Guide-linked quiz · Medium Level
Racial Bias in Healthcare Algorithms Quiz
Tests the study’s cost proxy, observed disparity, counterfactual allocation result and limits on generalizing one healthcare algorithm study.
Question 1 of 8
What did the studied population-health algorithm predict as a proxy for health need?
Preview every question8 questions
Answers and explanations appear as you take the quiz.
- What did the studied population-health algorithm predict as a proxy for health need?
- A patient’s preferred clinic
- The number of available physicians
- Race itself
- Future healthcare costs
- At the same risk score, what did the researchers observe about Black patients?
- They used fewer medications by design
- They were considerably sicker than White patients
- They had lower disease burden in every measure
- They were excluded from the algorithm
- Why did predicting costs understate Black patients’ health needs in the studied setting?
- The model used only age
- The system directly prohibited Black patients from receiving care
- Historical differences in access meant Black patients generated lower costs at similar illness levels
- Costs are never related to health
- What happened to the share of Black patients identified for extra help when the study’s target was changed?
- It increased from 17.7% to 46.5%
- The study did not report a comparison
- It fell from 46.5% to 17.7%
- It stayed exactly the same
- What was the key proxy mismatch in the study?
- A skin-tone scale was used as a diagnosis
- The algorithm predicted race rather than costs
- Healthcare spending was used instead of a direct measure of illness need
- The model used only clinician notes
- Does the 2019 study prove that every healthcare algorithm is racially biased?
- Yes, every model uses cost
- Yes, because it was published in Science
- No, because the result was only theoretical
- No; it studied a particular algorithm, target and population
- Why can a model be accurate on its chosen target yet still fail to allocate care fairly?
- The target may predict spending well but poorly represent clinical need
- Clinical need is the same as spending by definition
- Accuracy always guarantees equal access
- A risk score cannot affect care
- What should a health system examine when auditing a similar support-allocation model?
- Whether the model contains a race field
- Whether the target reflects clinical need and how outcomes differ across groups
- Only overall model accuracy
- Only the user interface color
Keep testing yourself
More quizzes picked for your level.