Fee-for-service is the predominant payment method in healthcare, also seen in most service businesses. A discrete problem is followed by a service where the outcome is verifiable, and the relationship ends when the service is completed with the problem solved. While primary care does solve discrete problems, its scope of work extends beyond the constraints imposed by a fee for service payment modality. Primary care’s complete product has four parts: first contact, continuity, comprehensiveness, and coordination. Fee-for-service can price exactly one of them: the contact. The other three are where most of the value lives, and none of them has a billing code. 

Fixing rates does not fix the flaws of fee-for-service. Double what primary care physicians are paid per visit and every distortion described below survives intact. The argument is structural: five observations about how fee-for-service actually behaves in primary care, each verifiable on its own, all pointing the same direction. 

No Code, No Care 

In fee-for-service, a service exists only if it has a code. There are roughly ten thousand CPT codes, and everything a clinician does must be translated into one of them, documented to support it, and submitted to a third party for payment. Work that doesn’t map to a code is, economically speaking, volunteer work. Consider text based care. In direct primary care, texting is routine. Some examples include: patients sending a photo of a rash, asking whether a medication interaction matters, following up on a lab result. The issue is often resolved in minutes. In fee-for-service, text-based care is nearly absent, and sometimes overly problematic. A text-based exchange is generally either unbillable or billable only through e-visit codes whose reimbursement doesn’t cover the documentation they require. So the system funnels every question into the one unit it knows how to price: the office visit. The patient takes a half-day off work for something a text could have solved, or skips asking entirely. Office-based care is more of an artifact of the code book rather than a clinical decision. 

Winners and Losers 

When everyone bills a shared pool, someone has to decide what each code is worth. That someone is the RUC: the AMA’s Relative Value Scale Update Committee, a panel of 31 physicians, most of them representing specialty societies, that recommends relative value units (RVUs) to Medicare. Historically, CMS has accepted roughly 90% ofthose recommendations. A committee most Americans have never heard of is, in practice, the price-setter for American medicine. And because Medicare’s fee schedule is budget-neutral, valuation is a zero-sum game. A bigger slice for one specialty’s procedures means a smaller slice for everyone else. Procedures are easy to time, count, and survey; cognitive work, especially preventative work like the forty-minute visit that keeps a diabetic out of the hospital, is not. Look around and you can easily see which one the system values. Proceduralists out-earn primary care physicians by roughly two to one, medical students follow the money, and the country runs a chronic shortage in the one specialty consistently associated with lower total spending and better outcomes. The result shows up in the national ledger: the United States puts 5% to 7% of its health care dollar into primary care, about half the OECD average. The structure picks winners and losers, and primary care was never positioned to win. 

Every Visit an Invoice 

Fee-for-service puts a cash register in the exam room. While the patient describes their symptoms, the physician is running a parallel process: what level of visit is this, does the documentation support it, is that second concern a separate billable problem or a freebie? Meanwhile the patient runs their own: is this question going to show up on a bill? Both parties are assembling a transaction while trying to have a relationship. That friction changes the care itself. “While I’m here, can you also look at…” becomes a coding question instead of a clinical one. The follow-up call, the care coordination, and the five minutes spent actually explaining the plan are precisely the items with no line on the invoice, so they get squeezed. Any business leader recognizes that people optimize what the compensation system measures, and the compensation system here measures billable encounters, not health. The transaction itself isn’t free, either. Billing and insurance-related activities cost about $20 per primary care visit and over $99,000 per primary care physicianper year. Most industries process a payment for around 2% of its value. None of this overhead makes anyone healthier; it exists solely to move money through the fee-for-service pipe. In a capitation or DPC model, the question at the point of care collapses to one thing: what does this patient need? 

“What Will This Cost Me?” 

Ask any primary care physician about the patient who sat on chest pain for a week, or mentioned the lump they found “a few months ago.” Under fee-for-service with modern deductibles, every contact with the system carries an uncertain price tag, and patients respond the way rational consumers respond to uncertain prices: they wait. Gallupfound 38% of Americans delayed medical care over cost in 2022 and the figure has hovered around a third for two decades. Delay is the enemy of everything primary care is good at: prevention, early detection, and relatively cheap intervention. Capitation flips the psychology. Once the membership or per-member fee is paid, the marginal visit costs nothing, and sunk-cost reasoning starts working for the patient: “I’m already paying for this, might as well get it looked at.” That one sentence is an early-detection program. The blood pressure gets checked before the stroke, the mole gets seen before it’s a melanoma, the medication question gets asked before the ER visit. Prevention happens when patients stop doing cost-benefit math at the moment of a symptom. 

Billable Providers vs. Useful People 

Fee-for-service doesn’t just decide what gets paid; it decides who is allowed to matter. Revenue attaches to encounters with “billable providers,” so a nurse, health coach, pharmacist, or care coordinator is a cost center unless their work can be tortured into a claim through supervision and “incident-to” rules. They get hired reluctantly and cut first. This would be a reasonable design if primary care were still one doctor with a stethoscope. It isn’t. Modern chronic disease management is team work: titrating medications, coaching behavior change, chasing the follow-up, coordinating the specialist. Most of it doesn’t require a physician, and none of it requires a CPT code to be worth doing. The payment model, not the clinical evidence, is what determines whether team-based care is financially viable. In fee-for-service the answer is usually no. In a flat-fee model, the question inverts. Revenue is the membership, so anyone who keeps members healthy, answered, and retained is protecting it. The nurse who handles the texting queue isn’t a cost to be minimized; she’s the reason the practice works. Same nurse, same work, opposite economics. 

The Obvious Objections 

Anyone who lived through the 1990s will point out that capitation was tried, and HMO gatekeeping made it a dirty word. But the failure was in who held the fee, not the fee itself. The insurer kept the savings while patients were locked into networks, so stinting on care paid. In DPC the patient pays the practice directly and can leave any month. Under-treatment shows up as churn, not profit. As for potential overuse, primary care demand is self-limiting. Nobody sees their doctor recreationally, and heavy use of the relatively inexpensive service is precisely what prevents use of the expensive ones. 

The other objection is that fee-for-service just needs better codes. It has had thirty years of trying with chronic care management codes, transitional care codes, annual wellness visits, e-visit codes, MIPS, etc. Each patch added complexity without changing the unit. The portal-message codes produced surprise bills, not text-based care. Meanwhile, the counterfactual already exists. The Society of Actuaries’ evaluation of DPC found lower downstream utilization, driven by fewer emergency department visits, and most peer countries already pay for primary care wholly or partly by capitation. 

The Proof, Assembled 

Put the five observations together. Fee-for-service can only pay for what it can code. It prices those codes through a zero-sum committee that structurally favors procedures. It inserts a transaction into every clinical moment, then spends six figures per physician processing it. It teaches patients to delay the very care that is cheapest when delivered early. And it restricts the care team to whoever can generate a claim. Each is a flaw on its own. Together they demonstrate something simpler: the unit of payment is wrong. Primary care follows the gestalt principle where the relationship maintained over time is greater than the sum of all interactions. What to do with this depends on which side of the check you sit on. 

1. Buyers (Individuals, employers, benefits consultants, or plan sponsors): buy primary care the way it actually works: a flat fee per person for unrestricted access. Save fee-for-service for the care that really is transactional. 
2. Clinicians: understand that your payment model is your practice model. No workflow redesign, scribe, or burnout seminar will fix incentives that are pointed the wrong way. Changing the payment unit changes everything downstream.