subscription or payment may be required)<\/a><\/p>\n