Dissecting Racial Bias in an Algorithm Used to Manage the Health of Populations.¶
Obermeyer, Z., Powers, B., Vogeli, C., & Mullainathan, S. (2019). Dissecting Racial Bias in an Algorithm Used to Manage the Health of Populations. Science, 366(6464), 447-453.
Cited by¶
1 citation across 1 artifact.
Each citation links to the sentence it supports in the citing article.
Primes¶
- Garbage In, Garbage Out
- Investigation traces the failure not to the model but to the training target: the proxy for "future medical need" was "future medical-care spending," which diverged from need along access lines, so the input's validity floor-bound the output's fairness, and no model sophistication could fix it; the fix was to redefine the target — an input intervention.
This sourceA clinical risk algorithm under-served one population because its label (future medical-care spending) was a biased proxy for medical need; the fix was redefining the target, an input intervention no model sophistication could supply.
SupportedVerified against the work's full text
Obermeyer et al. trace the bias to the cost label standing in for health need under unequal access, and say the remedy is new labels, not new fitting methods.
“we fit algorithms (for instance, by using a new statistical technique for decorrelating predic- tors with race or other similar solutions) is not required. Rather, we must change the data we feed the algorithm—specifically, the labels we give it.”
- Investigation traces the failure not to the model but to the training target: the proxy for "future medical need" was "future medical-care spending," which diverged from need along access lines, so the input's validity floor-bound the output's fairness, and no model sophistication could fix it; the fix was to redefine the target — an input intervention.
Verification¶
Does it exist? Confirmed. This work's DOI resolves to a registered record, which fixes its identity. That is all it fixes.
Does it back the claim? Read against the text for 1 of 1 citation: 1 supported. Each verdict is shown under its citation below, with what in the work backs the sentence.
Support is checked per citation rather than per work — the same source can be cited soundly in one article and wrongly in another. Per-citation recording began recently, so a citation with no recorded check is a gap in the record rather than evidence it went unchecked.
See how references were verified.
Registry ID ref:2a514fe6279d · see in the full table