Kvíz spojený s průvodcem · Střední úroveň
Kvíz stratifikace zdravotního rizika AI populace
Tests risk groups versus diagnoses, CMS terminology, access bias, intervention outcomes, calibration, and supportive use.
Otázka 1 z 8
What does a population-health risk score represent?
Náhled každé otázky8 otázky
Odpovědi a vysvětlení se zobrazují, jak budete v kvízu.
- What does a population-health risk score represent?
- A guarantee that a person will be hospitalized
- A complete picture of unmet need
- An estimate tied to a defined outcome and time point
- A diagnosis or measure of personal worth
- How does risk adjustment differ from risk stratification in CMS terminology?
- Stratification always determines payment
- Adjustment corrects measure comparisons; stratification groups populations for analysis
- They are identical terms
- Adjustment assigns a clinical diagnosis
- Why can low recorded utilization mean high unmet need?
- Few visits always mean low health need
- Models cannot use utilization data
- Access barriers disappear in claims data
- A person may face barriers that limit access to care
- How should teams use an elevated risk category?
- Offer appropriate support and keep individual assessment available
- Deny services to people with low scores
- Treat the score as a diagnosis
- Restrict access without review
- What should be evaluated after outreach based on a score?
- Only model size
- Only time to compute a score
- Who receives support, access outcomes, and whether the intervention helps
- Only prediction accuracy
- Why is calibration important?
- It measures user satisfaction only
- It guarantees equitable predictions
- It replaces subgroup analysis
- It checks whether predicted probabilities match observed outcomes
- What should happen if a patient’s status changes?
- Remove the patient from all care
- Update the risk estimate and reconsider the support plan
- Keep the old category indefinitely
- Wait for a vendor model retrain
- Which use of population-risk scores could cause harm?
- Using them to close access or deprioritize people without review
- Using them to offer outreach
- Comparing predicted and observed resource use
- Checking errors across groups