引導式測驗 · 中等 等級
醫療保健演算法種族偏見測驗
測試研究的成本代理、觀察到的差異、反事實分配結果以及推廣一項醫療保健演算法研究的限制。
問題 1 的 8
What did the studied population-health algorithm predict as a proxy for health need?
預覽每個問題8 問題
當您參加測驗時,就會出現答案和解釋。
- 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