Capitation Contract Basics: Reading a Risk Transfer
Read a capitation agreement as a price and the important terms look like boilerplate. Read it as a transfer of risk and the same clauses become the whole document: what the payment is buying, who is counted as covered, and what happens when the care costs more than the rate assumed. Federal regulation reads it the second way — it puts a ceiling on how much risk can move before the plan has to protect the practice against it.
Updated 9 min read
On this page
Key takeaways
- The payment is earned by a patient being enrolled, not by anything being done. Revenue becomes a function of the roster rather than of activity.
- Scope is the term that decides everything else: whether the payment covers only the practice's own services, or also services it refers out, or all medical services.
- Reconciling the roster is a billing function. A patient who left the panel but stayed on the list is the mirror image of an unbilled claim, and retroactive adjustments make it a running task rather than a monthly one.
- Where an arrangement puts a physician at risk for referral services beyond a defined threshold, federal regulation calls that substantial financial risk and requires the organization to assure stop-loss protection.
- That obligation is the plan's, not the practice's — but nothing makes it visible unless the practice knows to look for it in the agreement.
- The definition of a physician incentive plan is broad enough to catch withholds and bonuses too, not only capitation. The test is the effect on services, not the label on the payment.
- Encounters still have to be recorded even though they no longer generate revenue. Without them the practice cannot tell whether the rate was adequate, which is the only argument it will have at renewal.
What changes when the payment stops following the service
Under fee-for-service the claim does three jobs at once: it bills, it records what happened, and it is the thing the payment is checked against. Capitation breaks that bundle. The payment is a set amount per patient per period, made to cover a specified set of services without regard to how many services are actually furnished — so it is earned by coverage, not by care, and there is no claim for it to be checked against.
Two consequences follow immediately, and both land on the billing operation rather than on the clinicians. The roster of covered patients takes over the role the remittance played, because it is now the only thing the revenue can be verified against. And the encounter record loses its financial function while keeping every other one — nobody is paid for recording it, and the practice cannot survive a renewal without it.
This is the contract, not the posting
Term one: what the payment is actually buying
Scope is the term everything else hangs off, and the word “capitation” by itself settles none of it. The regulatory definition is explicit that the services covered may include only the physician's own services, or extend to services the physician refers out, or reach all medical services. Those are three very different businesses sharing one label.
| What the payment covers | What that means for the practice |
|---|---|
| The practice's own services only | The risk is utilization of the practice's own time and capacity. Cost is largely within its control, and the failure mode is a panel that is larger or sicker than the rate assumed. |
| Its own services plus services it refers out | The risk now includes decisions made by other people. This is the arrangement federal regulation is most concerned with, and the one that can put a practice at substantial financial risk — with the protections described below attaching as a result. |
| All medical services, professional and institutional | Usually described as global capitation. The practice is carrying a share of costs it does not generate and often cannot see in real time, which makes data access from the plan a term rather than a courtesy. |
Whatever the scope, the carve-outs are the other half of it. Services excluded from the capitation are paid some other way or not at all, and a service that is neither clearly inside nor clearly carved out is the one that produces an argument twelve months in.
Term two: the roster, and why reconciling it is billing work
If the payment follows enrollment, then the list of enrolled patients is the invoice. It is produced by the plan, it changes constantly, and it is frequently wrong in both directions — patients who have left still appearing, patients who joined not yet showing. Neither error announces itself, because there is no denial and no remittance to read.
Reconcile the roster against the practice's own record
Someone has to compare who the plan says is covered against who the practice believes it is responsible for. A patient on the roster the practice has never seen is a question; a patient being treated as covered who is not on the roster is a service that will not be paid for by anyone unless it is caught.Establish how retroactive changes are handled
Enrollment moves backward as well as forward, and the agreement should say what happens when it does — whether payments are adjusted, over what period a correction can reach, and how the adjustment appears. Without that term, retroactive terminations arrive as unexplained reductions.Know how patients are attributed in the first place
Attribution methods differ and are rarely negotiable, but they are always knowable: whether a patient is assigned by their own selection, by an algorithm reading claims history, or by default. It decides who lands on the panel and therefore what the rate has to cover.Keep recording encounters
This is the discipline that quietly erodes, because nothing is paid for the record and nobody chases a missing one. The practice needs encounter data to know what its covered population actually cost, which is the only evidence it will have when the rate is renegotiated — and plans commonly require it for their own risk and quality reporting regardless.
The metric problem, worth naming before it distorts something
Term three: risk, and the ceiling federal regulation puts on it
This is the part that is genuinely regulated, and it is the part practices most often do not know to ask about. Where a plan's compensation arrangement could have the effect of reducing or limiting the services provided to its enrollees, federal regulation treats it as a physician incentive plan — and the definition is deliberately about the effect rather than the label, so withholds and bonuses fall inside it just as capitation does.
Within that framework there is a defined threshold. Risk for referral services — the services the physician sends out rather than performs — beyond that threshold is what the regulation calls substantial financial risk. Several routes get there: a withhold above a defined level, a bonus above one, the two combined, capitation whose payments vary by more than a defined amount, and a catch-all for any other arrangement with the potential to hold the physician liable beyond the same line.
And the consequence is a protection the plan owes the practice
Two honest limits on that. The framework above is Medicare Advantage regulation — it does not automatically reach a commercial capitation agreement, though the structure it describes is the right shape to look for in one. And the specific thresholds, coverage shares and panel-size deductibles are numbers in a regulation that is amended: they are read from the current source, not from an article. What is stable is the architecture — a threshold exists, crossing it triggers an obligation, and the obligation sits with the plan.
The practical version, for a practice looking at a draft: find the clause that describes what happens when utilization exceeds the assumption. If there is none, that is the finding. Stop-loss, risk corridors, reinsurance and a reopener on the rate are the mechanisms that answer it, and an agreement that transfers risk with no mechanism at any of those points has transferred all of it. Reading a Payer Contract covers locating the clauses; this is the one to locate first.
Common questions
Is capitation the same thing as value-based care?
No, and conflating them causes a specific error. Value-based care is a category defined by tying payment to quality relative to cost. Capitation is a payment mechanism: a set amount per covered patient per period. A capitation arrangement with no quality measurement in it is not a value-based arrangement — it is a fixed budget with utilization risk attached — and it can be a perfectly rational deal on those terms. What matters is reading which of the two a given agreement is, because they are renegotiated on completely different evidence.
What is the single most important term to check?
Scope: exactly which services the payment covers, and what is carved out. The regulatory definition itself allows for three very different arrangements under one word — the practice's own services, its own plus services it refers out, or all medical services — and the risk profile is not comparable across them. Everything else, including how much rate is enough, is downstream of settling what the rate is buying.
Are we entitled to protection if utilization runs high?
In Medicare Advantage there is a defined answer: where an incentive arrangement puts a physician or group at substantial financial risk for referral services, the organization must assure that they have either aggregate or per-patient stop-loss protection, with per-patient deductibles determined by panel size. The obligation runs to the organization rather than to the practice. Outside that program there is no general entitlement, and whether an agreement contains stop-loss, risk corridors or reinsurance is a question for that agreement — but the absence of any such mechanism is itself the answer about how much risk is being transferred.
Why keep coding encounters if we are paid the same either way?
Three reasons, and only one of them is the payer's. The practice cannot tell whether the rate is adequate without knowing what the covered population actually consumed, and that evidence is the entire basis of any argument at renewal. Plans commonly require encounter data for their own risk and quality reporting, and an agreement may make it a condition. And a practice that stops recording because nothing is paid for it loses the ability to reconstruct its own history at exactly the moment it most needs it.
How should capitated revenue appear in our metrics?
Separately from fee-for-service revenue. Collection-rate measures are built on the idea that a claim was submitted for an amount and some proportion of it arrived; a per-member-per-month payment has no such denominator, so blending the two produces a figure that describes neither population. Keeping the capitated panel in its own view — and measuring it on cost against the rate rather than on collections — is the version that stays interpretable.
Key terms in this article
Defined once, on their own pages.
Continue learning
What the payment does when it lands, and the clauses around it.
Posting Capitation Payments
Why it is not a payment on a claim, and what treating it as one does to the A/R.
Value-Based Contract Basics
The other arrangement this one is confused with, and the federal definition that separates them.
Reading a Payer Contract
Locating the clauses this article says to look for, in your own agreement.
Payer Contract Renegotiation
Where the encounter data a capitated practice keeps recording eventually gets used.
Terminating a Payer Contract
The exit, and the obligations that outlive it — including for patients mid-course.
Payer Contracts & Reimbursement
The cluster: what clause types exist, what each governs, and how to find yours.
Authoritative sources
- 42 CFR § 422.208 — Physician incentive plans: requirements and limitations (opens in a new tab)
Defines a physician incentive plan as any compensation arrangement that may directly or indirectly have the effect of reducing or limiting the services provided to a plan enrollee; defines capitation as a set payment per patient per unit of time covering a specified set of services without regard to the number of services provided, and notes that the services covered may be the physician's own, may include referral services, or may be all medical services; defines global capitation as covering both professional and institutional services; sets a threshold above which risk for referral services is substantial financial risk, and lists the arrangement types that reach it; and requires the organization to assure that physicians at substantial financial risk have either aggregate or per-patient stop-loss protection, with per-patient deductibles determined by patient panel size and patients permitted to be pooled.