Massive Bio has made its inaugural strategic investment in Rivvi. Read the announcement

← Glossary

Definition · Value-based care

Risk adjustment factor (RAF) score

A risk adjustment factor (RAF) score is a number that estimates how costly a patient is expected to be. The average Medicare beneficiary scores about 1.0. CMS calculates it by adding demographic factors such as age, sex, and Medicaid status to the weights of the patient's documented HCCs. Higher scores mean higher expected costs.

What a RAF score is

A RAF score turns a patient's profile into one number. The average Medicare beneficiary is normalized to about 1.0. A patient at 2.0 is expected to cost about twice the average, and one at 0.6 well below it. Plans and ACOs track the average RAF across their whole population.

The score is additive. It starts with demographic coefficients for age, sex, Medicaid eligibility, and whether the patient lives in the community or an institution. Then it adds a coefficient for each documented HCC, applying hierarchies so only the most severe condition in a family counts. Some disease combinations add interaction terms.

In Medicare Advantage, a plan's payment for a member is roughly a base rate multiplied by the member's RAF score, with further adjustments. In ACOs, RAF scores adjust the spending benchmark. In both, a missed annual visit can quietly lower the score, because undocumented conditions drop out of the next year's calculation.

Questions

RAF scores, answered

How is a RAF score calculated?
Add the patient's demographic coefficients, such as age, sex, Medicaid eligibility, and community or institutional status, to the coefficients for each documented HCC. Hierarchies are applied first so only the most severe related condition counts. Add any interaction terms. CMS then normalizes and adjusts the total before payment.
What is a good RAF score?
There is no good or bad RAF score. It should reflect how sick the patient actually is. A score near 1.0 means average expected cost. What matters is accuracy: every condition the patient has should be assessed and documented each year, and nothing unsupported should be coded.
Can AI help find RAF gaps?
It can sort the data. In Rivvi, staff upload a claims extract or payer report. Then they ask which patients had a chronic condition coded last year but no visit this year. The action layer can then call or text those patients. Coding decisions stay with clinicians and coders.

Try it on your own data today.

Free to start. Most teams are using it the same day.