Capitation flips the direction utilization risk runs. Under fee-for-service, the plan pays the provider for every visit, test, or procedure delivered — more utilization means more paid claims. Under capitation, the provider is paid a flat PMPM rate for every member attributed to their panel, whether that member is seen zero times or ten times that month. The provider absorbs the utilization swing; the plan gets a predictable, fixed line item instead of a variable one.
The mechanics
A capitation rate is set per member, per month, for a defined scope of services — commonly primary care, sometimes a broader set including certain referrals or ancillary services. The plan pays that rate against its attributed member count, typically monthly, independent of claims activity. Services outside the capitated scope — specialist visits, hospital stays, anything the cap agreement doesn’t cover — still get billed and adjudicated as normal fee-for-service claims. This is the piece that trips people up: capitation replaces claim-by-claim payment for a defined slice of care, not the entire relationship between the plan and that provider.
A worked example
A primary care group is capitated at $30 PMPM for a panel of 2,000 attributed members. The plan pays $60,000/month, flat, regardless of how many of those 2,000 members actually walk in that month.
Compare that to what the same panel would cost under fee-for-service, at a realistic average of 0.3 primary-care visits per member per month — roughly 3-4 visits a year, in line with actual U.S. utilization — and $70 per visit:
2,000 members × 0.3 visits × $70 = $42,000/month
In this example the capitated rate runs higher than the fee-for-service estimate, not lower — the plan is paying a premium for predictability, not buying a discount. That’s the actual point of capitation: cap rates are priced actuarially to approximate expected fee-for-service cost, not to undercut it. What the plan gets in exchange is a fixed, predictable line item instead of a variable one, plus a transfer of utilization risk to the provider. If the provider manages the panel below the utilization the rate assumed, it keeps the upside; if utilization runs hotter than assumed, the provider carries that downside instead of the plan. Neither number is “correct” in isolation — they’re two different risk allocations for the same population, and which one actually costs less in a given month depends on how the panel behaves, not on which payment model was chosen.
Why running both models on one platform matters
Most self-funded plans and TPAs aren’t purely capitated or purely fee-for-service — they’re capitated for primary care and fee-for-service for everything else, often within the same provider network. That means a claims system has to do two different jobs simultaneously: route capitated-scope encounters to a monthly PMPM reconciliation instead of individual claim adjudication, while still adjudicating every non-capitated claim — specialist visits, hospital claims, referrals — through the normal eligibility, pricing, and cost-sharing sequence. It also still needs the encounter data from capitated visits, even though no claim payment is happening, because that utilization data feeds stop-loss reporting, care coordination, and accumulator tracking for the member’s deductible and out-of-pocket exposure.
A platform that can’t carve out capitated arrangements forces you to track them outside the system, which is exactly the kind of manual reconciliation an adjudication engine is supposed to eliminate, not create. Claimaro’s adjudication engine handles capitated and fee-for-service claims within the same rules framework, which matters directly for a TPA administering multiple client books that don’t all use the same payment model on the same network.