Geographical Variations in Medicare Spending FREE

In this issue, we have taken the unusual step of asking three experts to comment on the two lead articles. We did so because we thought that these articles contribute new and important information to our national dialogue about one of the great dilemmas of American life: the high cost of medical care. The articles provide insight into the reasons for variations in costs of care in different parts of the United States and help us to understand what we gain from greater expenditures. These articles are pertinent for us as internists because we decide whether to order many of the services whose frequent use distinguishes higher-spending regions from lower-spending regions.

We asked Kenneth Shine, MD, to discuss the implications of the study findings for physicians. Dr. Shine is a cardiologist. He recently completed 10 years as the president of the Institute of Medicine of the National Academies. He now works on health security issues.

Charles Phelps, PhD, discusses the issues from the perspective of a health economist. He helps us to understand the strengths and the limitations of the methods used in the two articles. Dr. Phelps is the provost of the University of Rochester in Rochester, New York.

We asked Gail Wilensky, PhD, to discuss the articles from the perspective of a Medicare administrator and health policymaker. She was the administrator of the Health Care Financing Administration (now the Center for Medicare & Medicaid Services) from 1990 to 1992. More recently, she chaired two expert panels that advise the U.S. Congress, the Medicare Payment Advisory Commission, and the Physician Payment Review Commission.

We expect these articles to generate a vigorous discussion, and we're grateful to our three editorialists for setting the stage.

Harold C. Sox, MD Editor

Location, location, location. As in real estate, what you pay for health care varies significantly by where you make the purchase. But there is no guarantee that the quality of the product will vary with price.

In this issue, Fisher and colleagues (1 - 2) have provided compelling evidence that 5-year mortality rates, functional status, and quality of care for three conditions (acute myocardial infarction, hip fracture, and colorectal cancer) do not vary significantly from high-cost to low-cost hospital referral regions (HRRs). Costs varied primarily by the number of consultations, tests, and hospitalization days rather than by the evidence-based services required. If anything, mortality rates were somewhat greater in the highest-cost areas. There was wide variation in use of intensive care unit beds, emergency intubations, and feeding tubes during the last 3 years of life. Influenza and pneumococcal immunizations and Papanicolaou smears were performed less frequently in regions with higher expenditure indexes. Patients in areas with higher expenditure indexes were more likely to see medical subspecialists, and those in HRRs with lower expenditure indexes were more likely to see family practitioners. Although the differences are small, the authors point out some evidence that access to care was poorer in higher-expenditure areas.

This study is consistent with evidence that the more hospital beds, physicians, laboratories, and subspecialists are available in a region, the more they will be used (3). Efforts in the United States to distribute physicians and beds more appropriately for the country's needs have met with limited success. However, our professional organizations must initiate efforts to help physicians and patients understand that more is not necessarily better. Where evidence is available for the proper approach to diagnosis and treatment, it should be applied appropriately. For conditions or circumstances in which evidence is not available, it must be collected. Evidence should drive diagnostic testing and treatment behaviors, and physicians must carefully examine the added value for the patient of each costly decision.

We know all too little about how clinical reasoning differs in regions with different health care expenditure patterns. Development of algorithms for treatment of a specific condition by physicians in high- and low-cost HRRs might be very revealing. Explicitly describing the decision tree (or clinical guidelines) would provide a mechanism for comparing how decisions vary from high- to low-expenditure areas. For example, do the indications for a procedure used in the care of a patient with myocardial infarction differ in different settings? The opportunity to observe the decision-making process for a number of standardized patients in various parts of the United States might characterize the process by which these remarkable variations come about.

The Institute of Medicine report Crossing the Quality Chasm(4) emphasized the need for “continuous healing relationships.” A physician providing such a relationship has a responsibility to see that the patient receives appropriate evidence-based care but does not receive care that adds no value. The physician can advise on the number and nature of the subspecialists consulted and the frequency of examinations, tests, and procedures. The electronic medical record can have an important role in reducing redundant and unnecessary tests. The patient's continuing-care physician often is aware of which subspecialists practice cost-effectively. This information should influence referral recommendations. General internists and family physicians can integrate the care process and provide preventive services, which, as Fisher and colleagues show, are not as well provided in areas with large numbers of subspecialists. These physicians can play a crucial role in the appropriate use of intensive care unit beds and procedures at the end of life. These responsibilities emphasize further the importance of the involvement of a generalist physician in the care of every patient.

Clues in the work of Fisher and colleagues suggest that poorer access to care may increase costs. Patients in the most expensive HRRs had the lowest “global” satisfaction and the highest interpersonal satisfaction with care. This suggests that a patient who went to the doctor in the most expensive areas had a more meaningful interpersonal experience but that there was overall discomfort with a system that has large numbers of subspecialists, teaching hospitals, and so on. Reliance on emergency departments rather than on a continuing-care physician may delay the initiation of care. If, in fact, access is decreased in more expensive areas, it may lead to higher costs for care of patients coming later to care—and perhaps a slightly higher mortality rate. May delayed access result in increased services? This hypothesis deserves further exploration.

As Fisher and colleagues point out, if all 306 HRRs had the same costs as the lowest-cost HRRs, as much as 30% of health care costs might be eliminated without adversely affecting health care outcomes. We should test this hypothesis rather than assume that it is true. In the absence of professional leadership, and in the face of rising health care costs, insurers and policymakers could respond with increased co-payments for visits and nonpayment for tests and procedures that are not part of the evidence-based requirements. It would be far better for physicians to reduce the rate at which health care costs are increasing through more efficient care than to invite the blunt bludgeons of purchasers and insurers. Medical educators can work closely with students and house officers to define and redefine the need for visits, consultations, tests, and procedures. The challenge comes when these trainees see their own role models, who are largely subspecialists, maximizing the flow of patients and income. Do patients in such situations have a “continuous healing relationship”?

Fisher and colleagues have convincingly demonstrated that excellent outcomes for patients can be achieved in regions that do less, but do it right. The challenge is to convince the public that this is not about rationing but about better care. The last time the profession ignored rising health care costs, we got “managed care.” Fisher and colleagues offer some clues that if carefully explored and elaborated might allow physicians to take a leadership role in controlling costs rather than wait for the next “market solution.”

Kenneth I. Shine, MD

RAND

Arlington, VA 22202-5050