A RAF score (Risk Adjustment Factor score) is a number that predicts how much it will cost to treat a specific patient compared to the average Medicare patient. A score of 1.00 represents average expected costs. Scores above 1.00 indicate a sicker, more expensive patient, while scores below 1.00 indicate a healthier one. Medicare uses this number to decide how much money to pay insurance plans and providers for each person they cover.
Why RAF Scores Exist
Without risk adjustment, insurance plans would have a financial incentive to avoid enrolling people with serious health problems. A plan that attracts mostly healthy members would look profitable, while one serving sicker populations would struggle. RAF scores level the playing field by paying plans more for patients who genuinely need more care. Providers who treat complex, high-cost patients receive larger payments than those treating relatively healthy ones.
This system is central to Medicare Advantage, the private insurance alternative to traditional Medicare that now covers tens of millions of Americans. CMS (the Centers for Medicare & Medicaid Services) defines risk adjustment as “a way to help make sure doctors and other health providers are paid fairly for the people they treat.”
How a RAF Score Is Calculated
A RAF score is the sum of weighted factors tied to two things: who you are (demographics) and what conditions you have (health status). Each factor carries a numerical weight reflecting its estimated contribution to total healthcare costs. Add them all up and you get a single number.
The demographic baseline starts with age and sex. Older patients and, in some age brackets, female patients carry higher baseline weights. For example, in the marketplace risk adjustment model, a 30-to-34-year-old woman carries roughly double the demographic weight of a man in the same age range. Factors like whether someone qualifies for both Medicare and Medicaid can push the score higher, since dual-eligible patients tend to have greater healthcare needs.
The bigger driver of a RAF score, though, is medical conditions. Each qualifying diagnosis maps to a Hierarchical Condition Category (HCC), and each HCC adds a specific coefficient to the score. More conditions, and more severe conditions, mean a higher number.
What HCCs Are and How They Work
HCCs are groups of related diagnoses organized into a hierarchy. The “hierarchical” part matters because it prevents double-counting. If a patient has both a mild and severe form of the same disease, only the most severe version counts toward the score. This keeps the system from inflating costs for conditions that are really just different stages of the same problem.
HCCs are also additive across unrelated conditions. A patient with diabetes, heart failure, and chronic kidney disease would have each of those conditions contribute separately to their total score. The more distinct health problems a patient has, the higher the RAF score climbs.
A concrete example from the American Academy of Family Physicians illustrates this well. A 68-year-old woman with uncomplicated type 2 diabetes and hypertension might have a RAF score of 0.428, below the 1.00 average. But that same woman, with her diabetes reclassified to include diabetic nerve damage, plus documented congestive heart failure and morbid obesity, would jump to a score of 1.327. Same patient, same visit, dramatically different score based on the specificity of documentation.
Why Medical Coding Matters So Much
RAF scores depend entirely on what gets documented in a patient’s medical record and translated into diagnosis codes (ICD-10 codes). If a provider knows a patient has a condition but doesn’t document and code it properly, that condition won’t factor into the score. The patient looks healthier on paper than they actually are, and the plan receives less funding to cover their care.
This is why clinical documentation integrity has become a major focus in healthcare. The American Academy of Family Physicians recommends that every diagnosis be documented using the MEAT principles: each condition should be actively monitored, evaluated, assessed, or treated during the visit. Documentation that doesn’t meet this standard can be rejected in an audit.
Specificity is critical. Coding “diabetes” generically produces a lower score than coding “type 2 diabetes with diabetic polyneuropathy.” Providers are expected to capture the type of condition, its underlying cause, severity, location, and any associated complications. Preventive screenings for things like depression can also uncover additional diagnoses that affect the score.
One important detail that catches many practices off guard: RAF scores reset every year. Even chronic, lifelong conditions like diabetes or heart failure must be reported again each year during an encounter. A condition documented in 2023 but not re-documented in 2024 simply drops off the score.
What Different Score Ranges Mean
The 1.00 benchmark represents the average Medicare beneficiary’s expected cost of care for the coming year. In practice, most healthy patients fall well below 1.00, often in the 0.3 to 0.5 range. Patients with multiple chronic conditions routinely score between 1.0 and 2.0. Those with the most complex medical profiles, such as patients with advanced cancers, organ transplants, or multiple serious organ-system diseases, can score above 3.0 or even higher.
For Medicare Advantage plans, higher aggregate RAF scores across their membership translate directly into higher monthly payments from CMS. This creates a real financial stake in accurate coding. Undercoding means a plan is underfunded for the care its members actually need. Overcoding, on the other hand, draws scrutiny from federal auditors and can result in significant financial penalties.
The V28 Model Update
CMS periodically updates the model used to calculate RAF scores. The most recent major change is the transition to the 2024 CMS-HCC model, known as V28, which replaced the older V24 model. CMS phased this in gradually, and as of calendar year 2026, the V28 model is being used at 100 percent with no blending of older model results.
V28 reclassified and removed certain condition categories, changed some hierarchies, and updated the cost weights assigned to various diagnoses. For providers and plans, this means some patients’ scores shifted during the transition, even without any change in their actual health. Conditions that were previously counted may no longer qualify, and some that carried high weights may now carry lower ones. The practical effect is that organizations had to review their coding and documentation practices to align with the new model’s requirements.