Search papers, labs, and topics across Lattice.
The paper introduces FRAME, a two-step framework designed to audit fairness bias in medical imaging by separating sampling variation from representational causes. By deriving a fair-model reference and testing the remainder with specific operators, the authors demonstrate that a significant portion of observed subgroup performance differences can be attributed to sampling variation rather than inherent bias. Their analysis of 702,206 images reveals that the fair-model reference accounts for a median of 41% of race differences and 22% of age differences, highlighting the importance of distinguishing between true bias and sampling effects in fairness evaluations.
FRAME reveals that up to 41% of racial performance differences in medical imaging may stem from sampling variation rather than actual bias, challenging conventional fairness assessments.
Subgroup performance differences are the standard evidence for fairness bias in medical imaging, and the usual response removes the demographic information that a model encodes. Here we introduce Fair-model Reference And Mechanism Evaluation (FRAME), a two-step framework for auditing such a claim. The first step derives a fair-model reference, the distribution of the difference under exact fairness at the observed subgroup sizes. In the second step, we test the remainder with two operators in representation space. One operator cannot change a within-group ranking by construction. Across 702,206 images and 36 encoders, the reference accounts for a median 41% of the reported race difference and 22% of the age difference. Injecting demographic decodability leaves the remainder unchanged, while entangling the group with the disease direction raises the race difference from 0.077 to 0.118. No intervention we tested changes the remainder more than a change of random seed does. Those interventions reduce a difference at the operating point and leave the within-group ranking difference at a median of 0.000. Applied to 89 differences in 9 published studies across 6 medical imaging modalities, the reference accounts for a median 25% of a rate difference and 70% of a difference in the area under the receiver operating characteristic curve. Image-text pretraining instead raises worst-group performance by about 0.05. Applying FRAME before choosing an intervention could distinguish differences that need a mechanistic explanation from differences compatible with sampling variation at the current cohort sizes.
