Appendix C: How Medicare sets prices today, and how Japan and Maryland do it

This appendix supports the sections "How prices get set" and "How hospitals get paid" in the plan. It explains the machinery the plan replaces and the two working models it borrows from.

Where the price of a doctor's visit comes from

Until 1992, Medicare paid physicians on the basis of what they customarily charged, which meant that whatever a specialty had historically billed became the benchmark for what it would be paid. Prices for procedures, which had been set high when they were new and difficult, stayed high after they became routine. Prices for talking to a patient, which had never been high, stayed low. Congress commissioned a Harvard team led by the economist William Hsiao to build something more rational, and the result, the Resource-Based Relative Value Scale, became the basis of Medicare physician payment from 1992 and, because commercial insurers key their own contracts to it, the basis of nearly all physician payment in the country.

The scale works like this. Every medical service in the country, some 10,000 of them, carries a score called a relative value unit, describing how much physician work, practice expense, and malpractice risk that service entails compared to every other. A standard 15-minute office visit scores about a 2. Cataract surgery scores close to a 14. To turn a score into money, the Centers for Medicare & Medicaid Services multiplies it by a single dollar figure called the conversion factor, currently about $33, and then adjusts for the cost of doing business in the physician's part of the country. Every price in the schedule is therefore two numbers: a relative score, and a national dollar value per point.

Who sets the scores

The scores are supposed to reflect the real work in a service. In practice they are recommended by the Relative Value Scale Update Committee, universally called the RUC, a committee of about 30 physicians convened by the American Medical Association, most of them appointed by the specialty societies. When a service needs a value, the specialty society that performs it surveys its own members on how long the procedure takes and how much skill and stress it demands, presents the results to the RUC, and the RUC votes on a recommendation to CMS. CMS has historically adopted about 90 percent of the recommendations, and the reason is not deference. It has no data of its own to check them against.

The problems with this arrangement are well documented and have been for 20 years. The surveys are answered by the doctors who will be paid the resulting price, and they are answered by small numbers of them. Time estimates have repeatedly been found to exceed what the procedures actually take when someone measures them; studies that timed surgical procedures against their RUC-assumed durations have found the assumed times were overstated by a wide margin, in some cases so far that a surgeon's assumed working day exceeded 24 hours. The committee's membership is weighted toward procedural specialties, and the scale is a zero-sum system inside a fixed budget, so every point added to a procedure comes out of the value of everything else, above all the office visit that is the whole of primary care. The Government Accountability Office reviewed the process in 2015 and found that CMS lacked the data to evaluate the recommendations it was adopting.

Congress noticed. In the Protecting Access to Medicare Act of 2014 it gave CMS explicit authority and a small budget, about $2 million a year, to collect its own data on the time and resources involved in physician services. CMS commissioned studies, and the studies found the same thing the critics had, but the standing data collection the law envisaged was never built. The government still prices 10,000 services on the word of the people who bill for them.

How the scores get revised

Each year Medicare and the RUC screen billing data for services whose circumstances have changed, for example a procedure that used to be done in a hospital and is now done in an office, and flag them as potentially misvalued. Anything flagged goes back to the specialty society that bills it, which surveys its own members again and proposes a new number. The process catches prices that have drifted. It never catches prices that were wrong to begin with, because the baseline it drifts from is the RUC's own earlier estimate, and the correction still depends on input from the doctors whose incomes it sets.

The result is a schedule that has been carried forward, with patches, for more than 30 years, and that has never been rebuilt from measured data.

Why the dollar value matters as much as the scores

The relative scores decide how the money is divided among services. The conversion factor decides how much money there is. Since 1992 the conversion factor has been set by formula, then by a formula that Congress overrode every year for more than a decade because it would have cut payments sharply, then by a statutory schedule of small annual updates since 2015. In real terms it has fallen for most of that period, which is why physician groups regard the schedule as a cut every year even as the money paid for procedures keeps rising through the scores. Nobody negotiates it in the open. It is the product of a law, a formula, and an annual lobbying fight over the patch.

What the plan does instead

The plan keeps the structure, a relative scale and a dollar value per point, because it is sound, and replaces the inputs.

The scores come from a public survey. CMS runs a standing survey of how long procedures actually take and what they demand, with teams observing procedures in a representative sample of hospitals and clinics, timing them, and publishing the data. This is the data collection Congress authorized in 2014 and never built, at the scale it needs.

Every score expires. Each relative value is rebuilt from the new public data on a fixed two-year cycle rather than carried forward. A price that was wrong in 1992 cannot survive to 2032.

Everything is published. The survey data, the sample, the methodology, and the reasoning behind every value, released before the new scores take effect.

The dollar value is negotiated in public. On the same two-year cycle, for the whole country, in open session between the payer and the organized bodies representing doctors and hospitals, inside a national budget for what the system will spend. One price per service, whoever pays and whoever is treated.

Japan: the model for repricing everything on a schedule

Japan has run its physician and hospital fee schedule this way for decades. Every service in the country carries a point value and every point is worth 10 yen, for every patient and every insurer, in every hospital and clinic. Every two years the Central Social Insurance Medical Council, a body called Chuikyo made up of representatives of payers, providers, and the public interest, revises the entire schedule. The revision has two parts. The government first sets the overall change in the money available, which is a political decision made in the budget. The council then reprices individual services within that total, working from a standing government survey of what hospitals and clinics actually earn and spend, so that a service whose cost has fallen can be marked down and a service the system wants more of can be marked up. The council can and does cut the price of a procedure whose technology has matured, and it has used the schedule deliberately to steer the system, for example by paying more for services delivered in the community than in a hospital bed.

Two features matter for the American design. Every price is reexamined on a known date, so no distorted value can hide, and the argument about what a service should be worth happens on the record, between organized interests, inside a budget. Japan has the oldest population in the world and spends about $5,365 per person on health care, and the schedule is a large part of why.

Maryland: the proof that global hospital budgets work in America

The plan pays hospitals a negotiated annual budget rather than a fee for each claim. That is how most of the developed world pays hospitals, and one American state has done it for a decade.

Maryland is the only state that sets hospital rates for all payers. Since 1977 a state commission, the Health Services Cost Review Commission, has set the price every hospital in the state may charge, and Medicare, Medicaid, and every commercial insurer pay the same rate, under a federal waiver. In 2014 the state went further. Under an agreement with the federal government, every Maryland hospital moved to a global budget: a fixed annual revenue, set in advance, adjusted for the population it serves, regardless of how many admissions it has. A hospital that reduces avoidable admissions keeps the money. A hospital that fills beds it did not need to fill earns nothing extra. In 2019 the model was extended to cover the total cost of care for Maryland's Medicare beneficiaries, including physician and post-acute services.

The results have been studied closely. Hospital spending growth in Maryland has been held below the national rate, Medicare has saved money against its own projections, and avoidable admissions and readmissions have fallen. Rural hospitals in Maryland have not closed at the rate they have elsewhere, because a global budget does not punish low volume.

The point for the plan is not that Maryland is perfect. It is that global budgets are administratively possible here, inside Medicare's own rules, with a state commission of modest size, and have been running for more than a decade. A hospital paid a monthly budget has no use for a 200-person revenue-cycle department, which is why global budgets are also where the billing industry ends.