Dr. Atul Gawande has provided a chilling description of the problems facing true health reform in his recent New Yorker article. In The Cost Conundrum he describes how medical care is provided in McAllen, Texas, which is second only to Miami as the most expensive healthcare market in the country. McAllen’s per capita expenditures are twice those in El Paso, Texas, a city with similar demographics.
There are no good reasons for the differences. McAllen’s population isn’t demonstrably sicker and the care isn’t measurably better. There is also little understanding among the participants about what causes the higher spending. What is chilling is how easy the medical care environment in El Paso could become like McAllen’s.
Gawande refers to the accountable care organization (ACO) concept proposed by Elliott Fisher and colleagues at Dartmouth University. They propose that physicians whose practices are focused around a specific hospital be given incentives to lower the overall costs of patient care.
Payer Costs are Provider Revenues
The ACO has merit as a goal, but the challenge is in forming them. Getting very intelligent people such as physicians and hospital administrators to change their behaviors, especially if such changes may reduce their income, will be difficult. We need ways to encourage voluntary participation of both physicians providing care in the hospital and those who decide who should be hospitalized.
The Dartmouth data show that in areas like McAllen, there is much more interventional work, such as tests, procedures and admissions, than in areas like El Paso. With more access to, and time with, primary care physicians there is less need for interventional work. This means redistributing resources from the interventionists to primary care clinicians.
It is hard to imagine a new ACO with interventionists and primary care physicians achieving this redistribution. The interventionists often wield scalpels and have a ready ally in the hospital that depends on them to keep beds filled.The Answer Lies in Separation, Not Amalgamation
Interventionists should partner with the facility in which they do most of their work. Elsewhere, I describe these new care delivery teams (CDTs) that are effectively the inpatient side of Fisher’s ACOs. CDTs would be voluntary associations of a facility (usually a hospital) and those physicians whose work depends on the facility.
Unlike Fisher’s ACO, the CDT specifically excludes office-based physicians responsible for the ongoing treatment of patients. The CDT also need not include all eligible physicians at the hospital, just the voluntary paticpants.
The CDT may be a single entity with physician employees or a loose collaboration of independent physicians and a facility, collectively deciding its own governance rules. The key is that the CDT takes responsibility for an episode of care at a fixed price. Physicians might be compensated by salary, fee-for-time, or fee-for-service and may share in the gains or losses of the CDT.
CDTs will focus on how to provide inpatient care more efficiently and at higher quality. (Quality measurement is critical in any reform; see my overall proposal. Savings will be achieved not through lower net provider income, but through better management and clinical decisions. For example, instead of routinely repeating imaging, radiologists may review well-done MRI and CAT scans done elsewhere. Orthopedists can agree on the necessary implants, allowing the hospital to strike better deals with suppliers. Nurses may be empowered to implement routine procedures reducing infection rates. Lowering Interventional Costs and Rewarding High Quality Care
CDTs by themselves will not solve the key problem identified by Gawande — the overuse of interventional services. To address that problem, we need to redirect patients toward those physicians who provide high quality care at lower overall cost. This can be achieved by combining (1) a mechanism shifting resources from interventional care to effective outpatient management with (2) a way to identify those physicians who provide such effective care.
A comprehensive realignment of the payment system can accomplish this, but in the interim, a voluntary major risk pool (MRP) can move us in the right direction. The MRP covers hospitalizations and chronic illness. This coverage for insurers eliminates costly underwriting. The MRP, however, is not simply reimbursing plans for expenses incurred; it directly offers attractive bundled payments to CDTs. These episode-based payments allow CDTs to do what they do best—high intensity acute care—and reap increased income. Higher provider incomes within CDTs are not inconsistent with lower costs to the MRP as the CDT reduces the resources needed from suppliers outside the CDT.
The MRP obtains electronic copies of claims from the insurers who are its clients and from Medicare, more information than the Dartmouth group has. After linking all the data and substituting coded identifiers, the MRP will make available the data under arrangements ensuring patient confidentiality.
The Power of the Electronic Matchmaker
Insurers and others accessing the MRP data will see there are local providers with efficient practice patterns, but not their names. An intermediary trusted by physicians will serve as an electronic “matchmaker,” transmitting messages from insurers seeking efficient physicians. By remaining anonymous until a “deal is struck,” efficient physicians will negotiate better remuneration—probably not just higher fees, but payment for ongoing patient management, telephone and e-mail consultations, and other innovations. Some physicians may band together, perhaps by sharing electronic medical records, forming real or virtual group practices—the outpatient component of the ACO.
The major risk pool is the mechanism reallocating dollars. More effective chronic illness management will lower admission rates and the MRP will transfer more dollars to those health plans directing more patients to efficient ambulatory care providers. To find those providers, health plans will negotiate better payment arrangements. To steer patients towards those providers, plans will provide new incentives and sources of information. We can create what Fisher and Gawande have in mind, as long as we think about how to manage the transition.
McAllen and El Paso are almost 800 miles apart—a long day’s drive. To move away from the expensive McAllen model of care, we need not just a destination but a plan how to get there. The self-interest of the players is currently driving us in the wrong direction. By harnessing that self-interest with realigned incentives we can reform the system. Without taking account of the incentives, we will never get to where we need to go.
Harold S. Luft is Professor Emeritus in health policy at University of California, San Francisco, and author of Total Cure: The Antidote to the Health Care Crisis (Harvard University Press, 2008). More information is available at www.haroldluft.com.