Risk Adjustment
Risk adjustment is the statistical method CMS uses to account for differences in patient characteristics when calculating home health outcome measures. Using OASIS and claims data, models predict each patient's expected outcome given their condition, and agencies are evaluated on observed performance relative to expected, so an agency serving sicker, more complex patients is not automatically penalized.
How risk adjustment works mechanically
For each risk-adjusted measure, CMS builds a predictive model from national data, using patient-level factors such as diagnoses, functional status, cognitive status, age, and prior care history drawn largely from OASIS. The model produces an expected value for each patient, for example the probability of improvement in bathing. An agency's observed rate is then compared with its expected rate, and the reported measure reflects that relationship rather than the raw rate alone. The practical consequence: two agencies with identical raw outcomes can earn different scores if their patient populations differ in measured severity.
Where risk adjustment shows up
Risk-adjusted measures run through the entire home health accountability stack: OASIS-based outcome measures on Care Compare, the Quality of Patient Care Star Rating, claims-based measures like acute care hospitalization, and the expanded HHVBP model, where performance drives payment adjustments up to plus or minus 5% of Medicare fee-for-service payments. HHCAHPS survey measures are adjusted for patient mix as well. Any time an agency compares its numbers with a benchmark or a competitor, the honest comparison is risk-adjusted, which is also why raw internal dashboards can disagree with publicly reported results.
Why documentation is the whole ballgame
Risk adjustment only sees what you document. If clinicians understate admission severity, or coders omit active comorbidities, the models compute optimistic expected values, and the agency is graded against a healthier phantom population. The fix is not gaming, it is completeness: accurate OASIS responses at admission, thorough and supportable diagnosis coding, and consistent documentation of functional and cognitive status. Deliberately exaggerating severity to inflate expected values is the mirror-image failure and a compliance risk. The goal is a record that reflects the patient you actually admitted.
Using risk adjustment intelligently
Operators can put risk adjustment to work:
- Compare observed vs. expected performance by measure to find true improvement targets
- Audit admission OASIS and coding completeness before blaming clinical performance for weak scores
- Educate referral sources that risk-adjusted scores make complex patients safe to send you
- Watch for measure-model updates in rulemaking, which can shift scores without any change in care
- Train QA to review severity documentation with the same rigor as payment items
Frequently asked questions
Is risk adjustment the same as case-mix adjustment under PDGM?
No. PDGM case-mix adjustment sets payment for a 30-day period based on patient characteristics. Risk adjustment calibrates quality measure expectations. They use overlapping data but serve different systems, one paying claims and the other scoring outcomes.
Which home health measures are risk-adjusted?
Most OASIS-based outcome measures, such as improvement in ambulation or self-care, and claims-based measures like acute care hospitalization are risk-adjusted. Process measures, which score whether the agency did something, generally are not.
Can an agency see the expected values used for its scores?
CMS reports typically show observed and risk-adjusted results for the agency's own patients, which lets you infer how your population's expected values compare nationally. Reviewing these reports regularly is one of the fastest ways to spot documentation-driven scoring problems.