Elliott Fisher and colleagues in their provocative paper published online December 5 validated an approach to quantifying the clinical and economic performance of physician communities clustered statistically around hospitals. Fisher describes the so-called extended hospital medical staff as “hospital-associated multispecialty group practices” or “virtual organizations.” While some physician markets do indeed function as “communities,” with norms and at least informal social control, in the vast majority of places, there is little “community.” These physician “communities” function more like colorful Darwinian ecologies, like a coral reef, with predators and prey, territories, hierarchies, pecking orders, and few rules of any kind. Holding the hospital accountable for the behavior of its community physicians outside the hospital is about as realistic as holding the Lebanese government accountable for the behavior of Hezbollah.

The hospital’s role in this ecology has been shrinking steadily for more than twenty years, as the advance of less invasive diagnosis and treatment has peeled off layers of clinical activity into physician- or investor-sponsored enterprises. Though hospitals have occasionally acted boldly in this space, their share of overall health spending has declined from 41 percent in 1980 to around 31 percent today.

A central flaw. High quality medical care requires good communication and accountability. The current fragmented payment method, which pays hospitals separately from physicians and pays multiple physicians separate fees for diagnosis or treatment of the same patient, fosters neither coordination nor accountability. Along with the piecework incentive inherent in fee-for-service, it is the central flaw in our current Medicare payment system.

The hospital/physician fee division within Medicare goes back forty-one years to the founding of the program, and perhaps thirty years beyond that to the separation of Blue Cross and Blue Shield, on whose template Medicare was built. Fee-for-service physician payment probably goes back to ancient Greece, or perhaps earlier. These payment silos are relics of an earlier medical age, not an adequate foundation for improving Medicare’s payment model.

Attempts to control costs. The current attempts to restrain the costs of Medicare’s Part B, a type of global budget based upon a so-called Sustainable Growth Rate, has become the health policy equivalent of an inflamed hemorrhoid, requiring the annual application of anesthetic and emollients. As procedure-oriented physicians continue running up the tab, growing their volumes at or near double-digit rates, the current SGR model produces annual, formula-driven, across-the-board reductions in physician fees that have virtually compelled congressional rescissions. The current SGR methodology does nothing to restrain economically motivated physician activism. Further, it redistributes wealth from primary care physicians to specialists, by cutting the former’s already paltry fees without restraining the latter’s MRI scans and arthroscopies. There must be a better way.

Fisher’s apparent intent is to use the hospital as a vehicle for limiting payment to both the institution and the community’s physicians. By limiting the total amount of money spent annually by Medicare in a specific physician community, Fisher presumes that both hospitals and physicians would reconsider both physician recruitment and facilities expansions that would lead to increased spending.

Abolishing the division between Part A and Part B and creating some form of unified, severity-adjusted payment methodology, based on per-episode-of-illness, has a lot of appeal. Focusing on the episode of illness (rather than Fisher’s annual communitywide spending) would be a superior method of shifting clinical/economic risk to providers compared to the much-reviled global capitation that proved unworkable in the 1990s. Like capitation, however, it would create incentives to minimize hospitalization as well as expensive diagnostic procedures and could potentially throw off “bonuses” for entities that could manage the care for less than the bundled payment.

Payment problems. However, to whom does Medicare write the check? If one writes it to the hospital and says, in effect, “You worry about how to get money to your physicians,” you’ve created a problem akin to giving the mayor of a town in Al Anbar Province in Iraq a trunkful of hundred-dollar bills and saying, “You pass it out.” At some point early in the process, you need armed guards and someone to start your car for you in the morning. If you write the check to a physician group practice, or a communitywide physician alliance of some kind, how do you assure that the hospital gets paid? The absence of robust organizations with sufficient governance backbone to enforce economic discipline is the fundamental flaw in any such effort, as earlier experiments along this line revealed.

Back in the early 1990s hospitals created “physician-hospital organizations,” in effect joint-venture revenue-gathering organizations, with their physicians. When the expected wholesale shift to capitation failed to materialize, these organizations withered. Those who collected capitated payment from health plans experienced serious problems. Weak governance and poor internal utilization controls doomed many of the organizations lucky enough to get contracts.

Possible solution. It may be that the solution is a two-track approach, with differential incentives to patients to use the more “organized” track. Medicare could continue paying on a per event (per admission, per visit, per procedure) basis, with separate Part A and Part B components, for a complex hospitalization, in which case the patient continues paying part of the bill through Medicare’s current structure of deductibles and limits. These payments could be systemically restrained with current methodologies (e.g., restrictions on hospital updates and on base Part B fees). Or patients could benefit from reduced cost sharing if they agreed to participate in a certified economic arrangement which received a unified, severity-adjusted lump sum for treatment of an episode of illness. (Of course, how one does this without creating windfalls for the patient’s Medicare supplement insurer would be a challenge. One could conceivably mandate that the savings be passed through to subscribers in premium rebates if the supplemental plans are to continue to be approved.)

One can share Fisher’s goals–better-coordinated care and reduced variation in practice costs (and complications)–without fully understanding how one operationalizes his idea of a unified community-based payment approach. Merely being able to conceive a statistical method for unified payments is simply the first step in implementing a complex and much-needed change in payment policy.